Yeah I know mtn ships to Europe now. But I have lots of stuff planned before trying out dual-pwm. Iâm not really interested in the EE A6 either as I think itâs way too small to be running DD. If it was a Convoy M1 I would have signed up for one though. I could of course get the A6 and move the driver to a M1, will have to think about that.
Thatâs interesting. Do you want the two emitters to light up at the same time, or at different times? Iâm not aware of any FET+1 drivers which were intended for two independent LEDsâŚ
Hah, I have same-sized lights running at 3X the power of the BLF A6. It just requires being more careful with the light⌠You can run 15 amps in an 18650 tube light for short periods of time; just donât leave it on high modes for very long.
I generally wonât run a light that size at more than one amp for any extended amounts of time, and usually the highest I use is 350mA (98% duty cycle) with a 1Hz full-power stutter (2% duty cycle), for an average of about 500mA or 600mA. This works well for biking, and the air flow keeps the light not only cool but actually cold.
Basically, I treat it as a 1-amp light with a really nice blinding turbo mode and bright party strobes.
I want them separate so one group of modes for one and another group for the other. Iâm using a FET +1 with cut traces to give two different LED - outputs.
Just given it another go building it slightly differently and it WORKS!!! So excited managed to finish my build!
Only thing is on the second group in the two lower modes both LEDs come on but not in high or in the first mode group. Odd but I can live with it for now.
Here is the new code I used.
Thatâs because you used fast PWM, which still turns on for about half a clock tick each cycle even at level zero. If you switch that to phase-correct, it shouldnât turn on the main LED while the second LED is on.
This only happens on the FET channel because the 7135 channel isnât powerful enough to make the effect visible.
I have a 32x7135 SRK driver which still lights up at fast PWM level 0. I guess with enough of them in parallel it can rise high enough to barely light the emitter at that speed. But yeah, the rise time is slower so itâs not normally visible until like six ticks or so on regular nanjg drivers.
Ever heard of search services? Nah just kidding, as the previous poster showed a pastebin is a place where you can put code to get it away from the forum flow and also get line numbering and syntax highlighting if wanted. You ask a question which only a few people can answer and the rest that canât donât have to scroll past your code in the forum if you put at a pastebin instead. If youâre interested you just click the link to the pastebin and you can see the code there.
1. I currently have blf-A6 configured so that âturboâ isnât the last mode in the group. I think because of this, the turbo stepdown isnât working. Is there any way around this, or does turbo have to be either the first or last mode?
2. This may not be something fixable by firmware, but I have noticed that OTC times speed up dramatically as the light gets hot. The values I have are perfect for forward/reversing normally, but after the light heats up even the shortest of taps makes the mode go backwards.
3. This firmware is in my light with red and white ledâs. The small circuit goes to the red. On the last mode in the second group, it should be full-on red and no white, but the white ledâs do light up at a moon level. Is this just leakage through the FET, or did I crew something up in the firmware?
Turbo knows what to step down from, but not what to step down to. You can change this by changing âmode_idx = solid_modes - 2;â to âmode_idx = mode_idx - 1;â. It looks like this should work in your case, even though it wouldnât have worked if you had a turbo in the hidden loop. You could also just comment out the â#define NON_WDT_TURBOâ to get rid of turbo step-down entirely.
2. Iâve been trying to fix that too, but I havenât figured out a way to do it. I tried to use memory decay in combination with OTC measurements, hoping that mem decay wouldnât have temperature issues⌠but that doesnât work either. The OTC discharge keeps the memory from decaying when itâs supposed to. In any case, the temperature scaling is annoying.
3. Change all your red-only modes to use PHASE instead of FAST, and it should prevent the white LED from coming on. You canât completely turn anything off in FAST mode.
Thanks, Iâll try that. Based on the code I posted earlier do you think thatâs what the problem was? Turbo stepdown was working in the first mode group, but not the second group. Btw, Iâve been wondering what does the âNON_WDTâ part refer to?
I guessed switching to PHASE would fix that 3rd issue, at least iâm starting to learn something!
You had turbo as the second-to-last mode in one of the groups, and it was configured to step down to the second-to-last mode. So it stepped down from turbo to turbo.
The reason it didnât do âmode_idx â;â before is that one of the default turbo modes is hidden / negative, and I didnât want it to âstep downâ from turbo to strobe. Your version doesnât have to worry about that though.
The âNON_WDTâ bit should probably be renamed. The code was originally based on STAR which used the watchdog timer to measure time for turbo, but it took too much room so I got rid of that and replaced it with a non-watchdog version. At one point there was an option to switch between the two methods, WDT turbo and non-WDT turbo, but I removed one.
BTW, why do you have both red and white turned on at the same time in one of the mode groups?
Ok, I wasnât understanding that. Makes sense now.
I was just playing around with stuff to see what it would look like. The light has two dedomed XP-Lâs and a deep red XP-E on the same 20mm triple noctigon. It actually helps bring out the red spectrum quite a bit. My âCRI testâ involves shining the light at my wifeâs clothes in the closet. You donât really notice much of a tint change, but the added red really makes the patterns âpopâ, even more so than my Nichia light. It does some cool/interesting things to shadows too.