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

Flashlight Modding and DIY Parts
561

It lives! Went with long LED 20 AWG wires so I can easily re-flash the 85. Ramping looks cool, temperature monitoring is working, max looks bright. Actually left LVP in using the internal ref and it blinks out a steady 5.2V , k, not very useful but it works as expected, roughly.

Really tough to get those 20 AWG wire solder pads down low though, working down the head of the L6. I covered them in kapton tape just in case. Would like to get the reflector sitting lower, LED more inside the reflector. Think some output in being lost. Think Dale machined out the head for a bigger MCPCB - it gets the wires cleared from the reflector, so can sand down the alignment piece to get it lower. Gotta do some measurements.

562

Nice! Any idea what kind of amps / lumens it is putting out? Is this the stock 3A emitter?

I am considering cutting out a grove in the reflector to clear the wires in the next DO1 I build as it is such a pain to get that reflector low enough to focus. I kinda doubt it would even effect the beam as I doubt any light would be able to hit that low on the reflector but won’t know till I try it.

563

Topping cells off - quick 1 second test in the PVC light box said about 6K lumens, using the stock N4 2B LED - this is the older black L6, think those 2B's later were not available. No amps yet.

564

Not bad, Curious how it handles the heat as well.

565

Ok, for numbers. On 2 BASEN cells (no spacers) @4.22V: 10.45A (tail-clamp meter).

lumens: 6,154 @start - 5,661 @30 secs, throw: 120 kcd taken at 5m, 693m

On 2 LK cells @4.22V using the Jaxman spacer w/bypass: 11.8A (tail-clamp meter)

lumens: 6,283 @start - 5,749 @30 secs, throw: 124 kcd taken at 5m, 704m

Gets warm after the 30 secs - no time right now to do more. Gotta get my Haberno's jarred, toilet leak fixed, etc...

Did not do any tweaking with the stock LED or centering piece, but have extra long 20 AWG LED wires to re-program the driver. Driver spring and tail bypassed w/22 AEG wires. Just fyi, this black L6 w/ N4 2B under performs compared to the N4 1C in my clear ano L6. Wondering if it's a true N4 bin, or something in the optics.

566

Wasn’t there some debate as to weather there ever was an N4 in anything but the 1C? Seems like everything else was N2.

Still 6000 lumens is not bad at all for an otherwise stock light.

567

Just tried another 21mm from first batch in TY-08. Had an XM L2 stock. Too much power on FET. Used mode 13 all regulated 5 modes. Perfect. Very versatile driver (Bistro) plus the extra blinky modes. Probably switch to XPL.

568

Yeah, I am finding that I swap to the all regulated group in most lights as well, I always have turbo in the hidden modes if I want it.

569

Hhmm, sounds familiar - could be right. I got one spare P2 1C from Richard a while back when he made them.

570

Finished prep'ing/jarring the peppers, didn't get to the leaky toilet yet though.

Did some checks on moon mode. I get 0.4 lumens on PWM level 3 (default for moon, 1st ramping level), and somewhere down below 0.1, maybe 0.04 lumens at PWM level 2. In Narsil, moon mode has a programmable level. At level 2, the LED basically glows in room light, but it's actually usable in the darkness.

Then did a run test. The temperature stuff probably needs tweaking for this big light. Here's a time line:

  • before run, reads 22C from the temp sensor in the ATtiny85 at room temperature (67F), trip point is 55C
  • dbl-clicked to max, tail standed the light (not that stable tail standing)
  • at 2' 38", heat monitoring tripped, dropping output - head was warm, not hot, outside temp on head (IR meter) read 48C
  • at 3'18" it tripped down again, outside temp read 42C
  • turned off the light at 4'30", outside temp read 39C - comfortably warm

Probably could raise the trip point from 55C to quite a bit higher. I want the temp tripping as a backup, before damage sets in, not because it's getting warm in your hand. Weird that the temp sensor still reads high 40 secs later, when the outside head temp clearly has dropped. Might be really what's goin on, who knows...

I think I read a production light had a trip point of 80C - that sounds crazy high though.

571

Not bad. The temp difference doesn’t surprise me. The TA drivers have a thick copper pour to under the MCU for this reason and as such the temp readings seem to remain quite accurate without the need for an external sensor. The temp inside the light should be hotter then the outside.

The only real issue I have had is the light getting “stuck between modes” and swapping modes each time the temp regulation kicks in but any temp sensor would do that.

80C actually doesn’t sound all that bad for a step down point, although some playing would be needed.

It would be quite hot to hold but it would still be well under anything being damaged. The LED’s are binned at 105C now days, meaning they are designed to be used 24/7 at 105C. The other components have similar ratings.

It is really more about the hand comfort then anything.

572

On the SRK version, has anyone tested our bright idea of pulling 1s voltage off of 2s/4s to run the mcu? Half of me says it is genius, half of me says it violates electrical law

573

No one has built one yet that I know of. I have some ordered and they said it shipped a few days ago. So I guess I will be the first to try it out once they come in. Might have to give the 8x XHP50 SRK a go afterall.

574

I might get a chance to try it (the concept, not your board) myself today, just depends how much time I have

575

Ok, here is the official release of the Bistro Texas Avenger edition V1. I have everything I could cram into the available space in here. If TK every has time to compact things a bit more then I plan to extend it with 31 mode groups and some more blinking modes along with a few other things but they will not really effect 98% of people.

This firmware will work on any of the Texas Avenger drivers when used with a clicky switch. For e-switch lights we have the amazing Narsil by TomE to take care of us (of which I have not been able to find a suitable light to try it out on myself yet).

So a quick rundown for the TA version of bistro.

It has 24 mode groups in all sorts of flavors
You can adjust the mode order and moon mode separately to further increase the options to 96 different combinations of modes!
Mode memory can be toggled in the menu
It has several neat strobe modes that are quite fun to play with
The modes have been optimized to work with the triple channel setup, putting as many as possible in non-pwn settings
The off-time settings have been recalibrate for the TA drivers which seem to be more stable and not drain the OTC as fast
Thermal regulation with a user calibration setting in the menu

A few other minor changes that won’t really matter.

Basically this firmware is setup so that 98% of people can simply flash it as is and tweak it in the menu to get exactly what they want out of it without ever having to touch a line of code!

Flashing the firmware is quite easy to do with less then $5 worth of parts and a computer. There are threads on this if you search with way more info then I could give. Flashing a pre-made hex file is quite simple though even for novices.

It works exactly like normal bistro, the only changes have been to the modes and mode groups to work best with the Texas Avenger drivers.

Before I show the mode groups here is a basic run down of how to interpret them:

Modes approx mA and lumen values for single XP-L LED
1 = 5.5ma = .5lm
3 = 19ma = 4lm
4 = 37mA = 12lm
5 = 66ma = 24lm
6 = 135ma = 60lm
8 (one7135) = 355ma = 150lm

Above this it will depend on how many 7135’s are installed
With 7x 7135
10 = 640mA
12 = 1A = ~335lm
14 (7x 7135) = 2.55A = ~830lm

Above this you are using the FET and it will vary wildly depending on the build.

16 = ALL7135s + 25% FET
18 = ALL7135s + 60% FET
Turbo = 100% FET

Now here are the mode groups as seen in the code, using the above lists it should not be too hard to figure out what numbers will deliver which current/brightness.

Click for larger version

Hidden modes are as follows:

Turbo > Batt Check > Biking Strobe > Strobe 10hz > Strobe 16hz > Old Movie Effect Strobe > Creepy fun house strobe

Here is the download link: Bistro Texas Avenger V1

It includes everything you need to either compile your own version with all the needed extra files already in place, or use the per-compiled version and simply flash it with AVRdude. It includes a batch file to make that much easier as well.

It also includes some useful python scripts by TK for calibrating things along with the voltage, OTC and temperature calibration firmware for those that really want to dig into things.

576

I know the feeling, been trying to know out a bunch of little things today. Going well so far. Got the firmware finished as good as I can with the space available. Got the Texas buck polished up and hopefully ready for release. and a few other things. Should also be able to list a dozen TA drivers for sale later as well as my first official batch.

577

Oh the buck is done? I haven’t checked that thread in awhile. It’d be nice for my TM16 XHP35 project, but I don’t think I can foot the bill right now.

I just realized today could be the last nice day to do an oil change so that has to have priority. I’ll report back if I get some testing done

578

Yeah, flint got the component list sorted out and I think the PCB design is done. Just got to get one last round of double checks due to the how pricey it is and then I will release it in the coming week I would guess.

It should be a really nice setup for XHP35 swap in the Q8 although it sadly could cost as much as the Q8 to build it. Although it would also offer the option for a buck driver for other custom setups as well.

579

Ok, pulling single-cell voltage off of a 2S or 4S physically configuration seems to work perfectly. Obviously you’ll still want to keep parasitic in check so that the cell voltages don’t become too unbalanced.

As expected, 7135’s do not like 16.8volts!
They don’t die immediately, but they don’t give anywhere near full output (I was getting less than 500ma from 8*7135) and weird stuff happens. They seem to lose their ability to switch off completely. For instance I was using a stock EE X6R driver for testing. On strobe mode, it would flash between one level and one slightly lower level, never going off. As I cycled through the standard modes (pwm based), it would only vary by about 5ma between modes. If I totally removed power to the driver (still 12v going to the LED, but switching off the 4.2v to the driver) so that the 7135’s were not getting any PWM signal, the LEDS would remain lit until I completely removed power. But I know I don’t have a short because if I connect power back to the LEDs, they don’t come on until I fire up the driver again. It seems like the extra voltage is causing a small internal short inside each 7135, or something like that.
Edit:
Here is a video just showing the battery configuration, maybe to help explain better what I can’t put into words (for those who don’t already know). It was filmed before adding the driver and LED.

580

Good info, I thought I had read that the 7135’s would work at 4S voltages just not very well. Good to know the truth.

Also good to have confirmation that pulling voltage off the single cell works as expected.

Hopefully the Texas Avengers I listed sell well so I can recycle the funds into building a batch of the SRK and L6 drivers.

It will let me finally try Narsil for myself as well!