Texas Avenger "TA" Driver series - Triple channel + Bistro or Narsil + Clicky or E-switch - The Ultimate open source driver!

Ok we’ll here’s v01

To do:
Go over it again, it’s an expesive board to find a mistake on 8 minutes after I pay (how it usually goes)
Delete all extra parts TA points out in his post just below this one.
swap to a solder jumper normally closed (that requires being cut to open it) since I expect it to need to be shorted*
Add BLF driver key programming pads for both MCU’s
Move bottom 7135’s in ~5mm for body clearance

*The jumper on the bottom separates the NET connecting the FET’s in case I dont like it IRL.

Next question: What about R10 (R5 in a regular TA driver circuit) and the polarity protection diodes (D1 and D3) can the pairs of those be consolidated into only one of each? Will just one R5 allow enough current to operate both MCU’s?

Yes, you could combine the input resistor for both drivers. Same for the voltage divider, input cap (C1) and input diode.

Thank you sir, anything else needing correcting jump out at you?

Nothing jumps out at me but I only looked it over briefly.

Updated to V04, now with 36 total 7135’s. Still haven’t ordered yet, redoing the back soon (the 7135’s sit to far out on the board to fit the head right) and adding a few more 7135’s back there too cause why not lol.

so many AMCs with no direct thermal path aka ground layer connecting the lights metal parts will be a problem heating the driver up a lot then go into thermal throttle

this is the reasom most SRKs have the AMCs poisitioned this way around

positioning the MCU more in the middle you can place around 21-22 on one side,
if you do 180° of driver one MCU and the other 180° with the other you got no problem signaling it on 2 layer

Just assembled a TA 17mm driver and it won’t turn the XM-L2 test LED on when connected on my test station.

- The assembled/soldered driver was not tested in a flashlight, but on a driver testing rig (if that makes a difference?). Hooked BAT+ to LED and BAT- to the driver ring. (rig resembles a “clicky light”, not eswitch)

- The backside (7135/battery side) is EMPTY, does this driver work without any AMCs on the back (only with FET +1)?

- I flashed “…/…/Bistro-Texas-Avenger-V1/Bistro Hex versions/bistro-texas-avenger.hex” from the .zip in the original post because I didn’t know what else to flash <- is this the correct file for a 1S setup? Which standard TA-Bistro FW should I flash for a standard 1S setup?

- Even though I double checked polarity on all parts, are there any issues? The Schottky is always my least favorite…

- attiny85 legs are bent, but not “wrapped” around or tucked under the attiny, they should make good contact (but might not?)

  • Checking/flashing with avrdude had no issues at all, flash command and fuses used:
  • Any glaring issues with wrong parts a fresh pair of eyes might be able to see?

I would be extremely grateful for any feedback. 4 pictures of the assembled driver are attached below:

Thank you so much for any help.
It would be great if this would work out somehow.

You putted on the diode reverse polarity.

Your diodes the wrong way.

Also you have the wrong FW if trying to run it as FET+1 (but I’m not positive what precompiled Hex to use for that, you may have to build it)

I knew it.
It has been a while that I reflowed a driver, I am out of practice.

Thanks! Will reverse it tomorrow.

I will put the additional AMCs on the back later, this was just a test before I build the other 2 drivers.
There are so many clicky bistro FWs, I wasn’t sure which to use.

If it is OK I will leave the “bistro-texas-avenger.hex” on the attiny.

Huge thanks for the solution.

The stock firmware has a few mode groups that will not use the back side 7135’s. You can of course compile a custom firmware designed specifically for this as well. I don’t think I ever did that since the mode groups were enough for most things.

Thanks, Texas_Ace.
I will put the AMCs on the back later, but might use this driver at least in one light as a FET+1.

Also, massive thanks for putting this project together in this thread, this is one of my most favorite drivers out there.

After reversing the diode I will report back.

proper FET+1 operation on 3 channel driver need you replace the defined values in the modegroup file

// 1 7135:
#define RAMP_PWM2 4,7,20,35,55,100,160,255,255,255,255,255,255,255,255,255,255,255,255,255
// All 7135s:
#define RAMP_PWM1 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
// FET:
#define RAMP_PWM3 0,0,0,0,0,0,0,0,15,35,55,85,110,125,140,160,180,205,230,255

Excellent, Lexel.

Thanks for adding this bit. I will copy the updated mode group settings and will attempt to compile new firmware.

Quick feedback (and huge thanks to everyone):

- Reflowed/reversed the diode because it was the wrong way

- Reflowed all 7135 on the back as well (pushed them a tiny bit towards the spring but the retaining ring clears either way)

  • Driver now works in a 219b triple S2+

Everything seems to work fine, except medium presses on this S2. They seem to be very long (3 seconds), but that is fine since I don’t need them.

I am now reading through the .txt file to understand the FW better, it is clearly the most sophisticated clicky driver firmware I have used so far.

Again, massive thanks to everyone who helped out.

Measured with a very cheap (free) Harbor Freight DMM (the copper in the leads seem like 1 micrometer in thickness), the S2+ 219b triple pulls 5.5A on the FET with an old Panasonic cell and the rear spring NOT bypassed. The tiny driver spring is bypassed.

Great, glad you got it working!

The 5.5A is a bit low but sounds reasonable considering non-bypassed springs and a cheap cell.

A triple led setup you need to measure with clamp meter. You will have big error with wires above 4-5A.

Thanks, unfortunately I do not own a clamp meter.

To make things worse, I am using this DMM:

The leads are the thinnest strings of wire I have ever seen. I want my 0 $US back! :smiley:

Will try to find two thick copper strips this weekend and measure again, and this time with a better/charged battery (30Q or 25R).

With a nichia triple it could be over 10A so maybe you can kill the fuse in that :slight_smile: