Ok, it made it 11 minutes and is still working. Body was very hot around the driver and pill, but it survived! I fried four emitters playing with this driver, but it turns out that most of my tinkering was likely unnecessary. I played around with every MOSFET on the board, but the only one I ended up leaving changed is the "C00C" MOSFET next to Q2 (I don't recall the number itself.)
I removed the stock "C00C" MOSFET and replaced it with four piggybacked IRLML2502 MOSFETs. With only one or two it was too hot to touch in a matter of seconds, with four it will get too hot to touch within about 30 seconds, but it's a big improvement and it's holding for now.
Here's what I have stacked on top of the stock sense resistors (I kept on adding until I got to 5.8A, so you can use different values to get the same effective resistance.)
(2) 100 ohm (1 on each bank for all)
(2) 75 ohm
(2) 25 ohm
The flyback diode is also getting pretty warm, but not too hot to touch. I will run it for a while longer tonight and see if it holds up.
I still think that (4) Knucklehead drivers running at 2A each would be pretty awesome...8A at the emitter, that's what I'm talking about. Will it work? As soon as my coils come in I'm going to try some running in parallel with 1 MCU to rule them all!
8amps would be simply incredible. I would love to see that happen! I have to think it’s possible as long as it’s really well regulated and there are Zero spikes.
I have to say I’m pretty happy with mine after adding the R025 but then again, who doesn’t want more?
I would check into your numbers btw. You measured 2,85A at the emitter and 1200 lumen OTF stock. Those numbers does not seem to go hand in hand to me. Even if the emitter saw 3A. Would you expect 1200 lumen OTF? Seems too high compared to Cree`s own data, Match, djozz, Fenix, Selfbult, etc.. For all I know, Fenix could have gotten U3 and not mentioned it, but I doubt it. Still, would seem a bit high.
It does seem a bit high to me as well, but unfortunately the stock light I was measuring was a different copy. I never measured the one I modded beforehand, so I don't know how high it really was. I have a suspicion that the current is actually a bit higher at the emitter than what my meter says, due to the extra resistance of all the alligator clips, etc. I was using.
Solder a loop in (basically extend the negative lead) and clamp around it. Piece of cake and little to no extra resistance is added. On the bench, it’s easy to measure directly from the leads between the driver and emitter, no splices cuts or clips required.
Took this out last night. All I can say is that my sight was limited only by the environmental haze / dust, not by the level of light. I think I might like this one better with the dome left on but with the current cranked up, that way I get a bit wider beam. The BTU Shocker cranked up has a better beam (MUCH wider, very similar intensity) than this IMO, but this is much lighter and the runtime is much longer!
If these didn't hold up, that was going to be my next step: remote mount a 70N02. I have run several 10-15 minute torture tests and it's holding strong so far.
RMM i hope you will stock some mosfets for this light, it will be very good to have an option to buy them from your store, instead of endless searching on mouser site.
I think I may try the 70N02 and try for 6.5A-7A with the stock driver. If it blows I guess I'll have a $150 paperweight, huh? Nah, that will just motivate me to try my 7A-8A option...
If 5,8A is peak. Im already headed for my 7-8A option.
But I could try 70N02 first. Just replace/remove "C001" from Q1 and connect the 70N02 in similar fashion + further resistor mod? I can try that if you want. But if you are going to do it anyway, I could just sit back and relax get some progress done on my scratch build.
Racer, I'll let you try it first! I'm pretty busy today. I would keep the resistors the same to start with and see how the current compares. I think that at a certain point we're really going to need a larger inductor, more input capacitance, etc. to make this reliable.
No, that's the correct way. The gate charge, PWM frequency, etc. may not be suited for the 70N02. I guess we'll have to stick with stacked "tiny" FETs for now. BTW I have run this one quite a bit more and it's holding up well. I feel confident with it around 6A.