Review: Zanflare B3 bicycle light- Mtbr.com
Page 1 of 2 12 LastLast
Results 1 to 200 of 232
  1. #1
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850

    Review: Zanflare B3 bicycle light - prototype

    Let make it clear, I've got sample light from Zanflare directly in order to make unbiased preproduction review. Product is published at their web page:
    https://www.zanflare.com/copy-of-vct...-led-motion-se

    They also decided to give it a try at crowdfund site:
    https://www.indiegogo.com/projects/b...ht-owl-bicycle

    I'm not here to decide if this is right business model or not I just want to inform the readers where they can find official product data.


    First impressions

    The light is nice looking, fins edges are nicely rounded so it feels good in the hands. As can be seen on pictures (currently not mine yet) the wire is 16AWG (overkill) and has SolarStorm type of connector which I don't like, still it fits on tipical battery pack. The mount is GoPro style.

    Modes operating. Not tested everything yet. Works more or less as described, except....see below.

    Back button
    turns leds on. Quick press swithes between one, two or three leds. Long press turns leds off and the last stage of leds is remembered even if accidentialy unpluged . There is small flaw. Immediately after turning off you can't switch the light on and you have to wait for about 2 seconds. Don't know the reason, but the feeling is not so good. You might think something went wrong.

    Top button switches between Lo, Mid, High modes. It correctly works without Off mode in the cycle. Preset levels are 15%, 50%, 100%. Personaly I would rather see it as 5%, 35%, 100% since eyes are logharitmic not linear "device". Brightness change from 50% to 100% is not that much as one might expect, but this is general problem with any light.


    Programing does NOT work as we anticipated. In fact there is no such thing. I was bit seduced by statement in Instruction Manual: "Double press SW2 to exit personalized brightness mode, note that the brightness level you have set will be saved." Zanflare confirmed it is wrong and they will correct the manual.

    So there is only "Personalized Brightnes Level Setting" acting as 4th special mode which is valid only until you power off the light or change to other mode. You can set it at any mode if you press top button for 2 seconds. The leds blinks 3 times so you know you have entered into personalized mode. Each click advance to higher level by 10%, so you can choose from 10 levels (similar as with KD2 light). After desired brightness level is choosen you need to double press top switch to exit that mode. Funny enough there is no sign you have exit from this mode. I wish to see some sign of confirmation. It's bit confuzing.

    There are other 3 special modes: Strobe, SOS, Beacon. To enter this mode you need to double press top button. With short press you can cycle between all three special modes. To exit that mode you need to double press top switch same as in Personalized mode.

    Buttons are little stiff and you need to press hard to switch works. On one side this is good cause it won't happen accidentialy, but anoying as you might change the light position pushing it so hard. I can't imagine how this will work using gloves in winter. I would love to see the swithes are easier to operate.



    Leds and beam profile

    Leds are Cree XP-L HD (High Density) V5. The tint is not super white but with bit of yellow hue, I would speculate it is around 1D. This is bit better than we have anticipated.

    Here is the comparation picture of leds tint. From left to right an old Yinding (about 4C tint), BT21, BT40S, Zanflare B3
    Review: Zanflare B3 bicycle light-tint-comparation.jpg

    Reflectors are OP (Orange Peeled) which disperse the light nicely. There is almost no defined hotspot. This is the result of quite small reflectors and pretty big led die (about same size as XM-L2). Diameters of front holes is only 13.8mm which means the reflectors are merely about 16-17mm. So 20mm TIR optics would not fit. Will be measured when the light is dissassembled.

    All in all I like the beam. Maybe it is more on wide side so pretty good for bar use. And it is pretty bright too. Will see how it compares with others.

    Thermal transfer
    The light gets warm even hot very quickly specialy on High modes. This indicates good heat transfer. Of course it has to be proven after internal inspection, but first impression is favorable.


    Current Measurments

    I've used fully charged standard KD pack with 4 Panasonic NCR18650B cells. Followed same technic as Garry below and got bit higher readings.
    #LEDs on Level Amps
    3 H 8A+ => shut off or switch to Low mode
    3 M 2.40A
    3 L 0.72A
    2 H 3.80A
    2 M 1.46A
    2 L 0.46A
    1 H 1.56A
    1 M 0.63A
    1 L 0.20A

    It looks like many battery packs will have problem. In reality is little bit different as there is substantial voltage drop over additional connectors and wireing. Leaving out measurment tools light still works on full. Later on we need to determine what are voltages when droped. Representative ZFL mentioned lowest allowed voltage is 6V which is bit high. I would set it to 5.6V to maintain voltage drop and sag at battery side.

    Waterproof?

    Yes, tested. I've submerged the whole light into water for more than 1 hour. It still works. Did not see any flaws. At first there were air bubbles between fins and even on the lens from outside of course.
    Review: Zanflare B3 bicycle light-light-submerged.jpg

    Using tothstick I managed to release those bubbles so I could see if any new appears. Finaly it has been turned on and runt in the water.
    Review: Zanflare B3 bicycle light-light-submerged-works.jpg

    Teardown

    In short;
    - everything is sealed (GOOD)
    - the reflectors are very small 16.8 x 7.45mm (well 20 x 13mm would be better)
    - the leds are mounted on single copper PCB (VERY GOOD). There is also some thermal paste behind
    - leds PCB sticks with another Alu backplate which is screwed to the housing (BAD).
    - there is quite small area (tiny lips and small part under the screws) for heat transfer to the housing and to surprise no thermal paste on that lips (BAD)
    - below central led there is thermistor soldered to the PCB which means temperature is controlled directly near leds (GOOD)

    Some most important pictures.
    Reflectors:
    Review: Zanflare B3 bicycle light-reflectors.jpg

    Reflectors size:
    Review: Zanflare B3 bicycle light-ref-diameter.jpg
    Review: Zanflare B3 bicycle light-ref-height.jpg

    Thermistor on PCB:
    Review: Zanflare B3 bicycle light-thermistor.jpg

    Driver consists of two PCBs:
    Review: Zanflare B3 bicycle light-driver-2.jpg

    Leds PCB:
    Review: Zanflare B3 bicycle light-led-pcb.jpg

    Alu backplate:
    Review: Zanflare B3 bicycle light-alu-backplate.jpg

    There is a lot more pictures but would overhelm this review.


    Driver

    It has three separate parts for each led and one common micro controller for modes. Leds are driven by FR9809 chip and sense resistors are R016 or 0.016 Ohm. Modes are controlled by unknow chip, but it looks pretty powerful and complex if you take into account all those pins it has. Here are some picturs from different sides.

    Review: Zanflare B3 bicycle light-driver-4-outer-leds.jpg
    Review: Zanflare B3 bicycle light-driver-back-side.jpg
    Review: Zanflare B3 bicycle light-driver-upper-switch.jpg
    Review: Zanflare B3 bicycle light-driver-cross-conection.jpg


    Status Led
    Additional measurments of voltage tresholds for status led at driver input. Keep in mind that actual voltage at the battery side would be higher than at driver side due to voltage drop. So I read the voltages that driver gets on input since the battery side could be variable ie. different battery packs and cables.

    Voltage treshold Status led
    6.85V blue to red
    6.03V steady red to red blinking
    5.65V red blinking to light completly off

    You need to disconnect the power to reset the status if the voltage for any reason gets higher. Between 6.03V and 5.65V after reconnecting red status led is blinking even if the light is turned off. It looks this is warning the battery voltage is below 6V. If the voltage is below 5.65V the light won't even turn on.

    To be continued....

    Some discussions about this light is going on in this thread too: http://forums.mtbr.com/lights-night-...s-1052103.html

    edit 20th Jul 2018: corrected tittle as final product is somewhat different
    Last edited by ledoman; 07-19-2018 at 11:15 PM.

  2. #2
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    My review may end up more of a jumbled mess of thoughts than a nice organized write-up, but oh well. I'm pretty busy these days and finding it hard to devote time to these things. Anyway, here we go.

    Initial impressions are positive. The light is small and lightweight (4.9 oz by my measurements). The fins are nicely done and well rounded so as not to be rough to handle. The buttons are a bit small and a little stiff to operate, but the stiffness feels like quality to me. I do agree with ledoman that using those buttons with gloves may be tough.

    The button operation is as ledoman described. Each LED must be driven independently because as you change between 1LED-2LEDs-3LEDs each single LED stays driven at the same level.

    Beam pattern is surprisingly good (wide and floody due to small OP reflectors) with no hotspot and also very few rings (which are only visible up close against a white wall). Even in 2LED mode (the outer 2LEDs) the beam pattern does not appear to have a hole of any kind.

    The beam pattern is floody, so it's more well suited for as a bar light. In my opinion a better design (at least for a design using reflectors) would have been to have the center reflector much larger in order to get more throw. The 2 side LEDs with smaller reflectors provide plenty of flood for up close. The light does have surprisingly good throw but that throw is produced purely from the total output of lumens. As you lower mode levels you lose throw pretty quickly.

    While testing current pulled from the battery pack in "3 LED Mode" on high on a fully charged pack, I saw a high of 6.40amps quickly dropping down to around 6.0amps! Wow! No wonder this light heats up quick. It can't maintain 6amps for long, but I haven't tested how far it drops yet (that's another future test during the temperature/thermal testing). Users will definitely need high quality battery packs with thick power leads and protection circuits rated high enough not to trip. (I imagine a lot of users will run into issues, even if using mediocre decent battery packs.)

    Current Measurments (From Battery Pack):
    #LEDs on Level Amps
    3 H +/-6.40A
    3 M 2.17A
    3 L 0.61A
    2 H 3.30A
    2 M 1.33A
    2 L 0.39A
    1 H 1.45A
    1 M 0.60A
    1 L 0.18A

    Here are my current measurements taken in "1 LED Mode" for output at each 10% increment (current measured pulling from my homemade 6-cell battery pack consisting of Panasonic/Sanyo NCR18650BE cells and a 7amp protection circuit with 20awg power lead). Keep in mind I have two extra connectors in-line during measurements which could be causing some voltage drop.

    Level Amps
    1 0.18A
    2 0.24A
    3 0.36A
    4 0.49A
    5 0.63A
    6 0.77A
    7 0.93A
    8 1.09A
    9 1.26A
    10 1.45A

    I took beamshots of each LED mode and each mode level and compared to my Solarstorm X3 (driver stock, has pills, modded to Neutral white) and my Nitefighter BT21S (stock driver and LEDs, only changed optics to LEDDNA 10º with white holders). There was a lot of moisture in the air, but they seemed to come out ok. Beamshots in my standard locations. First set across my backyard (approx. 185 feet to the two trees, approx. 275 feet to the house off to the left of the trees, utility pole in the distance is at approx. another 100 feet away from the trees), second set in the woods at my "simulated bike trail" environment:

    (BEAMSHOT PICS TO COME)

    Here is a video showing how the two buttons operate (except for "Personalized Brightness Mode"): https://youtu.be/07fd-h6r9CA



    STAY TUNED FOR MORE REVIEW

    (It's coming, probably in stages.)

    -Garry

    Last edited by garrybunk; 08-22-2017 at 01:50 PM.
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  3. #3
    mtbr member
    Reputation: Cat-man-do's Avatar
    Join Date
    May 2004
    Posts
    6,351
    This post reserved for my review....

  4. #4
    ZFL
    ZFL is offline
    mtbr member
    Reputation: ZFL's Avatar
    Join Date
    Aug 2017
    Posts
    19
    Thanks for sharing

  5. #5
    mtbr member
    Reputation: MRMOLE's Avatar
    Join Date
    May 2011
    Posts
    2,871
    Well the B3 is definitely attractive to look at. Promise of good power for a reasonable price is there too. Looking at the current draw you guys are getting I can see why the battery power cable is such a large gauge. Current draw seems inappropriate for claimed lumen output though. Roughly double the readings I get from my GW XS or Ituo XP3. Will be interesting to see how much of that power is used to produce light and how much to produce heat. 8 cell batteries might be required to get much runtime if you like to use high all the time. Definitely has my curiosity level elevated!
    Mole

  6. #6
    RAKC Industries
    Reputation:
    Join Date
    Jan 2017
    Posts
    3,265
    "pretty big led die (about same size as XM-L2)"

    Not cree LEDs and 6+ amps with ledoman seeing even higher (8+) is bringing to light the issues we saw when this light was first in the works with GJHS. Almost sounds like they went direct drive for high mode on 3 led. Or they are trying to drive the lattice bright leds to max to try and get close to claimed lumens.

    Understanding the better cooling with reflectors now. The design needs better transfer to the head since the light can so hot.

    Sounds like it could be a fun light for us light gurus but I could see serious problems for the average rider.

    I am bothered by the cost with the chosen LEDs. Think they would have had an easier time with heat issues on this light seeing that cree emitters would be more efficient.

    Anxiously awaiting teardown pics.

    Thanks for this guys, nice to finally see something new from you guys.

    Sent from my SM-G950U using Tapatalk
    Life on a bike doesn't begin till the sun goes down.


  7. #7
    mtbr member
    Reputation:
    Join Date
    Feb 2016
    Posts
    710
    Great start on the reviews fellas! Subscribed.

  8. #8
    arc
    arc is offline
    mtbr member
    Reputation:
    Join Date
    Sep 2004
    Posts
    370
    Quote Originally Posted by ledoman View Post

    Current Measurments

    I've used fully charged standard KD pack with 4 Panasonic NCR18650B cells. Followed same technic as Garry below and got bit higher readings.
    #LEDs on Level Amps
    3 H 8A+ => shut off or switch to Low mode

    2 H 3.80A

    1 H 1.56A

    If these amperages are being measured at the pack something is wrong.

    Assuming 8 volts on the pack its 12.5 watts for one lamp on high, 30.4 for two and 64 watts for 3 lamps on high, that doesn't seem right. Using the cree PCT tool with V5 bin the led produces 74 lumens per watt at it's worst rating if both the optical and driver efficiency are at 80%. Even at those lousy efficiency levels two lamps should make 2250 lumens, the three on high doesn't make any sense at all.

    Either something is throwing the amperage measurements off or the driver is massively inefficient, almost looks like an internal short circuit is happening on 3 high. Maybe Zanflare should be contacted to see if the drivers are faulty.

  9. #9
    arc
    arc is offline
    mtbr member
    Reputation:
    Join Date
    Sep 2004
    Posts
    370
    Quote Originally Posted by RAKC Ind View Post

    I am bothered by the cost with the chosen LEDs. Think they would have had an easier time with heat issues on this light seeing that cree emitters would be more efficient.
    I thought they were xpl v5's. Where did you see lattice bright?

  10. #10
    ZFL
    ZFL is offline
    mtbr member
    Reputation: ZFL's Avatar
    Join Date
    Aug 2017
    Posts
    19
    Quote Originally Posted by MRMOLE View Post
    Well the B3 is definitely attractive to look at. Promise of good power for a reasonable price is there too. Looking at the current draw you guys are getting I can see why the battery power cable is such a large gauge. Current draw seems inappropriate for claimed lumen output though. Roughly double the readings I get from my GW XS or Ituo XP3. Will be interesting to see how much of that power is used to produce light and how much to produce heat. 8 cell batteries might be required to get much runtime if you like to use high all the time. Definitely has my curiosity level elevated!
    Mole
    When the users use our battery case on Indiegogo page, they will get the ideal voltage 8.4V (4 18650 Cells: 2 series and 2 parallel), run time and brightness. The acceptable voltage is from 6 to 8.4V. There is a risk if higher than 9V. And it will not work if lower than 6V.

  11. #11
    RAKC Industries
    Reputation:
    Join Date
    Jan 2017
    Posts
    3,265
    Quote Originally Posted by arc View Post
    I thought they were xpl v5's. Where did you see lattice bright?
    Xp-ls would be same size as xm-l2 (maybe i misunderstood ledoman), beam pattern would be almost identical. XP-l is nothing more than an xm-l2 in a xpg2 foot print.

    Xp-l v5 shouldnt need near than much power to get 2000 lumens out of 3 of them. Little over half that maybe.

    Definately hoping I misunderstood because then it should have its rated output (even if it cant run it for more than a brief period)

    Nice thing is i have packs that should be able to handle the insane current draw. Far gauge wire and all.

    Sent from my SM-G950U using Tapatalk
    Life on a bike doesn't begin till the sun goes down.


  12. #12
    mtbr member
    Reputation:
    Join Date
    Jul 2013
    Posts
    751
    Great start guys! How long have we been waiting for this?

    The light was sort of expected to have heat issues. As you recall the size vs lumens was a big debate in the original thread as well as the lack of a solid core. Things that GB sort of knew, that can finally be confirmed now in our testing.

    FYI, the case was originally supposed to be copper extending to the outside where the LED plate attached, though weight and I'm sure price ended that. In the original designs you could see the copper.

    My sample is coming and I will do my usual beam shots and teardown.

  13. #13
    mtbr member
    Reputation:
    Join Date
    Jul 2013
    Posts
    751
    Reserved for my review

  14. #14
    arc
    arc is offline
    mtbr member
    Reputation:
    Join Date
    Sep 2004
    Posts
    370
    Quote Originally Posted by RAKC Ind View Post
    Xp-ls would be same size as xm-l2 (maybe i misunderstood ledoman), beam pattern would be almost identical. XP-l is nothing more than an xm-l2 in a xpg2 foot print.

    Xp-l v5 shouldnt need near than much power to get 2000 lumens out of 3 of them. Little over half that maybe.

    Definately hoping I misunderstood because then it should have its rated output (even if it cant run it for more than a brief period)

    Nice thing is i have packs that should be able to handle the insane current draw. Far gauge wire and all.

    Sent from my SM-G950U using Tapatalk
    It's quite possible I'm missing something because nothing about this light is making sense to me. With all that current and surface area to radiate heat it should be making more light than an XP3, not less. The XP3 has it beat in beam pattern and colour temperature. Proven reliability, good run time and easy to use. I don't understand why you would be interested in this light. I figured that new Imalent BG10 would have your attention, it has more features and gimmicks than your old Fenix. Probably last just as long too.

    There were things I was hoping to see in this light that I haven't found yet, hopefully the driver gets sorted out. I really want to try a helmet light with some nichia's. They'd fit on the led boards but the low vf would make the driver melt. Xpl hi's would be ok, they would fix the colour temp and beam pattern but wouldn't be much different than my XP3.

    Bike lights just don't seem to be advancing, so much me too design. Maybe the new luxeon V and companies like Imalent trying crazy stuff will shake things up.

  15. #15
    mtbr member
    Reputation: MRMOLE's Avatar
    Join Date
    May 2011
    Posts
    2,871
    Quote Originally Posted by ZFL View Post
    When the users use our battery case on Indiegogo page, they will get the ideal voltage 8.4V (4 18650 Cells: 2 series and 2 parallel), run time and brightness. The acceptable voltage is from 6 to 8.4V. There is a risk if higher than 9V. And it will not work if lower than 6V.
    My speculation of the necessity of an 8 cell battery was based on the current draw measured by the reviewers with the use of a 8.4V 8 cell battery and a 2 hr. runtimie on high (what I consider minimum acceptable). Still a lot of unknown variables about this light on my end but guessing your opinion of an ideal runtime (@max) and mine differ. Thanks for your reply.
    Mole
    Last edited by MRMOLE; 08-22-2017 at 11:36 PM.

  16. #16
    RAKC Industries
    Reputation:
    Join Date
    Jan 2017
    Posts
    3,265
    Quote Originally Posted by arc View Post
    It's quite possible I'm missing something because nothing about this light is making sense to me. With all that current and surface area to radiate heat it should be making more light than an XP3, not less. The XP3 has it beat in beam pattern and colour temperature. Proven reliability, good run time and easy to use. I don't understand why you would be interested in this light. I figured that new Imalent BG10 would have your attention, it has more features and gimmicks than your old Fenix. Probably last just as long too.

    There were things I was hoping to see in this light that I haven't found yet, hopefully the driver gets sorted out. I really want to try a helmet light with some nichia's. They'd fit on the led boards but the low vf would make the driver melt. Xpl hi's would be ok, they would fix the colour temp and beam pattern but wouldn't be much different than my XP3.

    Bike lights just don't seem to be advancing, so much me too design. Maybe the new luxeon V and companies like Imalent trying crazy stuff will shake things up.
    Arc

    You are missing a lot

    First off see my signature. Trails I only ride Ituo, pavement is Ravemen (unless a long ride then my wiz20 with wide angle optics that Mole recommended)

    However:

    Before all that began around a year ago I was just another "light guru" aka light geek around here. User name, Tigris99. Did many reviews and such. I blame garry, cat, mole, ledoman and GJHS for that lol.

    Bike lights rekindled my infatuation with LED lighting. Modified chinese lights, machined bits for them, even toy with machining my own (really need to get back to work on those). So I have everything between the cheapest 808 clone up to gloworm X2 and XS. Some for review others like gloworm, 880 clone and others I bought. All on a shelf in my collection. All ridden once at least.

    This ZF light was born here with all of us discussing and pushing for it. GJHS came the closest, which is this head design and other details. Was abandoned by gearbest/zanflare (one in the same) for reasons unknown fully. Then suddenly it reappears over on BLF apparently.

    Zanflare may have this light all weird and some things way out of wack, but its existence started here with some serious time put in by these guys. So no matter how failed the design ended up by this point due to Zanflare, the light exists. Which means I have to have it. May never get ridden on a trail (though I always ride every new addition once) but it must be part of my collection.

    Many of these guys like Mole and Cat have a large collection of lights as well. Likely well surpassing mine. We all have higher end lights (Ituo, gloworm etc). Obviously I have a full collection of Ituo and ravemen.

    As for hopes that things are not what they seem, we had high hopes and realistic expectations of how this would be, so ya I have some hopes that these guys doing their thing will prove initial impressions somewhat incorrect. They are the best at what we do, so I anxiously await all their findings to be shared.

    As for advancements, there are limitations of current technology. Remember LED and batteries are designed for bike lights. They are designed for other purposes and happen to fit for this purpose really well.

    As well as latest and greatest tech is expensive. Bike lights are not something where there is big money like other bike components. So innovation is much slower.

    Also that Imalent brand hasnt done anything special. Bogus lumen rating, little battery and so on. Just more added things to try and act innovative. All of which has been done already. The emitter choice isnt widely adopted due to cost and heat issues.

    That Luxeon V has promise. Higher drive current rating, but seems per watt the output is cruising pretty close to cree. Just designed to be pushed at higher currents. Cree usually has an answer to this stuff though, they did to the nichia 219c (released xp-g3)

    A big break through would be batteries which is coming.

    Sent from my SM-G950U using Tapatalk
    Life on a bike doesn't begin till the sun goes down.


  17. #17
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Quote Originally Posted by RAKC Ind View Post
    Anxiously awaiting teardown pics.
    Planning to do it today or tomorrow eventualy.

  18. #18
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Quote Originally Posted by arc View Post
    If these amperages are being measured at the pack something is wrong.

    Assuming 8 volts on the pack its 12.5 watts for one lamp on high, 30.4 for two and 64 watts for 3 lamps on high, that doesn't seem right. Using the cree PCT tool with V5 bin the led produces 74 lumens per watt at it's worst rating if both the optical and driver efficiency are at 80%. Even at those lousy efficiency levels two lamps should make 2250 lumens, the three on high doesn't make any sense at all.

    Either something is throwing the amperage measurements off or the driver is massively inefficient, almost looks like an internal short circuit is happening on 3 high. Maybe Zanflare should be contacted to see if the drivers are faulty.
    arc, there are two things you need to take into account. First are shotcomings of measurment equipment - it makes additional voltage drop. And secondly there is voltage drop at battery side too + wires and connector. More cells in parallel less voltage drop. At lower currents this is not important but it escalates more and more (expotentialy) at higher currents. Unfortunately we don't have proffesional equipmet.

    Now if the driver is regulated (we don't know yet) the lower is input voltage the higher is current in order to maintain constant power. Since voltage drop over the measurment tools is significant the current gets higher. There is some current drop, but it might be due temperature change. Direct drive is also possible, but I doubt it. At least it shouldn't happen.

    I still write my things under "First impressions" so things can change. I plan to tear the light down and I might try to measure things more accurately and closer to the light or even to the leds.

  19. #19
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Well I got sphere measurements last night (used a Fenix HL55 alongside to check/refine my calibration), HOWEVER I'm unable to put them in my spreadsheet and post #'s as I will be heading out of the office today! DOH! (Why is it that work always gets in the way when you have personal business to attend to???) I'll get them posted up as quickly as I can - possibly later this afternoon. For now I will say WOW! I think 3LEDs on high greatly surpasses what I saw from my heavily modded BT40S pushing 2000 lumens! BUT, #'s drop very quickly (no surprise there)! Lower levels are very steady though.

    I'm waiting until all other tests/measurements are done before tearing open mine. I don't want an "Uh-Oh" moment and realize I can no longer get measurements and pics I should have got while it was working! I do hope to pass CAT a working sample .

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  20. #20
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    With permission of old4570 memeber at BLF I can post in advance his current readings between battery pack and light. He will do review of B3 at BLF. You can read his teaser (whatever it means) at this link: Zanflare B3 Teaser ( Bike Light ) | BudgetLightForum.com

    So his readings are:

    1 emitter ( Low – 0.4 Amp ) ( Medium – 0.7 Amp ) ( High – 1.56 Amp )
    2 emitters ( Low – 0.56 Amp ) ( Medium – 1.54 Amp ) ( High – 3.16 Amp )
    3 emitters ( Low – 0.84 Amp ) ( Medium – 2.14 Amp ) ( High – 5.25 Amp )

    Batteries were 2s 2200mAh RC Li-Po batteries with XT60 connectors, used two in parallel. So much better than me and Garry used.

    Still readings are quite similar if you take into account differencies in batteries, wireing, connectors and mesurment tools. Current on 3leds/High is still quite big, I would speculate about 35W since I don't know the voltages at a time. And yes light gets hot !!! Will test thermal protection too, but very likely more at the end since there is chance something goes wrong.

  21. #21
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    So do we suspect his current reading 3LEDs on High are lower because he has way less voltage sag?

    BTW - I tried two high drain cells (@4.14v) in my modded 2cell Fenic battery case and the light only stayed in High (3LED mode) for about 10 seconds before tripping what appeared to be low voltage protection. It dropped to a low mode (might have flashed too) and I couldn't change modes. So this is not a light to run on a helmet with a 2 cell battery case (unless you have a really great battery setup).

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  22. #22
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    If the driver is regulated (ie. constant power) then this should be the case, yes. Just on low mode he got higher rreadings I can't explain.

    I've experienced some similar step down when tested high currents. The voltage limit is supposedly at 6V which is to high taking into account voltage drop over wires and connector + battery voltage sag. Modern (Panny) cells can go down to 2.5V, 2.8V would be the best to use, so 5.6V limit per pack. The worst bottleneck now are becomming connectors. They are not designed for high currents and they might causing high voltage drop regardles we are using 20+AWG wires.

  23. #23
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Quote Originally Posted by RAKC Ind View Post
    Xp-ls would be same size as xm-l2 (maybe i misunderstood ledoman), beam pattern would be almost identical. XP-l is nothing more than an xm-l2 in a xpg2 foot print.

    Xp-l v5 shouldnt need near than much power to get 2000 lumens out of 3 of them. Little over half that maybe.

    Definately hoping I misunderstood because then it should have its rated output (even if it cant run it for more than a brief period)

    Nice thing is i have packs that should be able to handle the insane current draw. Far gauge wire and all.
    Right, XP-L (phosphor size) is same size as with XM-L2 just XP-G (3.5x3.5mm) foot size. This footprint size is shared among many leds now.

    Speculating the driver is regulated explains high current readings with our equipment. Big voltage drop over that additional connectors, measurment device and wires would rise the current in order to maintain constant power. In reality currents would be smaller. Still voltage sag from battery to light would drive input current. I've experienced such thing with driver I've found at DX. It had same 10 level programmning user interface as KD2 and pretty much same we wated to be implemented here with B3.

  24. #24
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Tested if it is waterproof. YES still works after one hour in the water. See the pictures in the first post.

  25. #25
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Just finished with tear down (dissassembling). Didn't desolder anything. If you would like to take it apart completly you would need to desolder input cable.

    In short;
    - everything is sealed (GOOD)
    - the reflectors are very small 16.8 x 7.45mm (well 20 x 13mm would be better)
    - the leds are mounted on single copper PCB (VERY GOOD). There is also some thermal paste behind
    - leds PCB sticks with another Alu backplate which is screwed to the housing (BAD).
    - there is quite small area (tiny lips and small part under the screws) for heat transfer to the housing and to surprise no thermal paste on that lips (BAD)
    - below central led there is thermistor soldered to the PCB which means temperature is controlled directly near leds (GOOD)

    Will try to add few pictures in the first post, soon.

  26. #26
    mtbr member
    Reputation:
    Join Date
    Jul 2013
    Posts
    751
    Great Job Ledoman

    I can't believe this light is finally here for testing

  27. #27
    arc
    arc is offline
    mtbr member
    Reputation:
    Join Date
    Sep 2004
    Posts
    370
    Quote Originally Posted by ledoman View Post
    If the driver is regulated (ie. constant power) then this should be the case, yes. Just on low mode he got higher rreadings I can't explain.

    I've experienced some similar step down when tested high currents. The voltage limit is supposedly at 6V which is to high taking into account voltage drop over wires and connector + battery voltage sag. Modern (Panny) cells can go down to 2.5V, 2.8V would be the best to use, so 5.6V limit per pack. The worst bottleneck now are becomming connectors. They are not designed for high currents and they might causing high voltage drop regardles we are using 20+AWG wires.
    I just don't understand how this can happen if the driver is functioning properly. My Wiz 20 with only two cells makes 1500 lumens for over 100 minutes. My convoy S6 triple makes more light than the Wiz on one cell and only steps down from heat. My 850 lumen Cygolite runs over 30 minutes on high on one cell. RAKC has ran his XP3 off a two cell pack strapped to his helmet. My Haikelite MT07 does 5000 lumens on four cells until thermal step down. The BT70 ran fine on a four cell battery. The XP4 does 3200 lumens on 4 cells. Seems like 800 lumens a cell should be easy.

    I think Zanflare should be contacted to see if the light is functioning properly, it seems like there is something wrong with the lights Garry and you are reviewing.

  28. #28
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Here is the answer when constant power is used: http://lygte-info.dk/pic/Batteries20...erLoadTime.png
    You can see how current rise while voltage drops. And this chart is without any wires and connectors, so in our case voltage drop is even higher so is the current.
    You need to study different types of drivers (linear, boost, buck and regulation if there is implemented)

    Will have to check with ZLF to confirm the driver is fully regulated.

  29. #29
    RAKC Industries
    Reputation:
    Join Date
    Jan 2017
    Posts
    3,265
    Not functioning wrong, i has multiple drivers. Very inefficient. More or less like running 3 single emitter lights off the same battery pack. Massive current draw vs lumens produced compared to a single driver powering multiple leds all at once. much more effcienct both in power requirements and heat generation. Though they have copied Ituo's design of the thermistor on the LED pcb.

    Sent from my SM-G950U using Tapatalk
    Life on a bike doesn't begin till the sun goes down.


  30. #30
    mtbr member
    Reputation:
    Join Date
    May 2005
    Posts
    3,837
    Quote Originally Posted by RAKC Ind View Post
    ......Though they have copied Ituo's design of the thermistor on the LED pcb.
    Thermal sensing of the LED mounting surface is not an Ituo "design". At best they are just implementing commonly available electronic technology. I've done LED mount surface thermal sensing on most of my lights starting ~8 years ago. Other innovators in the DIY world have done that and more including active cooling.
    Last edited by Vancbiker; 08-23-2017 at 09:31 PM. Reason: added content
    GoPro adapters for bike lights http://www.pacifier.com/~kevinb/index.html

  31. #31
    RAKC Industries
    Reputation:
    Join Date
    Jan 2017
    Posts
    3,265
    I know guys like you have, never seen it in production lights before Ituo. Too open ended of a comment on my part.

    Sent from my SM-G950U using Tapatalk
    Life on a bike doesn't begin till the sun goes down.


  32. #32
    mtbr member
    Reputation:
    Join Date
    Nov 2012
    Posts
    1,043
    Quote Originally Posted by Vancbiker View Post
    Thermal sensing of the LED mounting surface is not an Ituo "design". At best they are just implementing commonly available electronic technology. I've done LED mount surface thermal sensing on most of my lights starting ~8 years ago. Other innovators in the DIY world have done that and more including active cooling.
    How is that done? Is it just a thermocouple with two wires running to the driver board? I believe that there need to some sort of amplifying circuit to get a usable signal out of the thermocouple.

  33. #33
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    No it is thermistor - different resistance at different temperatures. Explanation on Wikipedia

  34. #34
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Checked with ZLF. The driver is regulated:

    "In a same mode, when the temperature is NOT more than 60 degrees Celsius, it will maintain constant power. If more than 60, it will lower the brightness (and power) slowly. In this process, if temperature is more than 80, it will turn to mid mode immediately to protect circuits."

    Somewhat confuzing. The manual also states:

    "Note: Continued use of the product might cause overheating. When the temperature is high, the bike light will enable overheat protection function and automatically enter the low brightness level, if the user tries to enable higher brightness level the bike light will anyway turn to lower brightness mode because of overheating protection function."

  35. #35
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Well I lost all the details I just typed about thermal management - UGH! Suffice to say I took video and it's coming.

    I have bad news though. At work I've been placed on a project requiring me to be put in the field days at a time. This means I'll have much less time to post to my review and keep up with this thread. These will be long days which also means I'll have less time for tests/measurements at home in the evenings. I'll do my best and will still try to get the light to Cat in a reasonable time.

    -Garry

  36. #36
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    If you have taken, just post a link to your raw beamshots, so we can get some impression. Mine sample is per partes now. My next step is to examine the driver and see what I can do with it. It won't be assembled back for quite some time to do any beamshots. Hopefuly it will survive my "surgery".

  37. #37
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Just checked and I haven't uploaded my beamshots yet (swear I did). I did upload the "index" pic showing which photo is which mode & light. I'll try to get them uploaded tonight and post when I've done so. Here's the album link: Zanflare B3 Bike Light by Garry Bunk | Photobucket

    I'll also post my approximate lumens as soon as possible.

    -Garry

  38. #38
    mtbr member
    Reputation:
    Join Date
    May 2005
    Posts
    3,837
    Quote Originally Posted by varider View Post
    How is that done? Is it just a thermocouple with two wires running to the driver board? I believe that there need to some sort of amplifying circuit to get a usable signal out of the thermocouple.
    An external thermister provides an analog input (voltage change relative to heat level) to the driver MCU (microcontroller). Some MCUs have an integral thermister that can do the same thing as long as there is a heat path to the LED mounting surface provided.
    Last edited by Vancbiker; 08-24-2017 at 07:42 AM. Reason: fix spelling
    GoPro adapters for bike lights http://www.pacifier.com/~kevinb/index.html

  39. #39
    mtbr member
    Reputation:
    Join Date
    Oct 2007
    Posts
    50
    Quote Originally Posted by garrybunk View Post
    Just checked and I haven't uploaded my beamshots yet (swear I did). I did upload the "index" pic showing which photo is which mode & light. I'll try to get them uploaded tonight and post when I've done so. Here's the album link: Zanflare B3 Bike Light by Garry Bunk | Photobucket

    I'll also post my approximate lumens as soon as possible.

    -Garry
    Photobucket has gone crazy:

    https://www.theregister.co.uk/2017/0..._3rdparty_pgs/

    I recommend to use imgur instead

    Imgur: The most awesome images on the Internet

  40. #40
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Quote Originally Posted by juhake View Post
    I'm on a paid level of service, but not $400/yr. Guess they're waiting until my renewal comes up in October. Yeah I should move to another service, but I do have a lot on PB.

    -Garry

  41. #41
    RAKC Industries
    Reputation:
    Join Date
    Jan 2017
    Posts
    3,265
    Ya get rid of photobucket asap. Doubt theyll be around in october.

    Sent from my SM-G950U using Tapatalk
    Life on a bike doesn't begin till the sun goes down.


  42. #42
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Ok, here is the link for the Thermal Analysis Video (title might be a little misleading): https://youtu.be/ZEcGCzAG7FM
    I'll need to add the link up in the review post at some point, but can't reliably edit that post right now.

    I still have to upload beamshots too, hopefully later.

    -Garry

  43. #43
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Beamshots are in my PhotoBucket album here. There is a "pic" of a screen capture with notes listing which pic is which light in which mode. I compared the SS X3 and BT21S alongside the B3. NOT ALL ARE POSTED BELOW. All shot with standard MTBR settings.

    Here's the B3 in 3 LED mode High:


    2 LED High:


    1 LED High:


    SSX3 High:


    Nitefighter BT21S Turbo:


    B3 3 LEDs High:


    B3 2 LEDs High:


    B3 1 LED High:


    SSX3 High:


    BT21S Turbo:


    Hope that satisfies the craving for beamshots for now.

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  44. #44
    arc
    arc is offline
    mtbr member
    Reputation:
    Join Date
    Sep 2004
    Posts
    370
    Quote Originally Posted by RAKC Ind View Post
    Not functioning wrong, i has multiple drivers. Very inefficient. More or less like running 3 single emitter lights off the same battery pack. Massive current draw vs lumens produced compared to a single driver powering multiple leds all at once. much more effcienct both in power requirements and heat generation. Though they have copied Ituo's design of the thermistor on the LED pcb.

    Sent from my SM-G950U using Tapatalk
    It looks like three individual buck drivers, one per led with a big inductor (compared to the load). Functioning properly efficiency should be excellent. Rip the domes off the led's and get the current under control might turn this into a decent light. Even switching to xpg2's, xpl hi's or 219c's would help the small reflectors.

    It's weird the current doesn't increase linearly, it's three identical channels in parallel. They shouldn't be able to saturate those inductors. Maybe something is not turning off properly when the load or switching frequency goes up and the buck drivers are turning into linear drivers. Hopefully someone posts a picture of the driver that shows the numbers on the switch line controller chips and the mcu.

  45. #45
    mtbr member
    Reputation:
    Join Date
    Nov 2012
    Posts
    1,043
    Thanks for the explanation ledoman and Vancbiker.

    Garry, how far is it to that treeline in the B3 (3 led on high) picture. It looks like it throws pretty far. It seems like it has a lot of near-field illumination with lots of spread to the side, maybe due to the short reflectors. The tint seems on the ugly side. Way too cool for me. That ssx3 looks better tint-wise.

    Also, thanks for going to the trouble of getting some beamshots guys.

  46. #46
    RAKC Industries
    Reputation:
    Join Date
    Jan 2017
    Posts
    3,265
    I love garrys "test trail". Good for giving a more real idea of what a rider can expect.

    Keep up the good work guys, enjoying seeing you guys at work again! Missed you guys being around all the time posting this stuff for us geeks.

    Sent from my SM-G950U using Tapatalk
    Life on a bike doesn't begin till the sun goes down.


  47. #47
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Garry thanks for the beamshoots.

    Yesterday I've taken some pictures of the driver. They are using completly different regulators than we are used to see. It is marked FR9809 which has 805mV typical feedback voltage. Sense resistors are R016 so very low 0.016 Ohm. Based on that I've tried to calculate the current to the led but it doesn't make sense to me 0.016 / 0.805 = 0.0199. Speculating this is 1% then current is 2A. Will need to ask Archie if he can help.
    EDIT: "FR9809 is a high-efficiency synchronous step-down DC/DC converter"..."continuous load current capability is 5A"

    Buck part of the circuit for one led:
    Review: Zanflare B3 bicycle light-driver-1led-control.jpg

    Modes micro controller is pretty much hidden between two PCBs and can't read markings. And I won't take driver apart to get that info.

    This is the best I could get so far:
    Review: Zanflare B3 bicycle light-driver-mcu.jpg

    Will add some pictures in the 1st post too.
    Last edited by ledoman; 08-25-2017 at 04:35 AM.

  48. #48
    ZFL
    ZFL is offline
    mtbr member
    Reputation: ZFL's Avatar
    Join Date
    Aug 2017
    Posts
    19
    Quote Originally Posted by varider View Post
    Thanks for the explanation ledoman and Vancbiker.

    Garry, how far is it to that treeline in the B3 (3 led on high) picture. It looks like it throws pretty far. It seems like it has a lot of near-field illumination with lots of spread to the side, maybe due to the short reflectors. The tint seems on the ugly side. Way too cool for me. That ssx3 looks better tint-wise.

    Also, thanks for going to the trouble of getting some beamshots guys.
    We are also selling the NW version on Indiegogo.

  49. #49
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850

  50. #50
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Quote Originally Posted by varider View Post
    Thanks for the explanation ledoman and Vancbiker.

    Garry, how far is it to that treeline in the B3 (3 led on high) picture. It looks like it throws pretty far. It seems like it has a lot of near-field illumination with lots of spread to the side, maybe due to the short reflectors. The tint seems on the ugly side. Way too cool for me. That ssx3 looks better tint-wise.

    Also, thanks for going to the trouble of getting some beamshots guys.
    The two trees to the front are approx. 185'. The utility pole to the right would be about 285'. There's also another treeline across the road beyond the pole but unsure on that distance - at least 325' I'd say. I should probably take some daylight pics so people know what they're looking at.

    I think my white balance setting made everything look slightly cooler than reality. I checked other settings but that was the
    closest I could get (I've not been successful with custom white balance settings). There was a lot of moisture in the air too. That SSX3 is stock driverwise (1.6A per LED IIRC), but LEDs are 5B1 tint which is the warm side of neutral white and appear more like a 4C tint in the pics. My B3 in reality looks very crisp cool white (no blue, yellow, or green that I notice), not blueish like in the pics.

    -Garry

  51. #51
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    About tint, my assumption is about 1D, 1C. I could add two even more cooler tints to my picture above. I don't find B3 tint bad comparing to most other CW lights out there. Of course real NW version would be much better, no doubt.

  52. #52
    mtbr member
    Reputation:
    Join Date
    Nov 2012
    Posts
    1,043
    Quote Originally Posted by ZFL View Post
    We are also selling the NW version on Indiegogo.
    Where? I don't see anything on indiegogo that actually shows which color emitter you are buying. I find the whole page to be confusing.

    Why can't you just sell the neutral white version on Gearbest with a price for $40 for pre-orders.

    Quote Originally Posted by garrybunk View Post
    The two trees to the front are approx. 185'. The utility pole to the right would be about 285'. There's also another treeline across the road beyond the pole but unsure on that distance - at least 325' I'd say. I should probably take some daylight pics so people know what they're looking at.
    ...
    -Garry
    That seems like plenty of throw. What's your take?

  53. #53
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    I've made some measurments at led side. Desoldered one wire on the middle led and put DMM inline. All mesurments taken at room temperature and cool light. There is some sag when PCB gets warm. No wonder the light gets hot. Here is why:

    Low 0.48A @ 2,83V = 1.36W
    Mid 1.45A @ 3.11V = 4.51W
    High 3.02A @ 3.41V = 10.3W

    So all 3 leds together on High runs at 30W. If you ask me this is bit to much for such small light.

  54. #54
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    I think the throw is a result of the mass of lumens being sent out the front, not so much because of reflector design. I think the throw diminishes pretty quickly as you lower the modes/emitters. I don't think most users will power high mode well enough to keep it in use to get that throw. It's definitely better suited for flood.

    -Garry

  55. #55
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Didn't we (MTBR) ask for 3A per emitter in the custom build? What's the input wattage? Something like 7.5v x 6.4A for 48 watts? Isn't that awfully inefficient?

    -Garry

  56. #56
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    We might wanted 3A per led (though maybe not all of us), but probably for bigger light as the 20mm TIR lenses have been discussed.

    Garry you are forgetting the current measurments we did are somewhat off because our test setup is not perfect and results in noticeable voltage drop. Since the driver is regulated it sucks higher current than normal (no measurment tools inline, connectors, wires). Correct answer you would get by measuring ciurrent and voltage at driver input.

    Got idea how to do it cause don't want to desolder input wire. Will need to put another source of power in parallel. I might try it in next days since I'll be absent from home for two days.

  57. #57
    RAKC Industries
    Reputation:
    Join Date
    Jan 2017
    Posts
    3,265
    I dont remember it really being requested. It was asked about and pointed out that it was probably much to push out of such a small light head.

    And omg yes, that efficiency is horrid, 48W in for 30W out. Talking what, 60% efficiency maybe. The bad cheap lights are 80ish. Good lights are pushing 92-95.

    Definitely explains the heat problems we were seeing in the prototype and has been hinted at in the discussions so far.

    Sent from my SM-G950U using Tapatalk
    Life on a bike doesn't begin till the sun goes down.


  58. #58
    mtbr member
    Reputation: MRMOLE's Avatar
    Join Date
    May 2011
    Posts
    2,871
    [QUOTE]
    Quote Originally Posted by varider View Post
    Quote Originally Posted by ZFL View Post

    We are also selling the NW version on Indiegogo.
    Where? I don't see anything on indiegogo that actually shows which color emitter you are buying. I find the whole page to be confusing.

    Why can't you just sell the neutral white version on Gearbest with a price for $40 for pre-orders.
    If you scroll down the indiegogo site till you get to the spinning 3D image of the B3 there is a chart directly below. At the top/right side of the chart is the emitter information and an explanation of how they intend to handle the neutral white lamps. The reason you cannot select neutral white is the only thing you are doing on this site is donation to the crowd funding project. Your not actually ordering a light so no guarantee you'll get anything (if I understand the disclaimers correctly). Personally will not order till I can select neutral white from Gearbest.




    Quote Originally Posted by garrybunk View Post

    The two trees to the front are approx. 185'. The utility pole to the right would be about 285'. There's also another treeline across the road beyond the pole but unsure on that distance - at least 325' I'd say. I should probably take some daylight pics so people know what they're looking at.
    ...
    -Garry
    That seems like plenty of throw. What's your take?
    Luckily Garry also included BT21 beam pictures along with the ones of the B3. I know this light very well and judging from what I see the B3's throw is better than I expected. While the two lights show similar throw the hot spot of the BT21 is noticeably more intense so probably would out throw the B3 if aimed higher (my guess). Not a bad thing since Garry's BT21 + Leddna 10's is one of the better throwing 2 emitter lights available (not anymore) but also indicative that the B3 will be outclassed by other triples (X2/XP3) for throw. The fact you won't be able to run the B3 very long on high will only worsen this situation.


    Mole

  59. #59
    mtbr member
    Reputation: Outbound's Avatar
    Join Date
    Aug 2017
    Posts
    442
    Quote Originally Posted by garrybunk View Post
    I think the throw is a result of the mass of lumens being sent out the front, not so much because of reflector design. I think the throw diminishes pretty quickly as you lower the modes/emitters. I don't think most users will power high mode well enough to keep it in use to get that throw. It's definitely better suited for flood.

    -Garry
    The main hotspot is definitely due to the reflector design, the reason the throw diminishes quickly is probably because they just did a linear step down in "Dimming" rather than a logarithmic which will result in a perceived brightness difference. 50% power to the light won't result in 50% perceived brightness. Stick a lux meter out at a set distance and should be able to tell how bright the step downs are.

    If you guys wanna do some interesting tests. Put a contact thermometer near the LED after it's been running for about 10-15 minutes, enough time to heat soak. Then can figure out the approximate junction temperature, and compare that to the datasheet of the LED to figure out approximately what lumens it's running at in the "hot" state. A small board like that with little thermal contact as been alluded to in the previous posts is indeed going to get rippin hot. Not only will you lose 20+% of the lumens, it'll also start to shift to blue.

  60. #60
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Yes they did not use logaritmic steps for modes as I've pointed out in my first post above. The cameras does not work same as eyes which can acomodate amount of light.

    I agree it would be interesting to measure temperature at led side, but it's very hard to do and since there is thermistor and active thermal management it will fall back to lower mode in minutes. At least it should -> see my post #34. Then again Cree make their charts for Tj 85 °C which we should not reach if thermal managemet is working right. Anyway to risky to do. If I would have spare parts (driver, leds,) maybe, but we already know the heat is no good.

  61. #61
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Quote Originally Posted by RAKC Ind View Post
    And omg yes, that efficiency is horrid, 48W in for 30W out. Talking what, 60% efficiency maybe. The bad cheap lights are 80ish. Good lights are pushing 92-95.
    This of course it's not the case. As explained in post above yours, our measurments are somewhat misleading due to our equipment and regulated driver (we had similar discussions before). I repeated this several times. The best results with good connectors (though the stock one is still bad for hi currents) and 2 Li-po batteries managed to get Old at his BLF review at 5.25A though we don't know the voltage.

    Will try to do some additional current/voltage measurments at driver input DC+, DC-
    Name:  Driver DC in.JPG
Views: 2336
Size:  78.7 KB
    This should tell more about driver efficiency and without looses towards power source. There is already one info we can use - FR9809 datasheet already states up to 90% efficiency. On page 6 you can see efficiency curve at 3.3V output. This of course it is not only part of driver which consumes power.

  62. #62
    RAKC Industries
    Reputation:
    Join Date
    Jan 2017
    Posts
    3,265
    Matter being though watts are watts. Math still being the same.

    Even if its 5.25A (which would mean a pretty nicely sustained voltage from the packs and wiring/connections, high discharge lipos with deans connectors could hold 5.25A over 8V no problem for a short time, LIPO battery voltages operate much the same voltage window as LI-Ion) we're still well north of 40W input. And its possible he is running those 2 packs in series to get the voltage (ill check the thread, he has pics up I should be able to get some more details if the pics are high res enough. Spent a lot of years dealing in Lipos for RC helicopters)

    Zanflares info does mention multiple driver operation. 2 drivers at 90% efficiency were at ~80%.

    The 60% was a really rough guess but based on what OLD has, were still in the 70s at best. Really hard on standard packs. And the battery case they provide is really going to have problems.

    They have some good ideas behind this light but I think its being pushed way too hard for whats available in batteries and for the light head size.

    To reach around 2500 lumens in a triple with that kind of surface area for cooling should be gravy. Thats more direct flow surface area than lights pushing closer to 3500 lumens have.

    Not to mention we've have seen lights pushing closer to 7A that have faired better (bt70) and lights that deliver way more lumens with less current draw (Ituo XP4) and lights delivering similar to claimed lumens at half the current draw (most tripples).

    So there has to be something thats generating the massive heat this light is dealing with.

    This of course is speculation based on what we know from prototype testing and collective of what im reading here.

    This would be a fun light for guys like us to play with that have the packs to feed it. I gave 16awg dual pin cables or deans connectors to really allow feeding this thing. Since waterproofing the battery connector isnt needed and truly pointless, a deans connector would be the ultimate in dealing with voltage sag issues.

    Sent from my SM-G950U using Tapatalk
    Life on a bike doesn't begin till the sun goes down.


  63. #63
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Ok, I finished my lumen calcs. Remember, I make no guarantee to my lumens values, but I believe they are generally in the ballpark and within reasonable accuracy. I did calibrate to a different Fenix HL55 headlamp from the one I originally calibrated to AND I also made a slight change in my sphere for these measurements, so this could account for why my other "calibration check lights" being slightly different. The "expected value" column on my check lights is based on prior measurements. My "sphere multiplier" always stays within a very tight range of values (shows how consistent it is). I also recheck a couple lights at the end of my testing to verify I didn't knock the luxmeter from the same identical position etc.

    Here's the full spreadsheet with all my calcs (clickable):


    As you can see, the B3 at 3 LEDs on High mode is starting out at almost 2,600 lumens and drops to about 2300 after 30 seconds. The 2,200 claim isn't unreasonable at all! I will say that 3 XP-L's at about 3A each should be producing way more light though I would think.

    I haven't had time to compare the other claimed output levels to what I measured - feel free to do so. I got my #'s, I'm posting them up, and now I'm rushing out to do other things.

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  64. #64
    RAKC Industries
    Reputation:
    Join Date
    Jan 2017
    Posts
    3,265
    10-15% higher at cold start is normal, nice to see rated output is in line.

    Your right, xp-l v5 should be pushing 3K+ out the front at those currents. Are the losses at the reflector/lens really that high????

    Didnt happen to catch if you guys verified XP-L emitters but assuming you did.

    Sent from my SM-G950U using Tapatalk
    Life on a bike doesn't begin till the sun goes down.


  65. #65
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Yes I agree the drop is within reason of expectations on a light this small. It might not drop so badly if they mounted up the emitter board to a solid wall.

    I didn't verify real Cree emitters, though I'd expect they are. Without taking macro pics I will say they look like pics of real XP-L's. Perhaps if I disassemble I can get some good close-ups, or maybe ledoman can.

    -Garry

  66. #66
    mtbr member
    Reputation: znomit's Avatar
    Join Date
    Dec 2007
    Posts
    1,981
    Quote Originally Posted by garrybunk View Post
    I didn't verify real Cree emitters, though I'd expect they are.
    I'd say start with the expectation of latticebright for any of the cheaper lights like this.

  67. #67
    mtbr member
    Reputation: MRMOLE's Avatar
    Join Date
    May 2011
    Posts
    2,871
    [QUOTE]
    Quote Originally Posted by garrybunk View Post
    Yes I agree the drop is within reason of expectations on a light this small. It might not drop so badly if they mounted up the emitter board to a solid wall.
    Garry
    I could live with that 15% drop it's the additional 35% drop that's going to happen soon after when the light goes into thermal protection mode that rubs me the wrong way. From watching your thermal properties video and reading OLD's review on the BLF it looks like high mode will be unusable unless it's down near 50F degrees which eliminates all but about 2 months of the year for me living in Arizona. Hopefully when Ledoman reassembles his B3 a skilled application of thermal paste will help some over the factory assembled units you and OLD have been testing. Glad these are prototypes but considering how long they've been working on them I would think there would be less problems.
    Mole

  68. #68
    RAKC Industries
    Reputation:
    Join Date
    Jan 2017
    Posts
    3,265
    I found a pic of ledomans from the OP, they look like XP-L possibly. Hard to say as his camera may be showing distortion from the dome that my eyes are seeing. Otherwise they are knockoffs. Good ones.

    Im hard pressed to say they are not XP-Ls at this point. Im going to say unless one of these guys says no they are not that based on pics compared to my stache of XP-Ls that they ARE.

    Im with mole, this light has been in the works for over a year.

    Yes paste and a better thermal path would help, but problem remains as to why this thing gets so hot and why are the lumens so dang low for the wattage it uses????? I can over drive 2 xp-l HI's at 3.5A in a bt21 light head cranking almost 2000 lumens (HI's, not standard XP-L which is well over 2k lumens) and have no heat issues. It gets hot if it sits long enough but thats about it.

    Hell I run a yinding at just under 3A per emitter with XP-Ls. That one doesnt like 80+deg nights for long on high without my big vanc finned mount.

    So what are we missing here? 3 of the most efficient emitters available for this application, insane fined head exposed to full airflow. Ya path is lacking and can be improved but doesnt explain the initial heat generation. Ledoman pointed out head heats up fast, so thermal path isnt total crap.

    Chomping at the bit here to get this in front of my equipment to help figure out what is going on lol.

    Sent from my SM-G950U using Tapatalk
    Life on a bike doesn't begin till the sun goes down.


  69. #69
    arc
    arc is offline
    mtbr member
    Reputation:
    Join Date
    Sep 2004
    Posts
    370
    Sometimes I think they are still using the same driver that made last year's prototype overheat, and put a bigger radiator on it to mask the problem. I really hope this is just a bad batch of drivers.

    I took a quick look at the spec sheet for the buck controller, the switching frequency isn't the fastest but it's not horrible either. They are using the largest inductors the specifications allow and are using roughly the same number of components as the reference design so it's not like they are cutting corners. I think the three buck drivers off the same power source is creating too much electrical noise and the driver is becoming unstable. You can see it in the amperage and lumen measurements. If you divide the 3 hi measurement by three it is more than the single high measurement.

    I have a very basic understanding of buck drivers, a guy like Del on BLF could figure it out pretty fast. If there was an easy solution the Zanflare guys would have already done it. The solution would likely be resistor networks or capacitors to cut down the noise on the mcu and each buck controller chip. No more warranty and there is the risk of shorting out and destroying the driver. Most people here probably don't want to put that much effort into a light with tiny reflectors.

  70. #70
    RAKC Industries
    Reputation:
    Join Date
    Jan 2017
    Posts
    3,265
    good solution would have been to follow what all the guys here were saying. The biggest issue is the multiple drivers to run each emitter separately. Running them on either a buck or boost driver as one unit would solve the issue. Likely why you dont see much of this on any other lights. Too many issues.

    Am curious, you have all the specs for the drivers and such apparently based on your post, why not share and maybe shed some light on the situation.
    Life on a bike doesn't begin till the sun goes down.


  71. #71
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    [QUOTE=MRMOLE;13309281]

    I could live with that 15% drop it's the additional 35% drop that's going to happen soon after when the light goes into thermal protection mode that rubs me the wrong way. From watching your thermal properties video and reading OLD's review on the BLF it looks like high mode will be unusable unless it's down near 50F degrees which eliminates all but about 2 months of the year for me living in Arizona. Hopefully when Ledoman reassembles his B3 a skilled application of thermal paste will help some over the factory assembled units you and OLD have been testing. Glad these are prototypes but considering how long they've been working on them I would think there would be less problems.
    Mole
    I'm surprised my small fan isn't enough to keep it running in high mode, other lights cool down and stay at reasonable levels with that same fan. Even though I do believe something is not right with the driver circuitry, I think the emitter board being mounted to a solid wall instead of a plate would help immensely! I wonder if that mounting area is getting heat saturated and not transferring the heat away to the body fast enough. How many times did we tell them "use a solid wall of the body behind the emitters"?

    By the way, I don't believe my light feels as hot as would be expected. It gets hot, but not like it's nearly burning your hand (like a good hard driven flashlight would). Though my meter is showing the outside case temperature it's reaching.

    -Garry

  72. #72
    mtbr member
    Reputation: MRMOLE's Avatar
    Join Date
    May 2011
    Posts
    2,871
    [QUOTE][QUOTE=garrybunk;13309734]
    Quote Originally Posted by ;13309281
    I'm surprised my small fan isn't enough to keep it running in high mode, other lights cool down and stay at reasonable levels with that same fan. Even though I do believe something is not right with the driver circuitry, I think the emitter board being mounted to a solid wall instead of a plate would help immensely! I wonder if that mounting area is getting heat saturated and not transferring the heat away to the body fast enough. How many times did we tell them "use a solid wall of the body behind the emitters"?
    Interior thermal path design definitely could be improved upon but probably would have been adequate if the B3 had more efficient usage of power and only required a more typical 3a to make its 2300 lumens. Still curious if Ledoman has better luck with heat sensitivity after he reassembles his light (at least we'll know thermal paste was correctly applied). Have you tried turning the light off after it activates the thermal protection and measuring to see if the case temp. continues to increase? This is usually a good indication heat is getting trapped inside the light because of an inadequate thermal path.


    By the way, I don't believe my light feels as hot as would be expected. It gets hot, but not like it's nearly burning your hand (like a good hard driven flashlight would). Though my meter is showing the outside case temperature it's reaching
    .

    From looking at the reviews you and OLD posted it appears the case temp. @ thermal step-down is extremely low on the B3 (<120F degrees) compared to the typical step-down case temps. on most lights I've tested (150F degrees). It could be Zanflare has adjusted the thermal step-down threshold very low or the lights actually getting hotter internally and unable to quickly rid itself of the heat, or both.
    Mole

  73. #73
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Guys I'm back. Being away for almost two full days. You have wrote so many posts I need to read them twice ;(

    Regarding thermal transfer I agree there is missing solid wall behind led pcb. No doubt this would be the best they can do to improve it. Alu plate they are using had not thermal paste on the contact areas with the housing. This at least should be added if they won't change design. That alu plate is fixed with 4 screws to tight it up which improves heat tranfer somewhat, but this is not enough. Since they used copper led pcb its funny they did not do the rest right. And I don't see any reason they could not do it.

    I've also asked Archie (our member if you can recall debates about Xeccon) about used FR9809. Unfortunately he is very busy and gave me just short answer. I was very surprised with his answer:

    "PDF you've noted is for a DC converter, not LED driver. And yes, it
    uses PWM technique for regulation. Technically, it's possible to use
    it that way (e.g., by connecting resistor in series), but it's a kind
    of "amateur" level of design, so I'm surprised if it's indeed used in
    some LED light."

    Don't know what to think, is this hitech innovation to use FR9809 or "amateur" design as he said? Indeed I found resistor R016 is in series with chip output.
    Review: Zanflare B3 bicycle light-driver-fr9809.jpg

    I think we should ask someone more knowledgable at BLF for second opinion.

  74. #74
    mtbr member
    Reputation:
    Join Date
    Nov 2012
    Posts
    1,043
    I wonder if this is going to be like some of the Nightfigter lights. Seems good at first, but after couple of months of use it cooks its parts to death.

  75. #75
    RAKC Industries
    Reputation:
    Join Date
    Jan 2017
    Posts
    3,265
    I would take Archies answer at 100% correct. He's an Electrical engineer more or less.

    Same time i would still ask the BLF crew their thoughts as they may have seen this before.

    Sent from my SM-G950U using Tapatalk
    Life on a bike doesn't begin till the sun goes down.


  76. #76
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Who do we have at BLF? Garry somewhere mentioned someone to be in review loop that might inspect the driver.

  77. #77
    mtbr member
    Reputation: MRMOLE's Avatar
    Join Date
    May 2011
    Posts
    2,871
    Would appreciate if one of the reviewers would run the light in the programmable mode (with a ambient temp. reference) to see at what level the light becomes thermally stable. Thanks!
    Mole

  78. #78
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    My BLF contact admitted he's no electrical expert and would only analyze basics. I think we need someone more knowledgeable (no offense to him).

    I would Mole, but my work situation took away the majority of my free time now. (I'm not happy about it.). I was going to try and get some internal temperature measurements and probably then pass it to Cat so I'm not holding him up.

    -Garry

  79. #79
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    I've PM'd my contact. Will see what he has to say.

    -Garry

  80. #80
    mtbr member
    Reputation: MRMOLE's Avatar
    Join Date
    May 2011
    Posts
    2,871
    [QUOTE]
    Quote Originally Posted by garrybunk View Post
    I would Mole, but my work situation took away the majority of my free time now. (I'm not happy about it.). I was going to try and get some internal temperature measurements and probably then pass it to Cat so I'm not holding him up.
    Thanks Garry, just a request if you had time. Internal temps. should be interesting. Look forward to hearing Cat's view on the light too! Hopeful the B3 will perform a little better on an actual ride compared to indoor test environment.
    Mole

  81. #81
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    I have mine still dissasembled. Right now done test with measuring input powers at driver input DC+/DC-. I've soldered small led voltmeter and additional input wire to those 2 points so I could put accurate DMM inline just in front of driver. Started with freshly charged 4 cell Panny pack from KD.

    At a time of pluged the pack it pulled 0.02A @ 8.45V

    Mode Low Medium High

    1 emitter ( 0.22A @ 8.40V ) ( 0.65A @ 8,28V ) ( 1.54A @ 8.08V = 12.44W )
    2 emitters ( 0.49A @ 8.32V ) ( 1.47A @ 8,10V ) ( 3.58A @ 7.58V = 27.13W )
    3 emitters ( 0.73A @ 8.25V ) ( 2.36A @ 7.87V ) ( 6.67A @ 6.83V = 45.55W )

    Didn't wait to cool down between test, so the results may be temperature dependable. Now if we recall my measurments at led side we can calculate efficiency:

    1 emitter Low Mid High
    Pin_ 0.22A @ 8.40V = 1.84W 0.65A @ 8.28V = 5.38W 1.54A @ 8.08V = 12.44W
    Pout 0.48A @ 2.83V = 1.36W1.45A @ 3.11V = 4.51W 3.02A @ 3.41V = 10.30W
    Effi. 1.36W / 1.84W = 74% 4.51W / 5.38W = 84% 10.30W / 12.44W = 83%

    Lower efficieny at low mode (low current) pretty much coresponds to Figure 6 in FR9809 Datasheet. Of course this is for one led only.

    At more leds turned on things might be bit different due higher tempeatures. But if we take into account all leds are runt at same power 10.3W then we get 20.6W / 27.1W = 76% for two leds and 31W / 45W = 69% for all three leds.

    I was unshure so I took another battery pack, noname 6 cell and got 44W input ie. 70%. So there is realy quite bad efficiency at highest power which adds to heat of course.

  82. #82
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    So as we were suspecting. Thanks ledoman. Have you ever taken a good pic of the inside with everything removed including the mounting "plate" so we can see how little thermal path there really is?

    -Garry

  83. #83
    mtbr member
    Reputation: Cat-man-do's Avatar
    Join Date
    May 2004
    Posts
    6,351
    Seeing I don't have one of these yet I'm betting the problem with the heat build up has to do with the multiple drivers. Keep in mind, this might have been done to keep cost down while still allowing the lamp to provide an excellent output. Then again the multiple drivers were likely necessary to facilitate the added 1-2-3 LED function. I think I can live with the loss in efficiency as long as the lamp can maintain a 2K+ lumen output over several minutes.

    I'm still confused about how the memory function works. I now know the lamp will work with all leds on and still have three steady modes. I know those modes can be reprogrammed but....will the lamp hold those settings if you continue to change modes? I know if you turn the lamp off the lamp goes back to default settings. Is this correct?

  84. #84
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Cat, no you can't program the modes. You simply enter into what they term "Personalized brightness setting" (think of it like another hidden "mode" with no memory). You simply change that "mode" brightness in 10% increments but can't save it - you exit back out to the "normal"/preset modes. The memory works to turn back on with the same # of LEDs at the same mode level you were last at.

    -Garry

  85. #85
    RAKC Industries
    Reputation:
    Join Date
    Jan 2017
    Posts
    3,265
    Would have been way cheaper to use a standard buck driver and run them in parallel. But yes your right, individual led controls require independent drivers. An "amped up" version of a kd2 driver would have been ideal. Everything was already there, modify firmware, use components needed to handle higher amps and print to a board that fits the housing.

    Would have allowed for better thermal path, desired output, heat wouldnt be a problem, cheaper, easier, programming that works, etc.

    Stinks they completely ignored the fundamentals that even lights light nightfighter and yinding got insanely right the first try.

    Props for the attempt and trying to do something different but they found a line you just cant cross, ignoring heat generated by low efficiency drivers and using more than one together.

    Sent from my SM-G950U using Tapatalk
    Life on a bike doesn't begin till the sun goes down.


  86. #86
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Mole, video coming later, but testing shows mode level 8 (80%) without air blowing across will stepdown, but with airflow will take quite awhile to stepdown. Mode level 7 (70%) with airflow is thermally stable. I could work up approx. lumens at those levels if you wish.

    Like I said, video coming (need time to download, join together, and upload). I also committed a cardinal sin during the filming of the videos.

    -Garry

  87. #87
    ZFL
    ZFL is offline
    mtbr member
    Reputation: ZFL's Avatar
    Join Date
    Aug 2017
    Posts
    19
    [QUOTE=MRMOLE;13308347]

    If you scroll down the indiegogo site till you get to the spinning 3D image of the B3 there is a chart directly below. At the top/right side of the chart is the emitter information and an explanation of how they intend to handle the neutral white lamps. The reason you cannot select neutral white is the only thing you are doing on this site is donation to the crowd funding project. Your not actually ordering a light so no guarantee you'll get anything (if I understand the disclaimers correctly). Personally will not order till I can select neutral white from Gearbest.


    Mole
    In the spec sheet we have a note that the buyer can choose NW or CW. We don't part the NW and CW in the perks because of too many catalogs.

    It is not a donation. It is shopping there with a special price. I don't think there will be a cheaper price than $39 later. Indiegogo is a famous website in the world. There are more guarantees than anywhere.

  88. #88
    mtbr member
    Reputation:
    Join Date
    Jul 2013
    Posts
    751
    Quote Originally Posted by ledoman View Post
    Regarding thermal transfer I agree there is missing solid wall behind led pcb. No doubt this would be the best they can do to improve it. Alu plate they are using had not thermal paste on the contact areas with the housing. This at least should be added if they won't change design. That alu plate is fixed with 4 screws to tight it up which improves heat tranfer somewhat, but this is not enough. Since they used copper led pcb its funny they did not do the rest right. And I don't see any reason they could not do it.
    This was a tough one for me, the just didn't get they idea of not drilling the case completely through and leaving a wall for the pcb. I sent them pictures of the disassembled Yinding and drew pictures to explain. Remember we originally started this custom light by asking Yinding to just add betters fins since they had the best case for Thermal transfer at the time.

    i am not as technical as you guys, though I have a few contacts over at GB/Zanflare. Let me see if I can get someone more technical to help.

    Mine should arrive within a few days, i admit I am not as tech as Ledoman and Garry and they have done an awesome job. I am also willing to send mine on to Cat if Garry can't.

  89. #89
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Quote Originally Posted by garrybunk View Post
    So as we were suspecting. Thanks ledoman. Have you ever taken a good pic of the inside with everything removed including the mounting "plate" so we can see how little thermal path there really is?
    No, it is hard to get that plate out. In fact I could not since there is input cable wich prevents me to do that. And I don't want to desolder this thick cable. I may ruine the driver with heat. Would need more powerful solder tool to do it very fast. Anyway will try again before applaying some thermal paste and final assembling back together. Will ask also what paste they are using - what thermal conductivity.

  90. #90
    mtbr member
    Reputation:
    Join Date
    Nov 2012
    Posts
    1,043
    Quote Originally Posted by ZFL View Post
    In the spec sheet we have a note that the buyer can choose NW or CW. We don't part the NW and CW in the perks because of too many catalogs.

    It is not a donation. It is shopping there with a special price. I don't think there will be a cheaper price than $39 later. Indiegogo is a famous website in the world. There are more guarantees than anywhere.
    What happens if you don't make your $20,000 goal. Do people get a refund? Why not make a separate buy link for neutral white? It's much simpler that way.

  91. #91
    RAKC Industries
    Reputation:
    Join Date
    Jan 2017
    Posts
    3,265
    https://support.indiegogo.com/hc/en-...my-money-back-

    Its up to the campaigner to issue refund after it has ended, no protection after that point.

    Sent from my SM-G950U using Tapatalk
    Life on a bike doesn't begin till the sun goes down.


  92. #92
    mtbr member
    Reputation: Cat-man-do's Avatar
    Join Date
    May 2004
    Posts
    6,351
    Quote Originally Posted by garrybunk View Post
    Cat, no you can't program the modes. You simply enter into what they term "Personalized brightness setting" (think of it like another hidden "mode" with no memory). You simply change that "mode" brightness in 10% increments but can't save it - you exit back out to the "normal"/preset modes. The memory works to turn back on with the same # of LEDs at the same mode level you were last at.

    -Garry
    .........Say it ain't so Joe....Oh well, big mistake on their part but good news for their competition, Gloworm and ITUO. Of course the biggest issue will be "usability". If the reflectors and the standard modes of the B3 provide the users with a usable platform they will still sell. Of course I won't know if that will be so until I see one in person.

    As it is, if someone finds the lamp too bright on medium ( 3-LED ) they can either work the personalized mode ( which would be a PITA ) or switch the lamp to the 2-LED mode and use the medium on that setting. Either way it's not something that can be done while riding unless the lamp is bar mounted and you don't mind pressing a lot of buttons to get the output that you want.
    Last edited by Cat-man-do; 08-28-2017 at 11:11 AM.

  93. #93
    ZFL
    ZFL is offline
    mtbr member
    Reputation: ZFL's Avatar
    Join Date
    Aug 2017
    Posts
    19
    Quote Originally Posted by varider View Post
    What happens if you don't make your $20,000 goal. Do people get a refund? Why not make a separate buy link for neutral white? It's much simpler that way.
    NW and CW are almost the same except the tint . So they are at the same page.
    if we don't make the $20,000 goal, we still can send the items ordered. Of course the buyer also can ask for a refund if he wants.

  94. #94
    ZFL
    ZFL is offline
    mtbr member
    Reputation: ZFL's Avatar
    Join Date
    Aug 2017
    Posts
    19
    Quote Originally Posted by RAKC Ind View Post
    https://support.indiegogo.com/hc/en-...my-money-back-

    Its up to the campaigner to issue refund after it has ended, no protection after that point.

    Sent from my SM-G950U using Tapatalk
    Zanflare is an integrity brand. So please don't worry about the money. Even if the project is ended, we also can refund via other methods, such as Paypal, T/T.

  95. #95
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Datasheet for FR9809 states:

    "FB - pin 6 - Voltage Feedback Input Pin. FB and VOUT are connected by a resistive voltage divider. This IC senses feedback voltage via FB and regulates it at 805mV"

    At page 10 you can see how to calculate output voltage. Vout = 0.805 * (1 + R1/R2)

    I'm suspecting (do not take it as pure truth) they are using voltage regulator and different paths over resistors to get different output voltages - resulting in different modes. For example R1 (4.99 KOhm) and R2 (10 KOhm) at Figure 2 defines output of 1.2V. Since those two resistors makes voltage divider their combination defines output. Folowing static paths over resistors prevents "programable" features I'm afraid. Don't know how the Customized levels are implemented, though.

  96. #96
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    My BLF contact says he thinks he can get us much further than we are currently on driver diagnosis, but the driver may not survive dissection. If Cat is willing, he could forward the light on when he's done his review. Or I can skip Cat and let GJHS ship his to Cat. Any thoughts?

    -Garry

  97. #97
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Garry, per your request I've managed to get out Alu plate (was bit hard). Here are some pictures. Dimensions first:
    Review: Zanflare B3 bicycle light-alu-backplate-dimensions1.jpg

    Review: Zanflare B3 bicycle light-alu-backplate-dimensions2.jpg
    EDIT: forgot to tell it's 2mm thick (on average, not uniform thicknes)

    Areas of heat transfer (around 4 screws and tiny 1mm lips, not all have contact)
    Review: Zanflare B3 bicycle light-alu-backplate-areas-heat-transfer2.jpg

    Areas of contacts marked with red.
    Review: Zanflare B3 bicycle light-alu-backplate-areas-heat-transfer.jpg

    As is (without solid wall) it would be better if little biger leds PCB would contact those areas directly. Of course those areas on leds PCB should be without dielectric layer so the bare copper will stick to the body.

    So either they should make solid wall behind leds PCB or make apropriately bigger and thicker leds PCB and this alu plate would be redundant. Of course thermal paste should be applied to corect places in either case.
    Last edited by ledoman; 08-29-2017 at 02:38 AM.

  98. #98
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Have you verified the PCB is actually DTP? (Direct Thermal Path) Oh, and how thick is that aluminum plate?

    That contact area looks horrendously small. I'm surprised this light hasn't burned up yet. Definitely would without thermal monitoring.

    -Garry

  99. #99
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Do you know the method to verify DTP without desoldering leds and file/scratch the solder away? Anyway at BLF I've seen there is not such big difference between DTP and Dielectric layer as long as you have copper PCB.

    Alu plate is 2mm thick on average, added it in the post above. Funy it's not exactly 2 mm thick over the whole area. Maybe for this test samples are more like hand maded.

  100. #100
    RAKC Industries
    Reputation:
    Join Date
    Jan 2017
    Posts
    3,265
    Tiny probes for DMM. Touch the center solder joint of the emitter and the back side of the pcb. Thats all DTP is, the center of the emitter is soldered more or less directly to the copper main plate

    Sent from my SM-G950U using Tapatalk
    Life on a bike doesn't begin till the sun goes down.


  101. #101
    mtbr member
    Reputation: MRMOLE's Avatar
    Join Date
    May 2011
    Posts
    2,871
    [QUOTE]
    Quote Originally Posted by garrybunk View Post
    My BLF contact says he thinks he can get us much further than we are currently on driver diagnosis, but the driver may not survive dissection. If Cat is willing, he could forward the light on when he's done his review. Or I can skip Cat and let GJHS ship his to Cat. Any thoughts?
    At this point what I want to hear about is how the light performs when riding. Whatever gets Cat the light quickest is what I vote for. Thanks for the thermal stabilization point info.
    Mole

  102. #102
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Thermal Stability Testing Video now uploaded to YouTube here: https://youtu.be/Zoc7ZjsvhPg (for those that need help falling asleep tonight).

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  103. #103
    mtbr member
    Reputation:
    Join Date
    Jul 2013
    Posts
    751
    [QUOTE=MRMOLE;13312001]

    At this point what I want to hear about is how the light performs when riding. Whatever gets Cat the light quickest is what I vote for. Thanks for the thermal stabilization point info.
    Mole
    I just got mine today, i will take it out for a real world ride in the next couple of days

  104. #104
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    [QUOTE=GJHS;13312146]
    Quote Originally Posted by MRMOLE View Post
    I just got mine today, i will take it out for a real world ride in the next couple of days
    Great! Would love to hear how it works out and how well it manages heat.

    -Garry

  105. #105
    mtbr member
    Reputation: MRMOLE's Avatar
    Join Date
    May 2011
    Posts
    2,871
    Garry, great youtube segment. All those different readings provided additional information about thermal threshold temp. and thermal path capacities. Limited time currently but tomorrow will post my thoughts on what it all means.

    GJHS, look forward to your ride review.

    Thanks guys!!!
    Mole

  106. #106
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Quote Originally Posted by RAKC Ind View Post
    Tiny probes for DMM. Touch the center solder joint of the emitter and the back side of the pcb. Thats all DTP is, the center of the emitter is soldered more or less directly to the copper main plate
    With my ordinary probes I couldn't get any contact with bare copper. It was late so didn't try it with another pin like probes. Will try it today. As said before I want to point out there is no such big difference between copper DTP and copper PCBs with dielectric layer as long as the heat can succesfuly flow away to some heatsink (ie. light body in our case).

    Based on djozz tests at BLF there is no real difference up to 3A. Take a look his chart and follow the green line. You can read the whole thread HERE. You may read also his tests of XP-L V6 and many others if you have time.

  107. #107
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Quote Originally Posted by garrybunk View Post
    Thermal Stability Testing Video now uploaded to YouTube here: https://youtu.be/Zoc7ZjsvhPg
    Garry can you or someone else write resumption of this test? It would be much easier to understand.

    I've also confirmed they are using DTP. I've got contact with led pad and other side of leds PCB. I just needed to use pin like DMM probes and have to scratch off flux residues from led base. So they have done that very good. Out of my head I can't recall any light that is using DTP. Not saying there is none, just I'm not aware of them.

  108. #108
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Quote Originally Posted by ledoman View Post
    Garry can you or someone else write resumption of this test? It would be much easier to understand.
    Personalized brightness setting of 70% (7th level) with no airflow and 69º room temperature reaches thermal stepdown (50% lower) in approximately 3 minutes 45 seconds. That same mode with a small fan blowing what is estimated to be 5 to 7 mph winds across the light will remain in that mode without stepping down (I believe). 80% level (8th mode) with the same fan takes approx. 15 minutes to reach thermal stepdown with an external case temperature of 138ºF.
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  109. #109
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Thanks Garry.

    Additional measurments of voltage tresholds for status led at driver input. Keep in mind that actual voltage at the battery side would be higher than at driver side due to voltage drop. So I read the voltages that driver gets on input since the battery side could be variable ie. different battery packs and cables.

    Voltage treshold Status led
    6.85V blue to red
    6.03V steady red to red blinking
    5.65V red blinking to light completly off

    You need to disconnect the power to reset the status if the voltage for any reason gets higher. Between 6.03V and 5.65V after reconnecting red status led is blinking even if the light is turned off. It looks this is warning the battery voltage is below 6V. If the voltage is below 5.65V the light won't even turn on.

  110. #110
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    I now realize that I forgot to explain that my Thermal Stability video got cut short due to the camera battery going dead. (I was in a rush to get it posted.) It's explained in the video description on YouTube. After the video ended I continued running the light in mode level 8 (80%) until it hit thermal stepdown - about 15 minutes at 138 degrees F (external case temperature).

    -Garry

  111. #111
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Wow! This light is not for the faint at heart to tear apart! Not a light that every user could/should open up to add thermal compound. It's very delicate inside. They really crammed things in there tightly. I'm trying to keep it in working order for Cat. Ledoman talked me into opening it back up to add the thermal compound (blame him if I break something!).

    I did get some pretty good internal pics. I'm impressed with the seal around the power cord entry (one of the best I've ever seen). Just continue to be baffled by the horrendous thermal path (or lack thereof)! If I had time to reassemble with my temperature probe at the emitter mounting (and then remove & reassemble again) I would, but that's going to eat up too much time.

    Oh, when I popped the driver out I found one of the positive emitter wires broken open, a short waiting to happen!

    I'll post some more pics tomorrow.

    -Garry

  112. #112
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    I have some very good news (at least for Cat)! The light is reassembled and it still works! It was very delicate and touch and go for awhile, but I did it. I powered it up on a freshly charged pack and I would say it feels like it's getting hotter faster and quicker (could just be my imagination though). I think I will run a comparative test to my initial "Thermal Properties" video where I run it on high (with fan) in my sphere and see if lux values are any steadier and time how long until thermal stepdown (it was 1 minute 11seconds in the initial test).

    Entire Photo Gallery here: Zanflare B3 Bike Light by Garry Bunk | Photobucket

    Some pics:

    Thermal compound in place as opened:


    Lenses & Holder (note: no AR coating seen on the lenses):


    Reflectors:


    Sealing of power cable:


    Pulling driver out and exposing aluminum plate:


    Cracked open wire waiting to short out:


    Driver:






    Aluminum plate removed showing minimal body contact (looking at backside of copper emitter board):


    Showing aluminum plate thickness:


    Both emitter board and driver unmounted:


    Cleaned thermal compound off and found emitter pcb extremely scratched:


    Lots of wires not to break joints on!:


    Thermal compound applied at limited mount contact area:


    Thermal compound applied around edges after aluminum plate is mounted back in:


    New thermal compound applied to emitter pcb:

    Messy, but functional:


    Completely reassembled and ready for use again:


    After this next thermal test I will ship it out to Cat - possibly tomorrow if I can find time to run the test tonight. Perhaps he can even receive it before the holiday (Labor Day here in the US).

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  113. #113
    mtbr member
    Reputation: MRMOLE's Avatar
    Join Date
    May 2011
    Posts
    2,871
    [QUOTE]
    Quote Originally Posted by garrybunk View Post
    Personalized brightness setting of 70% (7th level) with no airflow and 69º room temperature reaches thermal stepdown (50% lower) in approximately 3 minutes 45 seconds.
    Point of reference, my Yinding (w/stock mount) takes about the same time from a cold start to step-down. Thanks Garry for all those extra instrument readings (time, current draw, lux, along with temp.). Provided some good information on lighthead thermal performance/capacity. While teardown pictures show obvious shortcomings in the thermal path the 3.6A current draw @ a stable level 7 indicates the thermal capacity of the lighthead should be adequate to handle the B3's max power if it weren't such an inefficient light. Current draw @ max power on my slightly more powerful XS (3.1A) and XP3 (3.36A) are far less than the B3's 5.25A (lowest recorded by reviewers). Ya, improving the thermal path would allow the B3 to run stably at higher power settings but improving the lights efficiency to more normal levels (even if they have to eliminate the redundant separate emitters controls) would do the same + allow lower interior operating temps. (better reliability) and the use of a much smaller battery (lower weight/less cost). Additional point of reference, my C&B Seen 7up (similar cost +wireless 2 button remote) produces slightly less power than the B3 but is stable at that output in 100 degree temps. and only requires 2.7A to do that.
    Mole

    ***Garry, hope you'll rerun stability test after thermal paste addition + what are the approx. lumens for level 7?***

  114. #114
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    [QUOTE=MRMOLE;13314799]
    ***Garry, hope you'll rerun stability test after thermal paste addition + what are the approx. lumens for level 7?***
    Boy having these videos sure makes it easy to snag data you weren't paying attention to or didn't think you wanted . Anyway, my rough calculations are:

    80% no fan test: @ startup: 2,095 lumens, just prior to stepdown: 1,825 lumens
    70% with fan: @ startup: 1,915 lumens, at end of 70% level testing: 1,660 lumens
    80% with fan (light already hot): 1,845 lumens, at cutoff of video: 1,780 lumens.

    (I didn't catch lux reading at 80% stepdown with fan.)

    Hope that helps!

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  115. #115
    mtbr member
    Reputation: MRMOLE's Avatar
    Join Date
    May 2011
    Posts
    2,871
    [QUOTE][QUOTE=garrybunk;13314884]
    Quote Originally Posted by MRMOLE View Post

    Boy having these videos sure makes it easy to snag data you weren't paying attention to or didn't think you wanted . Anyway, my rough calculations are:

    80% no fan test: @ startup: 2,095 lumens, just prior to stepdown: 1,825 lumens
    70% with fan: @ startup: 1,915 lumens, at end of 70% level testing: 1,660 lumens
    80% with fan (light already hot): 1,845 lumens, at cutoff of video: 1,780 lumens.

    (I didn't catch lux reading at 80% stepdown with fan.)

    Hope that helps!
    Exactly what I wanted, Thanks!
    Mole

  116. #116
    mtbr member
    Reputation:
    Join Date
    Nov 2012
    Posts
    1,043
    Does that aluminum plate actually make contact with the side the of the light, or is just resting on the lip and contacting on the red patches that ledoman outlined? Garry, wasn't there some kind of compound that the BLF guys use to close that gap (with the side)?

    From what you guys are describing that plate is hardly doing anything.

    From the pictures it looks they did a pretty large amount of machining on the inside of the light to create the cavity where the led board slides into. That's exactly what we didn't want them to do. Ideally we wanted a solid wall, with the led board sliding in from the front. Here it looks like it slides in the from the back into that cavity.

  117. #117
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Quote Originally Posted by varider View Post
    Does that aluminum plate actually make contact with the side the of the light, or is just resting on the lip and contacting on the red patches that ledoman outlined? Garry, wasn't there some kind of compound that the BLF guys use to close that gap (with the side)?
    Nope, no contact to the sides, which is why I tried to push thermal compound in the joint to help a bit. BLF compound? Hmm . . . you referring to potting compound? Like posted here? I've not actually used or done any potting (other than just thermal compound smeared all over my BT40S driver). I do have the duct work compound linked in the thread, but never tried it. I'm afraid there's not enough room in the B3 head by the time you cram that driver in there too, though I guess you could get a little in there. Not sure how effective it would be in this case.


    Quote Originally Posted by varider View Post
    From what you guys are describing that plate is hardly doing anything.
    Exactly! It's amazing it doesn't burn up! Even with thermal monitoring!

    Quote Originally Posted by varider View Post
    From the pictures it looks they did a pretty large amount of machining on the inside of the light to create the cavity where the led board slides into. That's exactly what we didn't want them to do. Ideally we wanted a solid wall, with the led board sliding in from the front. Here it looks like it slides in the from the back into that cavity.
    Yes, this was numero uno on our list of specs for the custom build! They did pretty good with other areas (finning and using a DTP emitter pcb), but really messed this up!

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  118. #118
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Good job, Garry. I think the most heat is transfered from Alu plate at those 4 areas where it is screwed to the body. The contact is tight so less room for air gaps. Of course it would be much better if that area would be larger and thermal paste applied.

    I've also noticed the thermal paste under Led PCB in both cases (mine and Garry's) is smashed in the middle while at both sides looks bit thicker. Looks like on sides it doesn't stick to Alu plate that well.

    The difference with mine sample is also thermal compound over all 3 FR9809 step-down chips. Don't know why since it doesn't stick to anything to transfer heat away. I've noticed those chips are pretty hot too. Of course by specifications at least 10% of energy goes to heat.

  119. #119
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    They really should have screwed down the emitter pcb or at least had outer edge screws on that outer edge faceplate. Right now there's not much force on the outer (left & right side) of the pcb pushing it to the aluminum plate since you only have force from the 4 screws in the middle emitter area of the faceplate. It does seem to work ok though. The first time I opened things up and pried out the emitter pcb I bent the board trying to pry it out (it hardly has clearance around the edges of the body). I was concerned that the bent side emitter would burn up as it might not have had enough pressure pushing from the reflector to seat the pcb against the aluminum board. I did straighten it up after the 2nd disassembly.

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  120. #120
    mtbr member
    Reputation:
    Join Date
    Nov 2012
    Posts
    1,043
    Potting compound, that was it.

    So I watched your last video Garry earlier today and it appeared to be stable at 70% (with fan) for a long time. At 80% (with fan) the temperature of the case (you measured outside?) keeps climbing and eventually reaches thermal stepdown. So really anything above that, 90 or 100% (in that personal brightness mode) it would just step down faster.

    In the regular low-medium-high rotation, the highest mode (high) probably corresponds to 100% in that fourth mode. So the high mode might suddenly drop down because of thermal throttling when you are counting on that extra light the most, such as on a long downhill. That would not be good. Of course the extra speed might be enough to cool it down.Then again maybe the high mode is not usable for anything over a few minutes.

    If they had used a solid wall behind the led and made the driver way more efficient it probably wouldn't have these types of problems. If the driver actually allowed the modes to be programmed, then you could set high to 70% and not have to worry about it. 100% could then be a special turbo mode, although I generally hate those with a passion.

  121. #121
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Last measurments before assembling it back. I wanted to see what are voltage drops from the battery to the driver. I've used an old 6 cell pack consisted of salvaged cells from some old notebook. The reason I've used it - it had balancing connector so I could monitor voltages right at batteries output. At driver side I've already had voltage monitor set for previous tests at input terminal DC+/DC-. As

    Will add later all mesurments. Right now I can say there was 0.38V drop at full power and full battery. Since driver is regulated and current will rise as the battery gets weaker so will rise the voltage drop. This would lead to premature signalization of low battery and very likely power fall back. It will be more obvious with weaker batteries and packs with less cells.

    Edit: voltrage drops on all modes. From battery side -> driver side

    One Led
    L 8.36 -> 8.33 = 0.03V
    M 8.33 -> 8.29 = 0.04V
    H 8.31 -> 8.25 = 0.06V

    Two Leds
    L 8.32 -> 8.26 = 0.06V
    M 8.24 -> 8.13 = 0.11V
    H 8.15 -> 7.95 = 0.20V

    Three Leds
    L 8.22 -> 8.12 = 0.10V
    M 8.00 -> 7.78 = 0.22V
    H 7.71 -> 7.29 = 0.38V

    Note also how the voltage sags on the battery side as more power is pulled fom it.
    Last edited by ledoman; 09-02-2017 at 11:30 PM.

  122. #122
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Ledoman, what is the cause of the voltage drop? Just the power lead & connectors?

    My light has now been shipped off to Cat and delivery is expected this Saturday. Looking forward to his take on this light!

    I have just finished uploading my 2nd Thermal Analysis video (testing after adding more/better thermal compound). I will say, results were surprising and I'm wondering if I messed something up during disassembly (I don't think so though). I ran 3LEDs in high mode (same as my first Thermal Analysis video - but not to be confused with the last Thermal Stability video) with the small fan. In the first test (in stock form, battery pack slightly discharged) it ran for 1 minute 11 seconds until it stepped down. In this test, it only ran 1 minute 26 seconds until it stepped down! The battery pack was nearly fully charged, but I expected it to run longer. Then what is surprising is that the light constantly bounces back and forth between High, Mid, High, Mid, etc. and this is where I wonder if I caused an issue or if this is normal. Anyway, here is a link to the video: https://youtu.be/l6-sZw9sE68

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  123. #123
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Garry, yes it is just power leads and one connector. On battery side I've used 0.5m long 20AWG lead and standard MagicShine plug. On the light side I've used stock lead and male connector as provided by manufacturer.

    At higher currents everything count and we should use less resistant connector. Yup, I blame connector mainly. It is designed for up to 2A, not for 5 or more.

    Besides this voltage drop there is of course voltage drop of batteries itselfs. Higher the current lower is the voltage battery can provide. And this goes in the loop as the battery gets more depleted providing lower and lower voltage regulated driver pulls higher and higher current which lowers the voltage even more. This is the problem of konstant power this driver wants to deliver.

    As an ilustartion HKJ at BLF is testing batteries with constant power of 10W. Just as an example of dependancy between voltage and current I took one of his charts:
    Review: Zanflare B3 bicycle light-battery-constant-power.jpg
    Source: Test of Panasonic NCR18650B 3400mAh (Green)

    Very same dependancy is happening with regulated driver in our B3 light.

  124. #124
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    ledoman - thanks and good point that the connector is only designed to 2A (well I have to assume that is the case as I don't know for sure). What is the best way to handle the connector? I know the "standard one" only has those two small points of contact and not the entire barrel - could we ask them to revise the connector to one with the entire barrel having contact? Or is that still not good enough? Whatever change they would make, the connector would have to be the standard size in order to be universal. Only other option would be to bundle with a battery pack and use proprietary connectors.

    What is your take on my new thermal test and the light bouncing back & forth between high & medium?

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  125. #125
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Concerning heat and thermal stepdown, as a point of reference, my custom modded HD-016 light doesn't stepdown until it reaches 166º F (I believe that's an external case temp) and steps back up when it cools down to 120º (it's based on the KD programmable driver mcu).

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  126. #126
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Quote Originally Posted by garrybunk View Post
    ledoman - thanks and good point that the connector is only designed to 2A (well I have to assume that is the case as I don't know for sure). What is the best way to handle the connector? I know the "standard one" only has those two small points of contact and not the entire barrel - could we ask them to revise the connector to one with the entire barrel having contact? Or is that still not good enough? Whatever change they would make, the connector would have to be the standard size in order to be universal. Only other option would be to bundle with a battery pack and use proprietary connectors.

    What is your take on my new thermal test and the light bouncing back & forth between high & medium?

    -Garry
    To manage the high currents we would need another connextor like XT30 / 60, Dean or similar. MS type 5.5x2.1 is good for compatibility, but is not good for performanes.

    To be honest I had no time to look video. I'm packing for vacations, doing things around house, ... All I can say it happening with my sample too, steping down and up, but my tests were static with dissassembled light last few days. As I can recall driver somehow balance temperature to be under 60 °C. ZFL has described it as:

    "In a same mode, when the temperature is NOT more than 60 degrees Celsius, it will maintain constant power. If more than 60, it will lower the brightness (and power) slowly. In this process, if temperature is more than 80, it will turn to mid mode immediately to protect circuits."

  127. #127
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    I've not seen any "lowering power slowly", just immediate change to mid mode.

    -Garry

  128. #128
    mtbr member
    Reputation: Cat-man-do's Avatar
    Join Date
    May 2004
    Posts
    6,351
    Quote Originally Posted by garrybunk View Post
    Ledoman, what is the cause of the voltage drop? Just the power lead & connectors?

    My light has now been shipped off to Cat and delivery is expected this Saturday. Looking forward to his take on this light!

    I have just finished uploading my 2nd Thermal Analysis video (testing after adding more/better thermal compound). I will say, results were surprising and I'm wondering if I messed something up during disassembly (I don't think so though). I ran 3LEDs in high mode (same as my first Thermal Analysis video - but not to be confused with the last Thermal Stability video) with the small fan. In the first test (in stock form, battery pack slightly discharged) it ran for 1 minute 11 seconds until it stepped down. In this test, it only ran 1 minute 26 seconds until it stepped down! The battery pack was nearly fully charged, but I expected it to run longer. Then what is surprising is that the light constantly bounces back and forth between High, Mid, High, Mid, etc. and this is where I wonder if I caused an issue or if this is normal. Anyway, here is a link to the video: https://youtu.be/l6-sZw9sE68

    -Garry
    Garry, my take on this is that when you improved the thermal path of the emitters that this also speeded up the heat toward the original heat sensor ( wherever that may be in the lamp ). No way a lamp should go into thermal power-down so fast. The circuit that controlled the thermal power-down was likely calibrated for the poorer thermal path but this is just my guess. In your video the fan you used looked pretty small so I'm hoping when I do some tests with a larger fan I won't get the power-down in quite the same way. More importantly I certainly wouldn't want the lamp to power down that fast as I'm riding at a decent speed. If it does I would consider it useless unless there is some way to disable the thermistor ( or somehow recalibrate it to a higher temperature ).

  129. #129
    mtbr member
    Reputation:
    Join Date
    Jul 2013
    Posts
    751
    Tomorrow I will give a real world test on a ride and report on how long it takes to dim and how it behaves with a breeze. I meant to do it yesterday on a new trail that I completely underestimated. It was the most technical trail I have ever walked, I mean rode and there was no way I was trying it again in the dark. For the Blue Mountain Reserve riders who guided me yesterday: thank you again.

    Anyone around NYC who wants to see the light, PM me.

  130. #130
    mtbr member
    Reputation:
    Join Date
    Jul 2013
    Posts
    751
    OK I took it out for a ride. It is 72 degrees here in NY

    Within 2-3 minutes with 3 Leds on High, the light stepped down to what I think is Medium then quickly went back to High and this up and down continued until you stepped the light down to a lower setting. Next I tried 2 Leds on High and the light stayed constant to my eyes. I even tried the light at its lowest with only one Led on low and it was still a decent amount of light. My sample had very minimal rings in the beam and put a decent amount of light on the front tire while still giving good throw.

    What I didn't like was the mount, it needs to be slightly taller. The screwhead hits the bottom of the light making it hard to tighten well enough so the light bounces on bumps. The button layout is OK, though you will have to practice, its easy to confuse which button does what on the fly.

    BTW I ran mine on the Nitefighter battery. Itried the cheap ones that came with the Yinding and it didn't work long.

  131. #131
    mtbr member
    Reputation:
    Join Date
    Nov 2012
    Posts
    1,043
    That's awful.

  132. #132
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    I've found another design issue which lowers heat transfer. Ok it might be better with the new light, but based on Garry's pistures I doubt (see below). In the process of assembling the light back I've found the alu backplate got bent on sides making gap to the light shell while normaly it should stick to the lips to transfer heat away.

    Review: Zanflare B3 bicycle light-alu-backplate-gap.jpg

    Review: Zanflare B3 bicycle light-alu-backplate-gap2.jpg

    You may recall I've made a comment on Garry's pictures and centraly smashed thermal paste while at sides paste looks thicker (see his first picture in post #112). Until now I thought there was just not enough pressure at sides to smash the paste. Wrong!
    Back side of the light (with switch button) is screwed to the Alu backplate trough the outmost holes bending backplate backawards. No surprise, the alu backplate is pretty weak where the alu backplate is screwed hrough 4 holes.

    Click image for larger version. 

Name:	Alu backplate areas of heat transfer.jpg 
Views:	155 
Size:	68.2 KB 
ID:	1154646

    So the heat is transfered only through central areas without sides I've marked red.

    If there was solid wall in the shell many issues would go away, including this!

  133. #133
    mtbr member
    Reputation:
    Join Date
    Feb 2016
    Posts
    710
    Excellent job Ledoman/Garrybunk. All the free R&D from you guys should not be ignored from Zanflare.

  134. #134
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Good point ledoman, I didn't even think of that plate being pulled back by the rear cover. And I'm surprised the Nitefighter battery ran the light on full blast without tripping the battery's protection circuit (but maybe there is no protection).

    -Garry

  135. #135
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Yes, I've observed it easier cause I've applied less paste than you. Maybe if you look your vertical picture (one before last in post #112) you may see some gap betwen led pcb and bacground. Also it seems to me the screw heads might cause led pcb can't fit tight too.

    On next opening I might try to put washer and tighten all together at least if the screws are long enough.

    During assembling I might ruined thermistor when I've pushed leds PCB to against Alu backplate. I'm getting pretty quickly switching power back and fort on the highest mode. I know I've scratched it and don't kono if sill works correclty. By ZFL it is 10K thermistor sized 0603 and I'm planning to replace it some day.

  136. #136
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Good idea on the washers. They might have to be custom cut more oblong shaped to fit around the emitters. They should have designed it that way so the screws would have held the PCB down in the first place. Heck, they should have moved the aluminum plate to the emitter side so it would be pulled against the body. Of course a solid wall would still be preferred.

    -Garry

  137. #137
    mtbr member
    Reputation:
    Join Date
    Nov 2012
    Posts
    1,043
    Very good detective work ledoman.

    As one of the guys pushing for this custom build, I would really hate to see this project go down in flames. But as it stands now, that's how how I see it.

    The number one requirement of a 2000 lumen light is heat removal. If you don't have that, then you have nothing. I know the forum regulars already know this, excuse my rant, but I feel the need to say it.

  138. #138
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Yeah, they're so close yet so far away. They'd really need to go back to the drawing board on this one and rework everything, but you know they won't.

    -Garry

  139. #139
    mtbr member
    Reputation: Cat-man-do's Avatar
    Join Date
    May 2004
    Posts
    6,351
    Today I received the lamp that Garry has sent me. Right now waiting for darkness so maybe I can take it for a spin on the local paved trails. Just shining it around in my home I've noticed that the beam tint is not real cold. Matter of fact it might indeed be a NW tint.
    Looks like it'll be a bit of a flood beam so I'll try it out on the bars first but if the beam carries some distance it might indeed make a good helmet light, assuming of course that the lamp doesn't constantly power down. Right now temps where I live are in the high 50's so likely I won't see too much power-down....BUT...we'll see about that.

  140. #140
    mtbr member
    Reputation:
    Join Date
    Feb 2016
    Posts
    710
    Cmd, nice!

  141. #141
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Edited post #121 with full table for voltage drop from battery to the driver.

  142. #142
    mtbr member
    Reputation: Cat-man-do's Avatar
    Join Date
    May 2004
    Posts
    6,351

    B3; first impressions

    Well...I tried. Even though the weather was horrible ( for testing lights ) I gave it a run for the money. Before I went out I had some trouble with the lamp. Seems the Gopro mount on the lamp is not really top notch ( no surprise there ). My original plan was to use it with the Gopro bar mounts that came with my ITUO lamps. Much to my dismay, the mount extension on the lamp itself would not slide over of the middle part of the ITUO GP mount. That was my first disappointment. Since I didn't want to risk breaking the plastic GP mount of the B3 that meant I was going to have to use the included mount that came with the lamp. Well, I'll be blunt; the Gopro ( clone ) mount(s) that come with the B3 lamp are pure crap...and that's saying it nicely. It would of been so much better it they had supplied a solid bar mount that clamped to the bars. The included bar mount they give you is designed to either work with an O-ring or the "ladder type" rubber attachment. They only give you the "ladder type", no O-rings are included. Need I say, the rubber ladder attachments completely suck. I got it to work but the mount still ends up having a lot of wobble and flex. Then there's the problem with the GP extension that comes off the lamp itself.. Seems they didn't make it long enough to allow the users fingers to have enough space to easily tighten the adjustment knob. ( Boy did they screw that up! ). When I decide to take this out on a real trail I'm going to have to figure some way to make the lamp more stable. I just tried one of my cheap Chinese Gopro mounts and that didn't work either. I finally had to go back to the original ( included mount ) using one of my own O-rings and a bar extension but I already know that no way am I going to be able to lock down the mount itself so it doesn't move. That means it's going to have play which in essence means the adjustment knob on the Gopro mount is pointless. Not to mention it will likely vibrate to some degree...anyway, back to the ride I did tonight...

    When I went outside I found out that a misty drizzle had settled in to the area. That meant that the paved MUP I was going to take was going to be damp and/or wet. I also took the Rokke along as back-up along with my A6 torch on the lid. I brought my ITUO XP3 along as well, stowed in my pocket.

    When I found a suitable spot I turned the B3 on. I have to say, it is quite bright. On full power it makes for one bright bar lamp and provides a nice even spread of light. I brought along my ITUO XP3 so I could compare beam patterns. It's quite possible that the B3 is brighter than the XP3 but undoubtedly the XP3 with all spot optics has the better throw. Still the B3 is super bright and throws a lot of light to the sides. Some people will find that very useful. Keep in mind that trying to judge throw during misty conditions is not easy. I'm sure that when I get drier conditions the throw will carry a bit farther.

    Right now I'm not sure if the high setting was powering down or not. I have a feeling it was because at times I thought I was in boost but when I changed modes I saw the lamp get brighter. Didn't help that the lamp seem to move out of adjustment whenever I touched it. Whatever, this is all first impression stuff. I need more time, better mounting solution, drier weather and more observations before I can come to any conclusions. I do seem to like the beam tint though. Of course in drier conditions that might change but we'll see.

  143. #143
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Quote Originally Posted by Cat-man-do View Post
    ..... Just shining it around in my home I've noticed that the beam tint is not real cold. Matter of fact it might indeed be a NW tint....
    Cat, see my first post and tint comparation. No way it is NW, but as you stated not realy cold white either. As said some time ago I think it's around 1D, 1C wich adds some yellowish hue. In fact they have declared it as 6000K-6500K. It's much better than those cheapo tints oriented more towards blue.

    So for the CW it is very good (tint wise). For NW they have declared 4000K - 4500K wich is quite warm around 4C, 5B on ANSI White. I would wote for 4C wich is already bit warmer than we have got with BT21 lights.

  144. #144
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    The GoPro mounting is something I didn't even evaluate. I believe the mount attached to the bottom of the light is aluminum.

    -Garry

  145. #145
    mtbr member
    Reputation: Cat-man-do's Avatar
    Join Date
    May 2004
    Posts
    6,351
    About the tint; When it comes to really bright light it's really hard to evaluate tint, especially if it's raining. Since it wasn't suppose to be NW I like to call tints like this "Bright white". Cool white tends to be more stark and has more of the blue spectrum in it even though a good cool white shouldn't look bluish. Anyway, FWIW it's using a fairly nice tint even if it's not NW.

    About the mount; Complete Sucksville. If I was buying one of these I'd have to find another way to mount it. Actually I consider it worse than your typical plastic O-ring mount because not only does it have less surface area in which to provide stability but the lamp sits higher ( because it is GP mounting ) and this makes it tend to wiggle and move more so than your typical cheap O-ring mounted light. I might have to run down to REI and see it they have a Gopro bar mount that will work with this light. Having very little experience with GP mounts I naively thought all GoPro type mounts were going to be, "universal fit". I kind'a learned that wasn't so about a year ago when I bought my first cheap Chinese GP mount but I must of forgot about it. If these go out as is Vanc should make some bucks making new adapters and mounts for these.

    Right now the major problem with the mount is that it's just not stable on the bars. When it comes to bar mounting I'm very picky. I do not tolerate excessive movement or vibration. If I can't find a bar mount that is stable I might have to use some Gorilla tape to keep it from vibrating. Right now I don't have a lot of options so I'll have to make due using what I have on hand.

    Seems odd to me that someone would engineer such an intricate design for a lamp and then totally drop the ball when it came to the mounting. If I was paying someone to design a mount for a lamp like this they would find themselves being fired real quick. This lamp needs a good standard "clamp type" GP mount and a longer *extension piece ( *the piece that screws to the bottom of the lamp ) off the lamp so when you try to adjust the tilt using the GP adjustment knob your fingers have room enough to grip the knob and turn.

    The helmet mount is Sucksville too. Once again, a very small platform in which to help with stability issues. They give you a short piece of velcro to use and it just isn't enough. If I use my own velcro I can create a longer piece but that's still no guarantee that I can make it stable without a bigger, stronger base in which to work with.

  146. #146
    mtbr member
    Reputation: Cat-man-do's Avatar
    Join Date
    May 2004
    Posts
    6,351
    Lucky me, I go out to buy a $60 GoPro mount from REI and it too won't work. Okay, I get to return it but somehow When I tried my ITUO mount again I got it to work. I figure me trying to force the mount onto the REI mount perhaps forced the prongs a big wider. ( Thank God! ) Now I have a stable mount. I'm getting a late start but I'm going to hop out to one of the local trails and hope like hell they aren't too muddy. If I see some dry trails It will be a stroke of luck.

  147. #147
    mtbr member
    Reputation:
    Join Date
    Jul 2013
    Posts
    751
    I think the mount was an after thought, it stayed secure though bounced on every bump. As i mentioned the screw head is too close to the case to tighten properly.

    Gary, the light tripped all of my batteries except the battery from the Nitefighter BT70

  148. #148
    mtbr member
    Reputation: Cat-man-do's Avatar
    Join Date
    May 2004
    Posts
    6,351
    Okay, just got back from my first trail ride. As expected I had to deal with some mud but nothing that caused me to have to bail. More or less just enough to be a nuisance though.
    I started out with the B3 on low with my XP3 on the lid standing by but not on. I noticed that I could see very well with the lamp on low. After a couple minutes I switch to medium and medium looked very nice. If I were to use this lamp regularly I'd likely be using medium and low a lot. High mode is quite awesome but on the trails I was riding there was no point in using it because of the tall grass on the section I started out on.

    Eventually I came to a section that looked to be good for a couple beam shots. The B3 in areas that are more open is very impressive, even on low. After I finished up the beam shots I did some quick ballpark measurements by pacing out the perceived distance throw. As I stood behind my bike I looked ahead to see what objects I could see in the distance and then paced out the distance. Easily throws to 150 ft. and then carries a bit more as the light fades into the distance. After I got back to the car I took a quick shot on the paved forest road. On flat surfaces like a road the beam pattern carries much farther ( more like 200-250 ft. )

    On the way back I took another trail that had less foliage but more technical features. On that trail I decided to use the high mode. Not that I needed more light but I wanted to see if the lamp was switching output. Sadly, I hate to say it but it was changing output and quite constantly. That said it wouldn't stay on medium but for a few seconds and then switch back...then after maybe twenty seconds or so back to medium. This continued the whole time I was using the high mode. It was a cool night too, upper 50's. Oddly, the lamp head really didn't get that hot the whole time so damn if I know why it keeps powering down.

    More on the beam tint; This time I got a very good look at the beam tint and Ledoman is right, this is definitely a cool white beam tint. On sections where I was going over a lot of rock I couldn't help but notice the loss in definition of the trail surface. Of course with the shear output or in lower modes this wasn't much of a problem. I'll keep my Gloworm NW X2 on the bars though, at least for the time being. That said a cool white tint works very well on wet dark colored trails

    Other things to note; Not more than 15 minutes into the ride right after I switched over to see what the high mode could do the lamp suddenly cut out. At the time I was on flat ground with lots of high grass and running only the B3. Thankfully I wasn't going very fast because of all the mud but it did give me a scare as I thought the lamp just died. Thankfully I discovered it was just the connector that came ( slightly ) loose. After that I sured up the connectors and I was fine after that. To be on the safe side I ran a torch on low for back up ....just in case.

    Now after about no more than a half hour into my ride the red light ( voltage indicator ) on the lamp came on. This happened last night too. Of course I have no idea what this means or how long the lamp will run while in the red. I just hate looking at a red light. FWIW I was using a 6000mAh Li-Po battery so I should of easily had enough power for a three hour ride. My total ride probably only lasted a little over an hour. Now here's where things get strange; When using some lamps that use voltage indicators there are usually three levels, Green, blue, and finally red. This lamp only goes from blue to red, there is no green. The battery I used has been on the charger for little over an hour so it's taking more juice than what I expected. More on that as I go along. ( edit: Battery took 2.25 hrs to recharge, not too bad but more than I expected )

    Below are some pictures. These are taken with my phone camera ( which sucks ) and uploaded to Photo Bucket. Photo Bucket has a "night enhancement feature which I tried. I used the enhanced photos because I felt it more accurately depicts what I was seeing. Both photos on high. Click on each to view each separately.

    Review: Zanflare B3 bicycle light-b3trail.jpgReview: Zanflare B3 bicycle light-b3road.jpg

  149. #149
    mtbr member
    Reputation: Cat-man-do's Avatar
    Join Date
    May 2004
    Posts
    6,351
    Quote Originally Posted by garrybunk View Post
    The GoPro mounting is something I didn't even evaluate. I believe the mount attached to the bottom of the light is aluminum.

    -Garry
    I think you're right. While fiddling with the lamp / mount I also noticed that the screws on the bottom were coming loose. Damn! Those screws ( which are hex ) are extremely small too. Luckily I have some small hex wretches but they might need some Loctite to keep them from coming loose again. Don't ask me why they chose such a small size screw. I think they are the smallest screws I've ever seen.

  150. #150
    mtbr member
    Reputation:
    Join Date
    Nov 2012
    Posts
    1,043
    Nice beamshots. That high output is certainly impressive.

    It's probably not getting hot on the outside, because it's not conducting the heat away from the inside. It's more like an oven that's baking the led driver.

    The thermal bounce from high to medium and back again would drive me absolutely nuts. I can't see why anyone would buy it in it's current state, it's a waste of money. Unfortunately, I think their design is set in stone.

  151. #151
    mtbr member
    Reputation: Cat-man-do's Avatar
    Join Date
    May 2004
    Posts
    6,351
    Quote Originally Posted by varider View Post
    ....The thermal bounce from high to medium and back again would drive me absolutely nuts. I can't see why anyone would buy it in it's current state, it's a waste of money. Unfortunately, I think their design is set in stone....
    In it's current state no way they should be trying to sell these to anybody. Somehow they have to at least do something about the premature / spastic power up-down issue. Move / change the sensor or change the circuit that's cutting the power. Then they need to rethink about providing a better bar / helmet mount set-up and then maybe someone will actually want one of these.

    While their at it, offer a NW version for those who want it.

  152. #152
    Tribe Member
    Reputation: yetirich's Avatar
    Join Date
    Jan 2004
    Posts
    873
    So these are on pre order at GB for $49 USD but they don't come with a battery. Also they have only sold $691 USD towards their goal of 20k USD with 13 days to go. Seems like this should be a hard pass on purchasing.

    The in depth reviews on this one have been awesome. Thanks for doing the hard work!
    Last edited by yetirich; 09-05-2017 at 09:25 PM.
    Proud Tribe member since 1992 - looking for better singletrack to be ridden year round

  153. #153
    mtbr member
    Reputation: MRMOLE's Avatar
    Join Date
    May 2011
    Posts
    2,871
    Cat
    Oddly, the lamp head really didn't get that hot the whole time so damn if I know why it keeps powering down.
    All that cooling surface area is doing its job and keeping the outside of the light cool enough, its that inadequate thermal path not allowing the excessive heat produced by the B3 because of the inefficient driver(s) and high current going into the light. Garry's thermal stability tests showed the thermal path would only support level 7 consistently with ambient temps just under 70F and a case temp near 140F. That would have seemed hot to you but your cooler ambient temps. and increased velocity over the test fan kept the outside cooler but unfortunately not the inside because of poor heat transfer.


    In it's current state no way they should be trying to sell these to anybody. Somehow they have to at least do something about the premature / spastic power up-down issue. Move / change the sensor or change the circuit that's cutting the power. Then they need to rethink about providing a better bar / helmet mount set-up and then maybe someone will actually want one of these.

    While their at it, offer a NW version for those who want it.
    Definitely agree! It will be interesting to see if they try. Would lower my opinion of Zanflare/Gearbest to the point where it would be hard for me to do any business with them in the future. If they do the right thing and delay release till they can get the lights problems fixed they need to take care of their crowdfunding/preorder customers too.

    In addition to "Cat's" recommended minimum changes I would add to the fix list: 1) Poor driver efficiency and related thermal and power consumption issues. 2) UI issues (inability to alter preset levels). 3) Poor mode button feel (noted by "OLD" on the BLF). TIR optics would make this light more useful allowing it to work effectively as a helmet light but as a bar light the reflectors seem to do a good job so not absolutely necessary to change (but preferable).
    Mole

  154. #154
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Thanks for your evaluation Cat! Those screws on the GoPro mount (if the same as the ones on the front bezel) should be
    a 1.5mm allen/hex key.

    I keep questioning whether I did something to mess up the thermistor, but then again GJHS has reported similar behavior.

    -Garry

  155. #155
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Quote Originally Posted by yetirich View Post
    So these are on pre order at GB for $49 USD but they don't come with a battery. Also they have only sold $691 USD towards their goal of 20k USD with 13 days to go. Seems like this should be a hard pass on purchasing.

    I'm in the market for a new light set for winter fatbike riding, commuting and was hoping this would be it. Any recommendations on a light now?

    The in depth reviews on this one have been awesome. Thanks for doing the hard work!
    Your questions would be better suited in a new thread where we could discuss without derailing this one. Post up your budget and requirements (throw, flood, runtime, etc.)

    -Garry

  156. #156
    ZFL
    ZFL is offline
    mtbr member
    Reputation: ZFL's Avatar
    Join Date
    Aug 2017
    Posts
    19
    We will improve some points for mass products

    1. The inside heat transmission.
    We will change the inside structures to increase the contact area of shell and heat dissipation panel. The heat will transmit to shell much faster and be emitted by fins.

    2. Good mount.
    We will improve the reliability the current mount. We recommend the users buy a extra metal mounts if he wants a better one. There are many Gopro mounts in the market.

    3. Switch.
    We will make the switches easier to use.
    The diameter of top switch is expanded from 7.5 mm to 10.5 mm, height from 2 mm to 3 mm.
    The diameter of rear switch is expanded from 7.5 mm to 12mm, height from 1 mm to 2 mm.

    4. Battery pack.
    We won't offer a battery pack as a standard accessory. We recommend the user buy a battery pack himself/herself and give him/her some recommend links. There are many good battery packs in the market.

    5. Good efficiency
    We tested and got the data 85%. It will increase a lot of cost if reaching 90%.

  157. #157
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Very glad to hear you will be addressing issues. Looking forward to these improvements.

    -Garry

  158. #158
    mtbr member
    Reputation: Cat-man-do's Avatar
    Join Date
    May 2004
    Posts
    6,351
    Quote Originally Posted by ZFL View Post
    We will improve some points for mass products

    1. The inside heat transmission.
    We will change the inside structures to increase the contact area of shell and heat dissipation panel. The heat will transmit to shell much faster and be emitted by fins.

    2. Good mount.
    We will improve the reliability the current mount. We recommend the users buy a extra metal mounts if he wants a better one. There are many Gopro mounts in the market.

    3. Switch.
    We will make the switches easier to use.
    The diameter of top switch is expanded from 7.5 mm to 10.5 mm, height from 2 mm to 3 mm.
    The diameter of rear switch is expanded from 7.5 mm to 12mm, height from 1 mm to 2 mm.

    4. Battery pack.
    We won't offer a battery pack as a standard accessory. We recommend the user buy a battery pack himself/herself and give him/her some recommend links. There are many good battery packs in the market.

    5. Good efficiency
    We tested and got the data 85%. It will increase a lot of cost if reaching 90%.
    ( Posted by the representative from Zanflare ^ )

    Nice to see some feedback from the manufacturer. If you're going to provide a better thermal path for the emitters that would be great but you still have to make sure that the lamp does not go into "Thermal step-down" mode too fast or that it bounce back and forth like it currently does. In it's current state it makes no sense to even use the three emitters on high. The lamp needs to be able to run at maximum output for speeds around 10mph @ temps near 80°F. for ( at least ) 10 minutes before stepping down. This I would consider to be an absolute "Minimal Standard" before I would even consider buying.

    Now, about the mounting issue; I think you need to take the mount issue more seriously. No one is implying that it needs a metal mount but it needs a mount that will prevent the lamp from shaking, vibrating, or moving too easily. I've seen many types of cheap hard plastic mounts that would do the trick and cost little more than the mount you current use. I recommend a mount with a pivoting clamp. Perhaps something like this...( see below )

    Selling this lamp with the current mount is like selling a high performance sports car with the cheapest tires you can find. Need I say, nobody does that if they truly want the product to sell.

    Review: Zanflare B3 bicycle light-61im3hnprxl._sl1200_.jpg

  159. #159
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    An update:

    So my BLF contact ImA4Wheelr is evaluating this light, specifically the driver. He is seeing the driver consuming 42 watts input (5.25A @ 8.0v) and outputting 33.58 watts for 80% efficiency which he says is typical of high current flashlight buck drivers in his experience. The output number is preliminary as the emitter leads were cracking away from the solder joint at the driver and need to be replaced.

    He's going to mount the emitters to an external heatsink, keep them cool, and confirm the thermal stepdown will not occur.

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  160. #160
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Today I've opened the light again and replaced thermistor with bigger one. I've used 10K 1% NTC thermistor. See red circle on the picture.
    Review: Zanflare B3 bicycle light-replaced-thermistor.jpg

    Now it behaves a little better. I've tested it only staticaly and when running on full power from cold light I've got few power offs of middle led and after that step down to mid mode. When started from already warmed light I've got just step down to mid mode.

    Another thing. I've bought NW version of Zanflare F1 flashlight (when it was in action). I suppose it has same NW 4500-5000K led that NW version of B3 would have so I've made tint comparation to the current B3 CW version. Here is how it looks:

    Review: Zanflare B3 bicycle light-tint-comparation-nw-f1.jpg
    CW B3 on the left and NW version of F1 on the right.

    I've made brightnes and beam of both to match with adjusting modes and distances. Tints on the picture are pretty much close to what my eyes can see on the white wall. You can notice CW B3 is not fully cool (plain) white as we have already stated before. I would expect NW version of B3 would be much more pleasant to the eyes and more compatible with lights such as BT21 and similar.
    Last edited by ledoman; 10-04-2017 at 02:07 AM.

  161. #161
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Sharing some of the information that BLF user ImA4Wheelr has sent me. He analyzed power consumption and produced this chart:



    (Yes, he is seeing more current pulled by the center emitter.) So efficiency isn't that bad at all on 2 LEDs High and 3 LEDs Medium; even on 3 LEDs High it isn't that bad.

    He also has the emitter pcb out of the light body and attached to a large block of aluminum. The thermal step down does not occur with a fan on. With no fan, the thermal step down does occur periodically. It was not expected that a fan would be necessary. Maybe they have the thermal step down set too low.

    He is finding the driver very difficult to work on with so many components on the opposite sides of the boards and the wires run through "through holes" where heating them up to remove will cause other components to heat up and come off. It's basically a "once and done" driver - one you assemble, but never disassemble. He is also struggling to find a better way to improve the thermal path, but still wonders if the improved thermal path will be enough for this light.

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  162. #162
    mtbr member
    Reputation: MRMOLE's Avatar
    Join Date
    May 2011
    Posts
    2,871
    Numbers on this light are interesting but just don't add up correctly. Using the lowest battery current draw @ max. output I've seen on this light (5.25A) it still requires over 50% more current than my Ituo XP3 to produce a similar amount of lumens with a similar 3 led layout. Hmmm!
    Mole

  163. #163
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Hmmm, my measurments of central led (posted on previous page) pretty much corresponds to measurments of left/right emmiter ImA4Wheelr got. I'm repeating so it's easier to compare. I've got same 10W power on highest mode.

    Low 0.48A @ 2,83V = 1.36W
    Mid 1.45A @ 3.11V = 4.51W
    High 3.02A @ 3.41V = 10.3W

    There might be some mistake in ImA4Wheelr measurments or it is just difference of preproduction sample. It could also be difference in the R016 resistor. Garry would you ask ImA4Wheelr the check those 3 resistors if he can?

    Quote Originally Posted by garrybunk View Post
    ... He is also struggling to find a better way to improve the thermal path, but still wonders if the improved thermal path will be enough for this light.
    I'm wondering too. It might work with better thermal path and enough air flow. Judging over the tumb at 3A current to each emitter I would say the light should be bit bigger, just few more mm to fit 20mm TIR lenses. There would be more room for the driver so it could run cooler and more efficient. It would look little bulky, though. This of course are directions for new version if any.
    Last edited by ledoman; 10-09-2017 at 02:11 AM.

  164. #164
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Inquiry sent to ImA4Wheelr with link to the above two posts. I too am baffled where the output is lost at.

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  165. #165
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Quote Originally Posted by ledoman View Post
    There might be some mistake in ImA4Wheelr measurments or it is just difference of preproduction sample. It could also be difference in the R016 resistor. Garry would you ask ImA4Wheelr the check those 3 resistors if he can?.
    ImA believes all three are R016 resistors, but will double check later tonight. He is curious as to other components on the driver (which are inaccessible without disassembling and ruining the driver) if they differ for the center LED. He will try to double-check his measurements as well. He has also commented that the thermal sag could be playing into the lower than expected output. I wouldn't have thought that would have kicked in so quickly from a cold start, but I don't know for certain. I measured 2,590 lumens at startup which should have been from a cold start on a fresh pack.

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  166. #166
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    ledoman, did you measure the current at the center emitter in 3 LED High mode? ImA says in 1 LED High mode it does measure 3 amps, it is in 3 LED High where it jumps higher.

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  167. #167
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Hmm no, I've measured only single led on all three modes. All my efficiency calculations assumed all leds runs at same power.

    Then this is even more wierd. The only thing I can think off is the obvious one. Both side leds have driver circuits on one PCB while central led has it on another PCB where the power comes in. Don't know if it's related, but looks suspicious.

  168. #168
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    ImA has joined MTBR so I don't have to be a middle man, but he can't post yet. He'll try again later.

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  169. #169
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Ok, good to know. Looking forward to read his opinions.

  170. #170
    mtbr member
    Reputation: MRMOLE's Avatar
    Join Date
    May 2011
    Posts
    2,871
    Quote Originally Posted by ledoman View Post
    Hmmm, my measurments of central led (posted on previous page) pretty much corresponds to measurments of left/right emmiter ImA4Wheelr got. I'm repeating so it's easier to compare. I've got same 10W power on highest mode.
    Sorry if I wasn't very clear in the wording of my post. I wasn't questioning that the test numbers didn't add up correctly. I was commenting that the efficiency ratings in the 80+% range didn't make sense considering the battery current draw/lumen output of the B3 and the fact that it requires over 50% more current to produce approx. the same lumen output as the two lights I have with a similar layout (XP3/XS).
    Mole

  171. #171
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    MOLE, According to ImA4Wheelr:
    Quote Originally Posted by ImA4Wheelr
    "Using my current measurements, output should be about 4000 before OTF losses with the V6 (bin). Cree data sheet indicates V5 is about 4.3% less output than V6. So lets say 3,828 for V5. Then 20% loss for optics (I don’t know if that is reasonable, but I have heard it before). So estimated 3,063 OTF."
    Those are simply based on current measurements and comparing to BLF user djozz's XP-L V6 emitter output tests. So yes, something is definitely wrong - perhaps it is the thermal sag playing in.

    ImA4Wheelr is still not able to post yet, so I will post over more information he has sent me:

    Quote Originally Posted by ImA4Wheelr
    I reviewed my testing and rechecked a few things. I’m confident in my reported measurements. I see nothing to indicate that I need to reperform the whole test again. Especially, when my measurements are similar to ledoman’s.

    I double checked the resistors. All three are labeled R086. I can’t make out most of the other components on the bottom buck driver. In center emitter only mode, the center emitter measures the close to or the same as the outside emitters at all three levels (L, M, and H). So it appears the MCU has basically 5 or 6 levels for the center emitter, but only 3 for the outside emitters.
    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  172. #172
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Hmm, hmmm, as you can see on my pictures (closeup in post #73) mine sample has R016 not R086. Is R086 mistake or true? Garry, I can't find any picture of yours driver where the resistors could be seen. Do we have different driver versions? Mine version can be seen on this picture.

  173. #173
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Response from ImA4Wheelr:

    Quote Originally Posted by gImA4Wheelr

    I don’t really have much to add beyond what you have said on my behalf. You may want to tell them that I cut battery pack lead out of the equation when measuring the driver. You can post my above picture to help show the difference (e.g. 12 gauge wire, power supply to measure input, etc). All you have to do is right click on the picture and then click on “Copy Image Location”. It appears that only added 1.2 amps of consumption, but I changed other factors at the same time.

    There can be several simultaneous contributing factors to the lumen loss:

    More than 20% optic loss. I have measured 7% gains from just switching to AR lens on several of my lights.

    Emitter bins not actually V5

    Thermal Loss as discussed (I may try to measure with the emitters well cooled and with and without optics, but it may take me some time to get around to that).






    I checked my own images and I never took a good photo of that side of the driver – DOH! I've asked ImA to confirm the resistor values.

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  174. #174
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Not sure why his image won't post correctly, but here is a link to it.

    (Perhaps the issue is the extremely long link text!)

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  175. #175
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Oh, another response already - yes a type-o, the resistors are all R016's.

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  176. #176
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Quote Originally Posted by ImA4Wheelr

    Oh, I don't think you have clearly communicated how bad the MCPCB is not thermally connected to the host. It actually floats and has no direct metal on metal contact other than some burs on the outside of the MCPCB touching the anodized intererior walls in spots. The shelf is .1mm away from the bottom of the MCPCB. Thermal paste can not make up for that and should only fill any micro voids.




    And as ledoman pointed out, the emitter shelf is under tension on both ends holding the back cover of the host in place. Here is the gap without the screws. With the screws in place, the gap is larger.




    NOTE: If images don't show, click them.

    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  177. #177
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Quote Originally Posted by ImA4Wheelr
    I just realized another difference in my measurements and the stock light. The driver in the host is not thermally connected to the host. So it can’t effectivly shed it’s own heat. It basically floats like the MCPCB does. In my test, there was a cooling fan blowing room temperature air over the driver. This definitely will make the driver more efficient, but I don’t know by how much.
    -Garry
    "My Bike Lights" Thread on BLF teardowns, measurements, and beamshots. Moving my photos, PM or post up if you can't see them.

  178. #178
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Thanks to both of you. Yes, driver is fixed only with 2 screws and no other direct contact to the housing. Shurely there is almost no heat tranfer away from the driver which means less efficiency.

    AFAIK they are going to redesign internals a bit to get better heat transfer from leds MCPCB. Not shure what that would be, we'll have to wait and see.

  179. #179
    mtbr member
    Reputation:
    Join Date
    Oct 2017
    Posts
    8
    ^
    Yay!, I can finally post. Hi guys. Really great collaboration you all have going here.

    I don't think I have much to add at this point. Garry asked me to analyze the driver. I have limited equipment and am no electronics expert, but I offered to do what I could. I intended to tear down the driver to consider the circuits and components, but it became very clear that I would ruin the driver in the process. After measuring it's performance, I see no reason to do so anyway. The performance outside of the light seems adequate to me. I could be wrong, but I don't think you can do much better moving this much current in such a small package.

    I think about 1.2 amps of input current needs to be added to my measurements to allow for the stock input lead and connector. I might try to verify that, but can't promise I'll get a chance to do that.

    Regarding where are the lumen losses is. I measured a tad over 3K lumens (near startup due to fiddling with modes) and 2,750 at 30 seconds in my integrated tube. Keep in mind this is after Garry had improved the thermal path. That is about what I would expect out of V5's OTF at that current. There is a light error in my OTF calculation above, but it's only about 4 lumens off. I don't know if my tube measures accurately. I feel it is ball park accurate, but that is base on my reference lights that may or not be accurately delivering to specs. I have more faith in Garry's measurements and his were consistent with another member's measurement.

    So assuming my measurements are wrong and there is lumen loss, I think that it would be due to any combination of the following:
    • Thermal issues. The light is a basket case of thermal issues as discussed above. Really needs some overhaulling in the design.
      • I don't know how fast the MCPCB heats up, but I bet it is very fast as it is quite light weight.
      • The driver is also not thermally connected. You would need very hardy ground planes with direct connections to the host to remove the heat. I really don't know if this is much of an actually issue though. I might try to test that by phsyically isolatating the driver from the LED so that I can try to measure how hot the inductors and diodes get. No promises there.

    • Inefficient optics. I'm not saying the optics are inefficient. Just that they could be.
    • LED's not actually V5 bin. I don't feel that ZanFlare would lie about that. I don't know if they test the binning. If they did, then they are probably real V5 bin.


    Regarding the very irritating step down/step up issue. It didn't manifest when the emitter mcpcb was kept cool. But it seemed they had to be kept too cool. I didn't measure temps, but touching the heat sink indicated that. So bad thermal path (certain of that) and maybe too low a step down temp threshold (not certain of that) may be the problem. Usually, the step down temp is hardware adjustable by changing resistor values in it's associated voltage divider. I don't know if that is the case in this light. I will try to see if I can see that part of the circuit, but I doubt I will be able to see in there good enough, let alone reach the voltage divider circuit resistors (if applicable to this driver).

    That said, this is a very compact light and 3 xpl's at those currents does not seem a viable. Maybe with more efficient emitters. That new Luxeon V emitter sounds promising.

    EDIT: Added content

  180. #180
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Hi ImA4Wheelr, finaly, nice to have you here. It looks new users have to be aprowed by someone or there is just time gap to avoid spamers.

    So you think for the power used the light host should be little bigger? I think so, too. And more efficient leds alone won't help as long as driver produce so much heat. Would be interesting to measure most heated elements if you can.

    BTW, do you think changing resistors from R016 to some higher value would help to run it cooler? Since those leds are voltage driven (in contrary to best practice current driven) what is impact of those resistors? In post #73 I've made brief sketch which might help you.

  181. #181
    mtbr member
    Reputation:
    Join Date
    Oct 2017
    Posts
    8
    ^
    Thank you for the welcome. I'm humbled by your questions because I think you have a deeper electronics knowledge base than me. Some of what I know, I learned from your various driver posts over the years.

    It does seem the light is too small as currently designed. I don't know how much heat the driver produces, but my sense was that it is very little in comparison to the emitters. I was directly touching the driver during testing to change modes. I don't recall ever feeling any heat coming off it. Granted, I was just holding the outside edges and pressing the switches. Also, there was air flowing over the driver. If I can, I will try to measure or at least get a better sense of how much heat it produces.

    Increasing resistance should reduce current to the emitters. I just looked at the data sheet for the buck converter. It says it's a Current Mode buck. It certainly behaved like one when testing. It provided the voltage the emitters demanded (e.g. current would hold steady, but voltage would vary). It does have a pin that using a voltage divider can chose between 5 different output voltages. I'm guessing that is to set a voltage output ceiling or to help the buck converter optimize performance. Just guessing there.

    Most voltage sense resister banks also limit output to the emitters as the current also flows through them. I've always seen them before the FET that feed negative to the emitter. So it is a bit different looking. But without seeing the whole circuit, I don't know what to think of it.

    EDIT: Added last paragraph and fixed a couple typos.

  182. #182
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    My knowledge came from the same source ie. watching/reading from others. ;-)

    Since you have driver open you could change one resistor to confirm it is only to regulate max. current. On the other hand how do they get modes if they set constant voltage ceiling (using voltage divider). The only way I can think of is PWM where each pulse is 100% peak, just it's length is different. It's wierd Zanflare representative said they are not using PWM (but he is not technichian).

  183. #183
    mtbr member
    Reputation:
    Join Date
    Oct 2017
    Posts
    8
    ^
    Sure, I can try it. The closest I have in resistor that can handle the wattage is R1. So I can stack 4 for .025 ohms or 5 for .02 ohms. I will need to make a new rig. I already disassembled the old one. It wouldn't have worked anyway because I designed that one to use the stock LED leads. That makes things too tight to swap out the voltage sense resistors. So may take me a few days. My evenings are pretty short these days.

    EDIT: added word "leads"

  184. #184
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Thanks, will wait. This should get us new experience regarding voltage driven drivers. I would go with 0.25Ohm or higher so the difference should be notable.

  185. #185
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Finally I'm not a middle-man! Glad to see you over here ImA! Thanks for all the help on this light!

    -Garry

  186. #186
    mtbr member
    Reputation:
    Join Date
    Oct 2017
    Posts
    8
    Thanks Garry. You're very welcome.

    Quote Originally Posted by ledoman View Post
    Thanks, will wait. This should get us new experience regarding voltage driven drivers. I would go with 0.25Ohm or higher so the difference should be notable.
    Wow, this thing is very different from what I expected. I tried to see as much of the circuit as I could by looking at traces and components and realized I was completely lost. Looked at the data sheet, but I can't glean from it how one goes about altering the current. It says this:

    "FR9809 is a constant-frequency current-mode step-down synchronous DC/DC converter."

    I take this to mean its a constant current buck. A table or chart in the data sheet indicates that the current is adjustable from 0 to 5 amps. I see were one can adjust the voltage it biases towards. For S&G's, I connected an electronic load to the left emitter leads. The unloaded output is 4.57v. I tried and it did indeed supply 3 amps at that voltage level. Nothing got hot on the driver, but I only ran it a couple minutes.

    The data sheet says that it monitors current through the inductor and when current gets to high, the buck chip drops voltage output until the desired current is reached. So it sounds like the inductor might be the key to altering current, but I'm just guessing at this point.

    I really don't want to hurt this driver. The light can potentially become a very nice light and I don't want to ruin it. At this point, I'm over my head. The data sheet doesn't have enough info for my limited knowledge base. I will try to see if there are any other reference materials out there that may help, but my time is pretty limited right now.

    I am curious though.

    EDIT: While squinting at the stuff between the 2 driver PCB's, I saw what appears to be a voltage divider for the tempature sensor. I'm not sure because I couldn't follow the traces. It had a resistor labeled "01C" I think going to ground and a resistor label "01B" going to a capacitor (IIRC).
    Last edited by ImA4Wheelr; 10-12-2017 at 07:40 AM.

  187. #187
    mtbr member
    Reputation:
    Join Date
    Oct 2017
    Posts
    8

    I think I see what is happening

    Studied the data sheet and driver pics that are in this thread. Not certain of this as I can't see what is connected to Pin FB and Pin SHDN/S, but I suspect what is going on is that Pin SHDN/S is used to turn on and off the F9809 to achieve the desired current. Pin SHDN/S is the bottom, right pin (next to bay R21) in ledoman's picture with the circuit notations in Post 73.

    Since the F9809 is designed to provide 5 amps of current, the MCU would turn it On 60% of the time and Off 40%. The Resistor at R21 appears to be a pull down resistor for Pin SHDN/S (similar to what would be used for a PWM feed to a FET gate).

    If that is the case, it appears the only hardware mod available is the large R016 resistors which appear to only be there to provide a static restriction of current to the LED. I don't think they are used to provide a voltage feedback function.

    This also is consistent with the Center Emitter driver appearing to be the same while delivering higher current in 3H mode. The outside drivers are turned on 60% of the time to get 3 amps. The Center is turned on 73% of the time to get 3.65 amps.

    I don't think changing resistance at R26 should cause any harm to the driver. No guarantees though. So I will try to change the resistance to see impact. Actually, I will await approval from Garry to do so as this is his light.
    Last edited by ImA4Wheelr; 10-12-2017 at 01:47 PM. Reason: fixed reference

  188. #188
    mtbr member
    Reputation:
    Join Date
    Oct 2017
    Posts
    8
    Hooked up an osciliscope to Pin SHDN/S. Definitely is receving a signal that changes with modes. This Pin is used to turn the F9809 on and off. So I feel more confident about my above theory on how this driver is regulating current. I don't know how to post pics here yet. Here are a couple links. I think the first is Low, second Medium, and third High mode.

    This can't be constant voltage. The driver pumps out the same current within each mode whether the voltage load is as high as 4.57v down to at least 3v. Probably would drive lower. I just didn't try. So it acts like a pretty good constant current power source. It feeds the targeted current and provides the voltage demanded by the load.

    It's probably me not knowing better, but this driver seems like a pretty cool design. It uses less components (In the buck sections), seems to make less heat, and is more efficient than the flashlight bucks I have used.

    https://lh3.googleusercontent.com/Je...w1920-h1080-no

    https://lh3.googleusercontent.com/Qt...w1920-h1080-no

    https://lh3.googleusercontent.com/7A-VQhco3-HA3WjLI7XRIYJ037rasFSm47KqIY2jQr-865OZRZWLiXFArW2mzW-2vO_S8lL-nY-jw4nMEPU_Kms6B-dr9FHR9h9KMRXJRiGPRYhdr-Cse6M5vEj7V09d_Wuu1bS69nRcickGBZ0A1S17Zp0G7iAWUC7l xrD6XGXSMVI9VCRPF-tYcKcxnwF6eGM1sXMQNIdya9zHFiTG6FjX3iYJazxlxPwKxSzI Zv1KIOQYlPlvFel5dZO4bbCXotir-INKyeoOU-9QtHQv5dIJc01jENX1NiKnY0RSbgVohgveZykyAhzOMlFN6pTx JRilTGlmtuQIszE9I3W9BX0U4vQgIzbXGCYeRrAII-yjDa32VO-IuUVomwe2QiJTrfHW686oVD2twOteNuhQqNKlugsgwA0h9pp_n pz4U_yoTufLiiwa8MqPRjU0_eQflQ_7fVJv8jFS_lzM_UAsN1g JWwn_fP9ZYgBC0gwCCl_3Up84V-gOH5SrRviNKoNirkP13YJyraVxjoXqQ7vWmlEf-ZtjUUKsS8pe3nEOYG9Vwb2nO_QWKXIcwYGbytlJZxe3ftEBALY X8gxUSHoU9WGFzcqUnkAgWZYchcsh6w=w1920-h1080-no
    Last edited by ImA4Wheelr; 10-12-2017 at 09:27 PM. Reason: Added content

  189. #189
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Interesting findings. To bad we don't have enough knowledge. It seems nice to you, but Archie who is EE said this "it's a kind of 'amateur' level of design" and he's surprised it's actualy used for driving leds.

    Probably we should get some info from manufacturer, but I doubt they are willing to do it since they think it's inovative. Not skilled to judge on that.

    Anyway, at least we can try to change R016 resistor and see what happens. Then we might suggest manufacturer to use more resistance to reduce the power and thus the heat which is bit on high side for this host.

  190. #190
    mtbr member
    Reputation:
    Join Date
    Oct 2017
    Posts
    8
    Hopefully Archie swings by and shows me the error(s) in my thinking. Until then, I have can only report what I know and stand by my statements.

    Regarding the R016's. I'm concerned about changing them. The F9809 is going to do its best to push it's designed current. So the LED will consume its voltage and the resister will consume its voltage. Obviously, increasing resistance in the path to the emitter enough will restrict current at some point. But in the process the driver will lose efficiency. The Resister will consume additional power and create more heat. The F9809 will work harder and make more heat. Something might get stressed and die prematurely. It would seem the best approach would be to alter the code in the MCU. I don't know why the R016's are there. They seem unnecessary. Maybe they help tame spikes or something. The not knowing is what give me pause.

    For already made lights, I think the best approach would be to fix the thermal issues and replace the emitters with much more efficient ones (when available).

    On a separate note. This driver has been trashed a good bit in this thread. ZanFlare has acknowledged most of flaws with the light and says they will work to correct them. They have pushed back on the driver criticisms. I'm my opinion (which should not be given much weight), I think they were right to push back.

  191. #191
    mtbr member
    Reputation:
    Join Date
    Nov 2013
    Posts
    24
    I've been watching the gearbest page on this light. Today they changed the ship date from sometime in November to "Oct 17 - Oct 20" I wonder if they fixed any of the thermal path stuff?

  192. #192
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    Sorry, I was unable to check email all day. Go ahead and do what you think is worth doing in testing this driver.

    -Garry

  193. #193
    mtbr member
    Reputation:
    Join Date
    Oct 2017
    Posts
    8
    ^
    I would like to. But based on my current understanding of how the driver works, I would need to set up a whole new test rig in order to gauge the impact. That takes time that I don't have right now.

    What I expect would happen is that the the driver would just use more power and still feed 3/3.65 amps. Not until the resistance is high enough that it combined with the emitter consumes more than 4.57v, will current finally decrease. I say this because I connected a 4.57 volt load to the driver and it was able to drive the targeted currents. This is pretty much equals to adding resistance to that point in the PCB So the result will be:

    • Additional resistance in series with the LED's which will consume watts and generate heat. This means less regulated run time.


    • The F9809's having to work harder because they would have to provide more watts. This will create more heat and reduce run time. Will probably be stressful for them too. Meaning, possibly shorter life span.

  194. #194
    mtbr member
    Reputation:
    Join Date
    Feb 2016
    Posts
    710
    Review: Zanflare B3 bicycle light-9a0c9b3e-a49a-4050-8707-305780cb0da1.jpg The postman brought me a present today. B3 in neutral white 4500-5000k.

  195. #195
    mtbr member
    Reputation: MRMOLE's Avatar
    Join Date
    May 2011
    Posts
    2,871
    Quote Originally Posted by Dirt Road View Post
    Click image for larger version. 

Name:	9A0C9B3E-A49A-4050-8707-305780CB0DA1.jpg 
Views:	90 
Size:	79.8 KB 
ID:	1172645 The postman brought me a present today. B3 in neutral white 4500-5000k.
    Definitely keep us informed! Curious what they fixed if anything. What are you going to use for a battery?
    Mole

  196. #196
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Interesting, how did you manage to get it? I've been told there are no further plans to do NW version of B3.

  197. #197
    mtbr member
    Reputation:
    Join Date
    Feb 2016
    Posts
    710
    I got it from Zanflare’s crowd funding effort (yeah I’m a trusting soul, live on gearbest, banggood, AliExpress and am yet to get burnt). I’d say my b3 has an awesome tint prolly closer to 4000k by white wall hunting. Hope the Ituo 4cell pack from my xp3 is up to the task. I doubt there’s any upgrades to my example, still has the op reflectors, no optics, mistake imo. It’s a heavy lighthead prolly throw it on the bars if the lake effect snow bands ever stop. KD 2cell pack is way outgunned, battery indicator turned red in a few minutes.

  198. #198
    mtbr member
    Reputation:
    Join Date
    Apr 2012
    Posts
    1,850
    Can you make some pictures and meassure size if the buttons. The newer version is ought to have bigger buttons.

  199. #199
    mtbr member
    Reputation: garrybunk's Avatar
    Join Date
    Feb 2014
    Posts
    1,850
    It would be nice to hear if it still has the issue of stepping down & back up constantly when hot.

    -Garry

  200. #200
    mtbr member
    Reputation:
    Join Date
    Feb 2016
    Posts
    710
    I’m out but when I’m home, I’ll charge up batteries and giving a go. And pics as well. My guess is this is the original design that zanflare unloaded. The heft of this lighthead should indicate some heat shedding capability.... or not.

Page 1 of 2 12 LastLast

Similar Threads

  1. light-bicycle or carbon-bicycle
    By superflyAL in forum 29er Components
    Replies: 9
    Last Post: 10-18-2019, 09:52 AM
  2. Replies: 134
    Last Post: 08-22-2017, 01:53 PM
  3. Light Bicycle, Carbon Bicycle or Nextie
    By dimitrin in forum Wheels and Tires
    Replies: 14
    Last Post: 01-05-2015, 08:02 PM
  4. Replies: 17
    Last Post: 10-20-2014, 04:11 PM
  5. Review: Niteye B20 bicycle light
    By h79 in forum Lights and Night Riding
    Replies: 3
    Last Post: 08-03-2012, 01:27 AM

Members who have read this thread: 1

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  

THE SITE

ABOUT MTBR

VISIT US AT

© Copyright 2020 VerticalScope Inc. All rights reserved.