LightMalls BLF Custom Circuit Board (Project Fails Seeking New Vendor [relic38/FlashPilot Taking Over Idea])

Bill Zeng sent me an email asking for specs on a Custom Build Circuit Board for BLF. Here is your chance, give me the specs…

So far this:

Hi Flomotion
tomorrow i will get 2 samples driver
at the mode: 100% - 35% - 10%, working current at 5A(the 3.7V battery must be good),3.7~4.2V
it can work on U3 and XM-L2 led emitter too
also please let me know what circuit board you need
i will make them separately for you
yesterday,i test the board at factory,i installed it with a U2 led emitter,and tested the lux,it is about 1000-Lumen(real,but not the max lumens)
for the sample driver,i want to use it on C12 and C8 flashlight,then i will check them
also i think this board can not be used at 501B and 502B,as its current is too much,5A,it is too much for the switch and others,i think if this board
work at the 501B,it will burn,but C8 and C12 will be no problem,right?

best regards
bill zeng

Any information about PWM?

17mm or 20mm? 20mm would be very useful to me. Good mode spread too. :steve:

Sounds nice. It’s going in a C8 so it must be 17mm. I think the low should be lower, more like 5%. A moonlight mode would be nice too.
Finally, make sure the PWM is 4kHz or higher.

17mm and please, lower the low.

What kind of driver? Buck, boost, linear?

What they said.

What he said. Oh yeah…17mm. The perfect driver if it has all those features. Amperage between 4A & 5A would be good.

Ok, we have something going here, it is very, very, important to get across to the Chinese engineers, what we want. You may have to be rude, and to the point. They are going to design a driver at our specs, so get it right, or whine later. They are reading every post, so put some muscle into this, it has to be worth it for them to recover the development costs, the more ideas the better, if you have technical knowledge about components that you want to see on this board, say something, that is thermal sensors, software/hardware programmable voltage/amps, audio warnings indicators, i/o I don’t know if it is economically feasible to get too many bells and whistles, but this is our chance to one on one with Chinese engineers…so, go for it…

17 mm NOT 20 mm

4 modes: moonlight; 0.1, low; 1, med; 30, high; 100

Max current; 4A

Buck driver; max input voltage 8.4 VDC minimum regulated voltage; 3.6 VDC

PWM; at least 3KHz, higher is better

Texas instruments has some chips that might be the right ones. But IIRC those dont go much over 3A. If we want to stay in the 17mm size (P60 and most small flashlights), we have to keep it rather simple. Of course a 5A buck/boost circiut for 3-4.35V is possible, but that needs more components. I would be very happy with perfectly regulated 1.5A each for 3-4 Nichias (about 3.7Vf at 1.5A) and 3A for the XML2 (3.7Vf at 3A on copper PCB).

It all depends on what they actually want to design. I think a regulated buck/boost driver for 3-4.35V input and 3A @ 3.7V output is already a rather big challenge in a small form factor.

Temperature sensor would be cool, but only if it can be turned off or set to a custom temperature. I wouldnt mind the light dimming at 100°C before the solder melts, 60°C would be rather annoying tho.

If they already have a driver that can do 5A with the XML2 (which I doubt because it has a Vf of 4.2V at that current) on one cell, get some more info. If it only can push that current from 3.7-4.2V its rather worthless. At 5-6A current draw even the best cells will fall below 3.7V within about 6minutes. Also, ask about efficiency. If its only 50-60%, its worthless again..

TL;DR: We need a single cell driver. There are enough great multicell drivers out there.

^^ This!

Why the hell do you want a buck driver for 3.6-8.4V? That would be useless for single cell lights.

The biggest issue with a 4A+ linear driver is that it will not regulate for very long, especially on XM-L2. Doing a driver that is essentially crippled on what will be the most popular LED in the next few months doesn’t make sense.

For a single cell driver, boost would make sense. It would be DD when the cell is fully charged, then switch over to boost when VBatt drops below Vf of the emitter. I don’t think the DD portion would be insane current either (probably less than 5A).
For multicell, it’s obviously a buck driver. To keep the cost down, don’t do a buck/boost. Do the single cell first then consider the multi-cell driver later.

For modes and functionality, my preferences:
High: 100%, 4 Amps, adjustable via resistor mod
Med: 30%
Low: 4-5%
Moonlight: 0.1%
Option for hidden strobe and SOS as well; similar to the new NANJG without the small flash, just time it yourself.
Low voltage detect/cutoff (2.8V); drop to low and quick flash every 10 sec. About 3-5 minutes later, moonlight or turn off.

Tracking the proposals through the thread could get nightmarish, so just go with mine :wink: so Flomotion, you may want to gather them up and tabulate the similar ideas. At some point you can get an idea for the most popular suggestions and submit the spec to them.
I also propose that, since this was your initiative, you retain final decision control.
Excellent work Flomotion! Looking forward to this one!

Edit: My free time is very limited, but I’d be willing to help with the MCU coding if necessary.

4A Turbo (4A)-High(2A)Medium(500) low (50). 17mm.

Me, likey… :bigsmile:

Of course I look at it differently,

4 amp with a timed boost to 5 amp, timed for 30 seconds then back to 4 amp

1% low, 20%, 50%, 100% = 4 amp and 5 amp boost, that's 5 modes. Hidden modes?, I don't care.

I don't worry about boost/buck because of using NiMHs. I would only use something like this in a 4AA NiMH light. Linear would be best in that case and 17mm diameter.

But, I'm the one percent, so in the long run, go for the best Li-ion driver and I'll keep adding chips to NANJG drivers. I think it sounds like a multiple cell driver if it's going to be all out 5 amps.

I moved away from incandescent barn-burner flashlights because while they were bright as hell, regulation was difficult to come by and expensive to achieve. It is crucial that output can drop out of regulation to extend runtime and avoid sudden cut-offs, but I simply do not want a light that cannot perform consistently for the majority of runtime.
Now I will get on my high-horse.

Well, well, well. While convenient, AMC7135-based drivers rely on brute force to regulate by burning off perfectly good wattage in order to regulate output (sounds barbaric, no?). Furthermore, we have reasonable confirmation (kudos to match) of the XM-L2’s higher operating voltage suggested by the XM-L2 datasheet. The forward voltage of the XM-L2 is decidedly high at the elevated drive currents that we desire—even on a normal, non-SinkPAD MCPCB.

High-horse dismounted.

Match’s data and the XM-L2 datasheet suggest that its Vf at just 4.2A will prevent linear regulators from regulating output for very long, if at all, with 1x18650. And 5A? Forgetaboutit. Needless to say, it will impact the performance of unregulated drivers with PWM’d modes. The resistance introduced from being installed in a flashlight exacerbates the issue.

For these reasons, amongst others:

  • Critically, our wunderdriver needs to regulate at least 3.2A to an XM-L2 for V-in between 3.4V and 3.9V to properly utilize 1x18650.
  • While I’d also be tickled to buy 17mm switching buck regulators that can accept up to 9V and can fully regulate >4A output on two cells, our driver must outperform linear regulators when used with 1x18650 and XM-L2, period

Edit: I agree with relic38’s spec recommendations. RaceR86 sums things up in a compelling manner below.