LED drivers and Accessories you want, but don’t exist

I measured the board the connections from the buck/Boost driver to the other parts are all like the board design

and no shorts between two pins of the buck converter measured

I build now all 3 drivers and test if one works


And just leave C8 open. Wait, is there a 206 resistor on R1? That is 20 mega ohm. You need 20 milli ohm there.

Was about to comment, R1 looks like sense resistor so I think 20m was 20 milliohm, 206 is 20 megaohm, so would make sense if no light with 20 meg in series with the emitter.

OK it makes sense was a bit wondering about 20MOhms but I thought with C8 some kind of protection for the OPAmp

but as I had no layout I thought it is 20MOhms

Shoki you may have a 0,02Ohms 0603 resistor that you can send me 3?
or I add there some 0.02mm² wire of a certain length

I got here only 3 0.1Ohms 0805 resistors in my box
I think I got some dead Klarus XT11 Drivers I may salvage a few 0.05Ohm 0805 resistors

and a small zip lock filled with stripped driver parts I may find some small resistors

I currently don’t have, but I can order some, wanted to make a mouser order this weekend for the new MP3431 board anyway. If you want I can buy some 20mOhm resistors and send them to you.

Do you have higher resistor values in the milli ohm region, just to check if the board works?

The 0.02mm² wire won’t work, 1.4A through that is a bit too much (50mA is the maximum current for that size). For now, maybe a 0.1Ohms 0805 resistor fits. Just to make sure it works (and it should put out around 300mA). I can order some resistors for you, no problem.

And with 20MOhm, you were driving an LED with around 1.5nA :laughing:

after looking on my dead drivers and an old motherboard I found suitable resistors

0.1+0.05+0.05 parallel are 0.2
and a way oversized 0.02 one


If you can solder the three resistors on top of each other, go for it. Let me know if you then still need 3 0.02Ohm resistors.

OK I stacked resistors on 2 drivers one without and one with C8 both do not light up

Ok found it I soldered positive to the wrong side of the spring pad on the one without capacitor

I get 1.65A on the one without capacitor

so I have to reduce R2?

To get less current, the gain of the Opamp has to be increased. So R2 has to be higher, or R3 needs to be lower.

stacked a 19.1k Ohms on R3 so I get 1.48A that seems OK for 1.4A UV emitter

if you do a order from Mouser or digikey soon could you get me 3 0603 0.02Ohms resistors?

Good to know. And the other driver with C8?

does not work even with removed capacitor, I think a contact of the boost converter is not connected

the 3. with only a 0.1Ohms resistor and capacitor works, at reduced current

Was that a driver you powered on without shorting out two pads at first? Or have you tried the now working one without shorting the pads as well? Maybe the IC is toast because you tried it without shorting the pads at first.

OK, the 3rd one works as expected.

I inspected it with magnification glass and it looked like the IC had no connection to the inductor on the side where it is hidden behind C1
I used very little paste on the outer pins of the IC to do no shorts there

OK reflowed it with a bit paste and now it works
this driver does spot on 1.41A, probably a lot variation on this 3 stacked sense resitors

The IC itself has a feedback variation of +–5mV, the resistors have variation (1% or so), and the offset by the OpAmp adds some more.

Nice, now all drivers work?

I have found out that the driver gets pretty hot and the output current drops with higher temperature slowly

on my bench one driver quit working
it was the one with the capacitor

removing the capacitor and the driver shows a fast strobe
adding it again and it is dead

it looks like a bad ground issue, need to add maybe a bit more paste to the IC

A fast strobe? How fast? And it quit working? Was it still hot or cooled off already when the fast strobe occured?

Did it get hot in a flashlight, or was it on a bench uncooled?

Even though it might have 90% efficiency, 0.5W or so will be heat.

Schoki, since you mentioned that 3x3 MCU is the max you can fit I did some research…
Would Silicon Labs C8051F52x/3x series work?
They:

  • are available in 3x3 mm 10-pin DFN package
  • have 2-8 KB flash
  • have 256 B RAM
  • have a 8051 core, max 25 MHz
  • work in 1.7V-5.25V range
  • work in 40 125 °C
  • have a temp sensor and PWM
  • cost more than ATTiny ($2-$2.5 in small quantities)
  • can be programmed with open source tools on Linux and closed source freeware on Windows (dunno about Mac / BSD / …)
  • the programmer hardware costs $35

Though that said, the new attinies which we can’t program may be close enough to make the porting effort not worthwhile…