Quad seoul P4- Mtbr.com
Results 1 to 11 of 11

Thread: Quad seoul P4

  1. #1
    mtbr member
    Reputation: zen bicycle's Avatar
    Join Date
    Mar 2007
    Posts
    463

    Quad seoul P4

    Would you run them all in series or in a series parallel configuration. Both would be using a buck puck. Both are possible, just looking for suggestions and recommendations. If I run in series the vf would be right at 14v for a 14v battery with the loss in efficency so it should have a very long run time, but in series parallel I could use a smaller battery pack since the vf would be smaller and I really don't need a run time over 3 hours

    Thanks

  2. #2
    mtbr member
    Reputation:
    Join Date
    Jan 2004
    Posts
    301
    I'd suggest using a boost converter instead of a buck and running them in series. If you run them in series with a nominal 14V battery (I presume actually 14.4) then it's almost a waste of time having the buckpuck at all, since they'll almost never be in regulation.

    I'm sure some would recommend series/parallel, but personally I'm not keen as even with LEDs from the same strip you'll get a lot of variation in Vf (unless you're lucky), so you'll tend to get one set taking a lot more current than the other. That and you'll only be able to run an average 500mA per LED since Buckpucks max out at 1000mA.

  3. #3
    mtbr member
    Reputation: zen bicycle's Avatar
    Join Date
    Mar 2007
    Posts
    463
    I hadn't thought of the boost option

    Thanks

  4. #4
    mtbr member
    Reputation:
    Join Date
    Sep 2005
    Posts
    100
    Quote Originally Posted by zen bicycle
    Would you run them all in series or in a series parallel configuration. Both would be using a buck puck. Both are possible, just looking for suggestions and recommendations. If I run in series the vf would be right at 14v for a 14v battery with the loss in efficency so it should have a very long run time, but in series parallel I could use a smaller battery pack since the vf would be smaller and I really don't need a run time over 3 hours

    Thanks
    Hi,

    I have made myself a 4 Seoul P4 light with L2 optics. The driver I am using is Fatman boost from www.taskled.com powered by 9.6V, 4.5Ah battery. The battery is made of 8 4.5Ah NiMH C type cells in 1/2 litter watter bottle.

    What you have to keep in mind is that both NiMH and Li-Ion batteries have quite big voltage deviations from the nominal voltage. I.e. 14.8V Li-Ion battery pack will give you 16.8V initial output when freshly charged and will slowly drain down to 12V with protection kicking in (you must use some kind of protection for Li-Ion packs). The same logic applies for NiMH packs with the individual cell voltage between 0.9V discharged and 1.35V charged.

    If you would need to power 4 Seoul P4s at 1A the Vf 14.1V. Add 1.1 voltage drop for the driver (bFlex in this example) and your battery should provide at least 15.2V. Make your calculations together with the above paragraph and you can have a decision.

    One more thing. Do not try to run the power LEDs in parallel, as they have different internal resistance one to another. This will make one LED to get more current than the other and you have a pretty good chance of frying it...

    Hope this helps.
    Stefan

  5. #5
    www.hahntronix.com
    Reputation: mhahn@hvc.rr.com's Avatar
    Join Date
    Mar 2007
    Posts
    252

    Quad seoul P4

    I'm always kind of partial to a buck configuration. Buck controllers tend to be much lighter (smaller, cheaper inductors) and way more effiecient (90% vs 75%) than Boost controllers. Though with 4 white leds in series you'd need a buncha voltz. Probably at least 19.2 volts.

    Why not consider a 9.6 V battery and 4 seperate buck controllers driving 4 seperate leds (or maybe 2 controllers driving 2 pairs of leds). The problem with 4 leds in series if one blows you're ridin' home in the dark. Redundancy is a good thing..

    I've been experimentin' with national's lm3404. The circuit is pretty simple, you need an lm3404, a diode, a small inductor, a sense resitor, another reststor and a small cap, plus your led. If you ignore all the math they throw at you in the data sheet and design a driver around the upper current limit sense voltage (.3 volts) and can live with the fact your led will be experience a triangle wave shaped current between max and 66% of max (your average comes out about 83% of max) you wind up with a very robust led driver (runs from 8 volts up to about 80) that also has a regulated 7 volt output (add a cheap voltage regulator and you can have a microcontroller drive the duty cycle of the led). The chip works just fine with luxeons and crees. Max (average) current is around .850 mAmps.

    I'm sure buying someone elses driver puck is simpler, but soldering is big fun.

    Mark

  6. #6
    I ride a Swarf
    Reputation: Stuart B's Avatar
    Join Date
    Mar 2005
    Posts
    1,314
    I would run 2x2 and run 2 buck controllers. The redundancey is a vey good point. I know at races that I have been flying down a section I wasn't fully familar with and it would really hurt if my light failed hehe. ! reason why I have bar and helmet lights offroad.

    2 Seould is going to be plenty for non technical sections too, saving battery juice.

    Stu
    What exactly is a rigid hard tail?

  7. #7
    mtbr member
    Reputation:
    Join Date
    Jan 2004
    Posts
    301
    Quote Originally Posted by [email protected]
    I'm always kind of partial to a buck configuration. Buck controllers tend to be much lighter (smaller, cheaper inductors) and way more effiecient (90% vs 75%) than Boost controllers.
    I don't know what boost converters you've been looking at, but the Fatman (which is what most of us using boost converters have) is typically 90-95% efficient, and uses just as small an inductor as a buck converter. In fact there's absolutely no reason why a boost should need a bigger inductor, since the sizing of that is all down to the required current, the acceptable ripple and the converter operation frequency, none of which factors are any different for buck or boost. Meanwhile this poor efficiency thing seems to be a popular urban myth. Yes in theory a boost is slightly (~1-2%) less efficient than a buck due to the loss in the Schokty diode, but you really won't notice that (and anyway differences between different designs and different chips are typically more than that anyway).

    Not just theory BTW - I've measured the input and output on my Fatman with my setup, and it is running at around 95% efficiency for most of my runtime.

    The problem with 4 leds in series if one blows you're ridin' home in the dark.
    I've never heard of an LED "blowing" (unless badly abused) so you're attempting to solve a non-existent problem.
    Last edited by chrism; 05-06-2007 at 04:56 PM.

  8. #8
    www.hahntronix.com
    Reputation: mhahn@hvc.rr.com's Avatar
    Join Date
    Mar 2007
    Posts
    252
    Buck converters use smaller inductors because they require less peak current (saturation current) thru the inductor. Look at it from a power point of view. A buck converter driving a load that needs 4 volts at 1 amp current from a source that provides 8 volts will need an inductor that can pass a 1 amp peak current 50% of the time. So average current (roughly rms current) is about 1/2 amp. 1/2 amp * 8 volts = 1 amp * 4 volts. A boost converter driving the same load from 2 volts will need an inductor that can handle a 4 amp peak current 50% of the time. So average current is 2 amps. 2 amps * 2 volts = 1 amp * 4 volts. An inductor with a saturation current of 4 volts will be quite a bit heavier and larger than an inductor with a saturation current of 1 amp.

    As to efficiency, a boost converter typically has 2 current sense resistors. One to measure the current thru the inductor while "charging" it up (to prevent overcharging when the input voltage changes suddenly). And one to measure current thru the led while "discharging" the inductor. So you have 2 places that power is wasted (the sensing resistors have a current thru them that produces a voltage, hence some power is spent measuring currents).

    And both topologies have some kind of diode-like switch. A synchronous switch (typically a mosfet that conducts current at just the right time) can be more efficient than a diode because the switch has a lower voltage drop. So it is possible that you can have a boost converter with very small sense resistors, and synchronous switching that is more efficient than a buck converter using larger a karger sense resistor and a diode. But if you are comparing a buck converter with a boost converter, and both use the same technology (same type of switch, same size current sense resistors), the buck will waste less power.

    I'd be very suprised if your fatman is really hitting 95% efficiency. Unless it uses synchronous switching and drives a long string of leds. If it drives a single diode and uses a diode as a switch, it won't be that efficient.

    Mark

  9. #9
    mtbr member
    Reputation:
    Join Date
    Jan 2004
    Posts
    301
    Yes OK, the current in the inductor is less with buck - however your calcs aren't exactly correct. Since with the standard buck topology the inductor is always inline with the LED, I'm not quite sure how a 50% duty cycle of 1A manages to pass 1A through the LED! You're missing the fact that when it is disconnected from the input power supply it is still driving current via the diode. Therefore the mean current through the inductor is actually 1A, with a peak a bit higher than that dependent on the frequency and the inductor value. Meanwhile for the boost in your example, whilst the mean current is 2A (you got that bit right at least), that doesn't mean the peak is 4A, since that will again depend on the inductor value and the frequency of operation - in fact as I mentioned above, the ratio of peak to mean is exactly the same as a similar buck converter, so the difference for your example is only a factor of 2, not the 4 you're claiming.

    Not sure what your point is with the sense resistors - you do know that a buck also has 2 current sense resistors for the same reasons? I suggest you check the datasheet for the LM3404 you're advocating!

    Anyway as I said, I have measured mine and it operates around the 95% efficiency mark with my setup which is 3 LEDs in series running off 2 series LiIon cells - so not that different to the original question. A normal Schottky diode, rather than a synchronous setup though. Good point about it being more efficient running more LEDs in series - remind me again why you were advocating 4 separate buck converters driving the LEDs separately (where you are incurring 4 times the losses, as most of the losses in the sense resistors, the diodes and the switching of the FET are the same whatever the output voltage)? In fact I'd be extremely surprised if such a setup was more efficient than a well designed system with a boost converter.

  10. #10
    mtbr member
    Reputation: zen bicycle's Avatar
    Join Date
    Mar 2007
    Posts
    463
    sstephanov what type of runtime are you getting from that setup, and are you having any problems with thermal managment?

  11. #11
    mtbr member
    Reputation:
    Join Date
    Sep 2005
    Posts
    100
    Quote Originally Posted by zen bicycle
    sstephanov what type of runtime are you getting from that setup, and are you having any problems with thermal managment?
    I am getting around 2.5 hours of runtime out of a full battery running at 1A.

    As for the thermal management, the 4 Seoul P4 stars are glued to a old Celeron heatsink (5x5x5cm) with Arctic Silver Thermal Epoxy. When the light is stationary the heatsink reaches about 75 degrees Celsius, but when on the bike or a small fan is providing airflow, it does not get above 30 - 35.

    Stefan

Members who have read this thread: 0

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.