Thanx for the pics Trout.
Waiting patiently for the beamshots.......I am thinking the XML will have a little bit more of a flood than the P7.
Also, the ground wire soldered to the B- spot in the last pic kinda scares me.
Look, whatever happens, don't fight the mountain.
Looks like an incremental improvement in the driver design.
R7 & R6 look like current sensing resistors (last photo). Still too small, they are 0805 smt packages so will get pretty darn hot. But having them both be .15 ohms will probably give an LED current of 2.67 amps, which is a bit more than the original bastid managed.
The inductor, while not shielded is at least mounter on the PCB and not dangling in space.
C1 looks to be a halfway decent tantalum capacitor, but I cant make out what voltage it's rated for. Would be nice if they put in something rated at 16 volts or more. It looks like a 1210 package, so it might be rated for 16 or even 25 volts.
D3 looks like a SS34, which should be able to handle around 3 amps, up to 40 volts. It will probably still get pretty warm (not enuff copper around it to spread out the heat), but this is the first bastid-like light I've seen with a diode that had halfway decent specs.
IC2 is probably the driver IC. Looks like it is labeled 5340, which may mean it's an improved version of the SM5241 used in the bastid, or some completely different chip. I can't find a data sheet for it.
The button side of the PCB looks interesting (next to last photo). It has about the right number of components that it may be using the microcontroller (8 pin soic to left of button) do figure out the battery voltage. The original bastid used a voltage level detection chip which is why the battery level LED stayed on all the time. Is that still the case Troutie? The traces I can make out on the PCB make it look like the microcontroller might be a PIC (easiest way to tell is pin 8 will be ground and pin 1 will be around 5 volts). If the micro is a PIC I may break down and buy one of these since it will be pretty simple to change the programming.
The quality looks a little bit better than the first bastid I tore apart, though the B- wire looks like it had a bit much stripped off it before being soldered down.
Troutie, is the user interface any better?
Is it wrong to say I prefer the shape of the original? There I've said it.
lol @ MrLee
..and nice post Mark, very informative!!
Originally Posted by troutie-mtb
This looks much like the difference between my cheap P7 and XML torches.
I think the extra OP needed on the P7 really hurts the beam.
Originally Posted by Cat-man-do
The tree was about 50 yards away
the current was what was coming from the power supply at 8 volts not what was coming from the driver to the leds Sorry I should have worded that bit better.
yes it blows the old P7 bastid into the dust for throw and looks a whole lot better in real life than on the pics .
I like its beam and so will most users I dont like the user interface and am betting it is why there are reports of switches failing as from hi back to hi is 4 presses on the old bastid would much prefer something like Georges Trio mode. and no flash
Nice beam shots. The beam from the new nameless bastid looks very nice.
Any chance you can buy just the light heads? These look like they'd be pretty easy to modify if the micro is something standard. If you see markings like you find at www.microchip.com, then it is probably a PIC. It would be pretty easy to gin up a new PIC with a better user interface.
If the markings aren't clear (sanded off), when you get a chance, look at the voltage across pins 1 and 8 on the 8 pin soic on the button side of the PCB. If pin 1 (the one closest to the dimple on the soic, on the lower left corner of the soic in your picture) is between +5 and +3.3 volts and pin 8 (upper left corner) is ground, the chip is probably a microchip PIC.
Assuming the current sense voltage on the driver IC is 0.2 volts (what it was in the original bastid) then a T6 XML would put out around 840 lumens, or maybe 700 lumens after optical losses.
I am a bit disturbed by the current values you measured. The initial reading of 1.5 amps implies an efficiency of around 75%. 75% is pretty dismal efficiency, the electronics board will be dissipating around 3 watts. I would suggest heatsinking the electronics is you can.
The reading the light settles down at: 1.38 amps, would imply a slightly better efficiency of maybe 82%.
Originally Posted by email@example.com
Took a bit of trickery to read the chip ended up brushing a bit of the wifes makeup powder on a bit like fingerprint powder .
top line 12F675
next line 1/SN103S
bottom H Q ( or could be 0) B
put the multimeter on the led wire and it read 2860 ma which settled down to 2700ma on hi as it warmed up
900 ma on med
260 ma on low
will see if the heads are available though the battery pack seem to be OK and in a rubber case
12F675 is a pretty standard 8 pin microchip PIC. 1K worth of instruction space (1.75K bytes of ROM if you prefer), 128 bytes of eeprom (can be used to store last mode of light), and a whooping 64 bytes of memory. $0.90 in quantity 25 from digikey.com for the soic version.
I might be willing to trade some custom programming for a couple of light heads and the email address of your source in China . A simple user interface with 2 or 3 modes, long push to turn off, and ability to remember last mode light was in would be pretty easy to program.
With a slightly more expensive 8 pin PIC ($1.07 in quantity 25) it might be feasible to add features like user programmable light levels, optional flashing modes (done in a way that you don't normally have to cycle thru them when all you want is a light), user programmable low battery levels (handy for using light with a 9.6V NiMh battery or even a 14.8V LiIon battery for example), and other such fluff.
What's the quality of the battery packs? Any chance they are more waterproof than the original bastid?
Does the charger look decent, or more like the original bastid charger that got so hot I was sure it was going to catch fire?
Last edited by firstname.lastname@example.org; 03-20-2011 at 08:38 AM.
Hmm, wonder if you could shove an LFlex in there (or get a new seat machined up). If you can get it lighthead only, swap out the driver, and hook up a nice little 1s2p 18650 battery... this might be a great little semi-custom job.
Thanks for the teardown, It still looks to be low quality (although better than the P7)
The clonewars continue...
The heatsink pill doesn't appear to have any tighter or larger contact area with the body that the old one, though the light being longer, it has a bit more air interface once the heat gets to the body. Unless cleaaned, no AA to help, either. So is the thermal path mostly unchanged as it appears?
Just saw it on DX, I'm tented...
Anybody know anything about the new "waterproof" battery with the MS clones?
Originally Posted by Ofroad'bent
yes now know lots
did a runtime test last night they say it is a 4.4 ah pack
full to empty on hi. they say 3 hours. I got 2 hours 35 mins
did the same with one of Smudges batteries and got 3 hours 35 mins. his was using samsung 2.6 ah cells so 5.2 ah
they are in a rubber sleeve which seems not too dreadfull .
Troutie was kind enuff to forward contact info to me and I wound up buying 3 lightheads to see if these things are any good (Thanks Chris). The pictures he took of the electronics didn't look too bad.
My short review of the lightheads I got is: save your money up and buy one of Troutie's lights (or one of mine ), or build your own. Any of those choices will be better quality than these lights. OK, so I may be a bit biased, since I sort of compete against these guys.
Here's the longer boring review details:
I bought 3 lightheads from a company named *st*r (will reveal name later, but if that pattern matches somebody you are thinking of doing a big buy from, please PM me).
I implied that I would like to buy a lot of these. I wasn't lying. A lot of people come to me wanting to buy lower powered lights in the 400 to 1000 lumen range. I make some small ones in that range, but with the economics of building 10 to 20 lights at a time being what they are, I can't compete with the price of the Magicshine/Bastid lights. I'm probably going to stop making smaller lights. The amount of work I put into them means I make something like a dollar an hour building them
If the quality of the lightheads had been halfway decent, I was going to replace the PIC on the back of the driver with one that could sense temperature and had enuff ROM for a decent user interface. With a bit of thermal epoxy and a way to regulate the LEDs output if it got too hot, I thought I might be able to sell these as low end lights and offer a longer warranty than most US distributors. I also figured selling them with NiMh batteries would get around some folks fears of cheap Chinese LiIon Batteries. And if somebody wanted LiIon batteries, they could always buy some of the "new and improved" packs avaialable from various distributors.
I'm not sure advertising them as "Cheap lights made in China, but re-engineered in America to be more reliable and less explosive" would have been a viable marketing ploy, but it might have worked if the electronics were well made.
When I opened up the lightheads my first thought was "Wow does the quality of the electronics suck." All 3 had problems like: cracked inductors, badly soldered parts, etc. On the board with the cracked inductor, I was able to pull the inductor off with my bare hands. I'm not superhuman , the inductor was never soldered down, it was being held in place by dried solder paste. I'm pretty sure this means you could buy one of these lightheads, test it on your bench, then have it fail while out on a ride because the solder paste eventually let go.
Pretty disappointing for samples sent to a prospective distributor. They all worked, but one had annoying flickering on the low setting.If you buy any of these, you pretty much have to dis-assemble each one to see if it is any good. It really is a shame that the driver board appears to be so shoddily constructed. The design isn't nearly as bad as the original bastid.
The drivers at least use two 0.15 ohm 0805 sized resistors (R6 and R7) for current sensing which is better than the original bastid. The current running thru each resistors is about 1.333 amps (the LED is driven at 2.67 amps, the bastid drove the LED at about 2.4 amps), so the resistors must dissipate about 0.2666 watts (watts = (I * I) * R). Thats kind of pushing it, but you can find 0805 resistors rated to handle up to 1/3 of a watt. 1/3 watt resistors cost around $0.50 cents each. 1/4 watt resistors cost maybe $0.15. 1/8 watt resistors cost about $0.05. My guess would be they used the cheapest ones they could find, but there is no way to tell just looking at the parts.
They use a halfway decent diode, a SS34. That can handle 3 amps at 40 volts. I'd probably have used a 4 amp 20 volts diode like the MBRS320T3G. I like my diodes to handle more current than I think they'll see. The Magicshine/Bastid uses something undersized like an SS22.
The input capacitor (C1) appears to be a 4.7 uF tantalum capacitor. I can't find who made it, so it's hard to determine it's maximum rated voltage, but I have found 10 and 20 volt capacitors that same size. My guess is it's a 10 volt cap.
The inductors (coils) are nice looking, at least the ones that aren't cracked.
The LED driver chip, labeled 5341D is almost surlely some variation on the SM5241, the one used in the bastid. The pinouts appear to be the same. I'ts probably got an input voltage range up to 36 volts. I googled for SM5341D but couldn't find a data sheet.
The MOSFET (an IRF9410 8 pin soic) looks pretty decent. I hung a scope on the gate of the MOSFET on the driver board that seemed to work well. It seems to be switching the MOSFET pretty cleanly, so the 5341D chip does a decent job of driving the MOSFET.
The 12F675 PIC (near the button) is being used to light the voltage LEDs, do PWM (pulse width modulation) to set the light level, and monitor the input voltage. Too bad the user interface sucks. As you push and release the button you get: high, medium, low, flash, off, high, medium, etc. I can't figure out why most of these cheap lights make you go thru a flasher mode when switching thru light levels. The 12F675 doesn't have a hardware PWM, so maybe the user interface (UI) sucks because the Chinese firmware engineers spent too much ROM figuring out when to turn the PWM output on or off and didn't have enuff room for a decent UI. They could have bought a slightly more expensive chip and fixed that problem.
I'll probably put my sample lights back together with a 4 amp diode, a higher voltage input cap and some physically larger current sense resistors with a higher resistance (probably 0.2 ohms in a 1206 package). I figure you can run the xml at 2 amps and still get more light out of it than the bastid, you get longer battery life and hopefully the current sense resistors won't burn up. With a higher voltage input cap, you can probably run these lights on anything up to a 14.8 volt battery. I have a few friends running homebrewed four 18650 cells in series battery packs, who need a cheap backup light. I'll probably sell a couple to them and see how they like them.
I might keep one for me. As Troutie points out the lights have a very nice beam. Too bad the rest of the light sucks.
I'll try to post some pictures later this week.
Last edited by email@example.com; 04-24-2011 at 09:44 AM.
thanks for the detailed review.
most of them still think, they could squeeze by with one light.
and anything new is cool, while 2x P7 is still better then 1 xml.
Nothing really new, since we still need
- better drivers
- better housings
- better battery protection circuit
- better batteries
- better battery housings
- better chargers
and to power 3+ LED's, everything needs to be in place.
Still looking to design everything from scratch.
By end of this year, there is going to be a lot more choices in the 2-3k lumen range,
at least for the lights, the Battery pack availability is still needs to catchup .
Since those things , going to need power.
Would need lots of input and recommendations. I'am good with pcb design, not that good with circuits. The hardest part, like pulling teeth, is getting parts, some stuff, like the crappy zetex, can't even get sample parts under 100 pieces, the real good stuff, they want to sell you 3000. As a little Garage hobby guy, I'am broke, don't have a CNC ether , argh.
Well, if anybody has a laundry list, of the 3 Stooges , Good, Bad, Ugly, + availability,
maybe we could start sorting things out.
cheers, Rob -> contact me for input , feed me
thanks troutie for the beamshots.
Originally Posted by troutie-mtb
the bastid's throw looks impressive, although the cri seems to be awful.
could you commend on the throw of the bastid compared to other optics. (laura, regina ..)
Originally Posted by reinholdus
Yes it has a better throw than Laura / Regina sharp spill cutoff like regina .
very cold white led no indication of what bin .
Good evaluation Mark Thanks .
Like you I am staying away from selling crappy chinese imports or even making similar output lights to sell .( its fraught with all sort of issues look at Geomans problem )
I managed to sell one of the samples I got for the price I paid and will keep the other as a point of reference in my archive of lights .
^^The thistle in the posey patch. what an archive...