Help with Current Sense Resistor Modding

I forgot to mention that in my list. Small wires are fine for the PWM lead, no harm there. I don’t know what AWG the stock wires are, but they are fine for PWM.

No LED+ is necessary coming in/out of the 105c. It’s fine to use it, but all the 105c does is pass BAT+ directly to LED+ so it’s just extra connections and resistance. Not a big deal.

Side note: Note that while I’m very confident that you can get away with leaving the FET in place, eventually if you hook enough stuff up to an output pin of an MCU, whether that stuff is being used or not, the MCU will struggle to send out a strong enough PWM signal.

I'll remove the FET. I also got the Attiny13a removed too (that was tough!). I've not got all 22ga wire installed ready to mount to the existing driver. I even piggybacked another 7135 My first attempt at that ever! And I found that to be MUCH easier than attaching these darn wires! Just hope I didn't burn up any of the 7135s.

Im gonna have to quit for now. Thanks for the help!

-Garry

FWIW, and I don’t do as many of these as others, I’ve never burned up a 7135 or had a “bad” one that I noticed. I’ve torn legs off of them several times, but it’s always been pretty obvious that I had physically damaged the thing.

A heat gun makes all this much better!

Well I couldn't leave it alone, so it's connected!

Best at pic I can post until I get a chance to upload my high res macros. Not too bad of a job! Now I hope it works! Wait - let me test quick. Well it works, however I'm getting 2.35A measured in each mode (it still ramps down then back up between modes) with 2 unprotected cells in parallel at 3.99v. Ugh! Why'd I lose mode levels?

Ill probably post high res pics Monday. I'm done for now.

-Garry

I am unable to see the PWM lead in your picture.

I took a closer look at the stock driver. It appears to be setup for current sensing after all. It’s probably still just doing PWM to the FET and using an RC circuit to smooth the sense voltage out.

so err, sorry! If I’m correct I’d say that this is not really a common setup. I’m not sure that the best solution is not to re-assemble the driver and do a sense resistor mod. EDIT: hold that thought. I’m taking an even closer look now. EDIT 2: OK, I’ll stand by that for now. Forgive me but this driver is only so interesting and I won’t stare at it all day. It looks like what I wrote in strike-through text may be the case.

Well I already broke the old FET getting it off :( .

-Garry

Hmm, I’m not making much sense here. Why would you be limited to 2.4A if you bypass the sense resistors? They’d have zero voltage drop across them and thus the driver would attempt to run full blast. Maybe it never provides a 100% duty cycle no matter what. Your 105c is fully populated with 8x7135?

Anyway you can probably just use a normal FET like the 70N02 or AOD510 if you have those.

My 105c has 9x7135's now. Possible I burned some up? I don't have any FETs here at all on me. I did connect the GND directly to the incoming NEG lead (connected across those resistors where they connect to GND. The PWM wire is connected to FETs gate pad (lower left). LED- from new driver to to the FETs drain pad. I also checked for shorts on the 105c while wiring up.

-Garry

If you would like to do an experiment let’s move the Nanjg 105c’s GND connection over to the right of the sense resistors and see if modes come back. I do not expect to see more current, but I do expect to see modes / levels.

You're right! Modes now show:

L-1.06A

M-1.38A

H-2.08A

Too closely spaced though. And low & medium fluctuate a good bit on the meter while high is steady at 2.08A.

-Garry

Sorry for the mixup earlier. So it’s as I said: this driver appears to use current sensing through an RC circuit to do “CC” (not really) PWM. I’d drop an FET in and do the resistor mod to get the desired current.

How badly did you damage the FET when removing it? I’ve been doing some serious bending on one or two DPAK FETs over here and they aren’t dead yet…

I broke a corner of it off and one of the legs. Is there anyway to keep using the 105c with it (and get more output)? I know I could always just use it as a contact board and run the light entirely off the 105c, but I'd have to wire another one up that still has the MCU attached.

And so those ARE current sense resistors and NOT current limiting resistors? So going back to an FET would give me regulated output or not?

-Garry

On a 7135 driver loss of modes sometimes means the pwm pin is pulled high (usually because the led + pad is shorted to the pwm pin closest to D1) but a momentary switch should not be able to interrupt that. Is there anything else on the trace we think is the pwm that might affect the signal ? If there’s unwanted capacitance there might it not prevent the 7135’s from turning off the way they should?

I’ll tell you what I think:

  • Either the maximum duty cycle the MCU (or whatever it is) will provide is <100% OR the 7135s just do not like something about the PWM signal (very high freq for example).
  • Those are current sense resistors but they are not used in a traditional way. This driver provides PWM output using an FET (like an East-092 driver) and varies the PWM to achieve a specified average current. That is not the same kind of output that a buck or linear driver generates. I’ve actually considered trying to do this but I decided that there must be some reason not to since nobody else was doing it. Kind of like the opposite of “if everyone else was jumping off of a bridge…” The average current is being regulated, this is not the same as what we normally refer to as regulated output.
  • You can probably grind the FET until that leg/pin can be soldered to. Using flux will probably help. Maybe carving with an X-ACTO knife could give better results than grinding/filing, depending on the situation.

I don’t think you have the ability to flash MCUs, but if you did you could use ToyKeeper’s STAR mod which features support for status LEDs. You’d just wire up the 105c to the status LEDs & their limiting resistors on the stock PCB.

I already threw the FET away and can't find it. Is there any easy cheap replacement (like an EBay listing?). Any FET work?

-Garry

Would the AOD510 from one of the DD boards work? It doesn’t look like you’ve done anything beyond wasting that one part.

This from Digikey

Any decent DPAK FET which we’d use on a regular DD driver should work, AOD510 included.

RMM sells singles of IRFR3711ZTRPBF, that should be fine and I think it’s the cheapest option if you just want to purchase a single unit. Should be <$4 shipped.

I've ordered 2 FETs from Mountain (along with other parts to keep on hand). In the meantime, here are those higher res. photos I promised:

My final driver wiring after changing over to 22ga wires and removing the + wire (the PWM wire is the red):

Ground Wire on the Back Side Due to Thicker Area to Attach (and my stacked 7135):

Closeups of my stacked 7135;

Few views of being wired up to the main driver board with FET removed:

Really tough to show the wiring with so much in the way! And those photos were prior to me moving the ground wire to the other side of the sense resistors. This was the wiring that gave me no modes.

I'll post back when I get the FET and get it remounted.

-Garry

^ Sweet driver mod you got going there. That 7135 stack is just showing off. So many of us wish we could stack those that well.