buying an SBT-70 and putting it in this TK61, letting the power fly. 
Ultimate throw be damned (I’ve already got 617Kcd in another light)
attiny13 just stops working if voltage goes over 6v. Besides, whatever component is getting the PWM signal operates from a digital signal (either on, or off) so more voltage or less voltage on the PWM pin should have no effect on the output. 6 volts is just as 'on' as 4 volts.
I don’t think this is the case. Comfy’s explanation is correct, a higher voltage on the buck controller’s PWM input will not cause the problem DBCstm described.
It’s more likely that the driver produces high voltage spikes at higher currents. We have that problem with other drivers as well. The “big” LEDs seem to cope with this better than the XM-L/L2.
I did think that the hx-1175b was shown to work OK driving an XM- series LED at higher currents than that though… Am I remembering wrong?
I thought some buck controllers do relative voltage adjustment.
Are you two talking generalities, or about this specific buck controller? If generalities, do you feel 100% certain that it's safe to rule out the PWM voltage as the potential issue?
“generalities” puts a negative spin on what we are saying. We are telling you how all PWM inputs work.
The only thing a PWM signal communicates is “high” or “low” (two states, like zero and one).
Not trying to cast dispersions on what you 2 are saying. Just trying to figure out what we know for certain in order to figure out how to help Dale.
The driver seems stable, but I have not used an oscilloscope or anything to search for spikes. Given no other options at the moment, I think Dale should grab the voltage from the point I recommend he use in Post 82 and see what happens.
Don’t worry, no offense taken. I just wanted to be plenty clear that we aren’t speculating about the PWM. PWM signals work (provide a duty cycle) or don’t work, but there is no device that gleans something additional from the voltage on a PWM input.
Can you remind us of how far this driver has been successfully pushed with an XM-L?
I haven't tried to keep track of that. I know some people have used it to drive xml's using it in it's stock configuration (7ish amps). It's a popular UF-T90 mod. Driving an xml/xml2 over 7 amps is touchy business with any driver. I did a quick search and saw a couple examples of driving xml's in other lights or just testing of 7 amps and less.:
Interesting. Thanks for the links, that’s the kind of thing I needed. So, while I’m confident about the PWM voltage not being the source of the problem… maybe it would have been better if DBCstm could have tested the stock driver on an XM-L/L2 before doing the ATtiny13A mod. My money is on the idea that he just got a bum driver!
I got 2 and have had the same issue with both. The issue is in the driver giving too much Voltage to the emitter and frying it when using larger wires. Richard has been having the same issue I’ve had and is working on solving it. He’s burned up 10 LED’s to my 5 or 6. So now I know it’s not just me.
So, apparently putting 20 ga wires on it lets the buck converter run the Voltage up too high.
So you’re saying that with the stock wires it doesn’t zap XM-L’s, just once you bump to a heavier wire gauge? (just verifying!)
That's right. I had one driver that I got away with the thick wires on...The subsequent ones have been blowing LEDs.
Bummer. Starting to sound like this is an MT-G2 emitter only driver. Thank you for the info guys. I was just about to build a light with this driver and xml2's. Need to go another route, it seems.
I will post some more details soon, but all it needs preliminarily is a little output capacitance. 22uF ceramic is enough to keep it from blowing LEDs with the factory sense resistors.
I also tried swapping FETs, diodes, and adding more inductance, none of which seemed to help at all. I was surprised that adding more inductance didn't work.
I will see how far I can push it with extra capacitance. Unfortunately I don't have an oscilloscope so I can only tell whether or not the voltage is staying within the limits of an XM-L2 but I suppose for our purposes we don't need a perfectly clean output. 
Good work RMM. It seems that an inexpensive scope might pay dividends if you are burning up a lot of LEDs in your testing? Just FYI, the PC based one comfychair is using is $75 shipped from the US.
I do have a scope, although I only get one channel. It seems that my QX5241 driver zapps LEDs at 6A, based on one potentially flawed test. I’ll get some more caps in hand and take a look at that on the scope ASAP… starting at 4A of course ;-).
Just a another quick update:
With two 25v 22uF ceramics on the output I can run 50 and 68 ohm sense resistors together with no issues, which delivers over 8A to the emitter with 18 AWG wires. My TK61 is pushing over 2000 lumens OTF now!
Sometimes the simple solutions are the best ones.
I know you are always pressed for time. Would it be possible for you to grab the 4 values required for an efficiency calculation? Vin/Vout Iin/Iout
Either way, thanks for the update!
I spend much time at the law office these days! If I remember correctly, it was pulling about 5A on the input for over 8A on the output... sorry but I don't remember the exact numbers but I am going to be building some more this week. I'm guessing that efficiency is around the 85% mark when it is run full bore.
I think I am going to order an oscilloscope. There have been a few things that I have been able to pin down by trial and error and lots of online searching, but it could definitely be a time saver (and time is what I am usually short on.)
In a few more weeks I will be done with my law office/internship assignments and back to regular school, which interestingly enough gives me more time experiment, list, etc. I thought I would have more time this summer but I have had significantly less because I decided to take on the second job along with the first.
RMM wrote:
With two 25v 22uF ceramics on the output I can run 50 and 68 ohm sense resistors together with no issues, which delivers over 8A to the emitter with 18 AWG wires. My TK61 is pushing over 2000 lumens OTF now!
Sweet! Congrats and thanks for the info.
EDIT: Please let me know if you start selling modded versions of these drivers. I will put a link in the OP.
They don’t have the same pinout, that much we can be certain of. I’m moving this conversation back into your thread though, this info is more useful there!
Based on the photos in RaceR86’s post #67 in this thread I’d say we’re almost certainly dealing with QX9920. If that's the case, I'm starting to see why they've got all those resistors around the PWM line. They appear to be abusing the EN (enable) pin rather than using the TOFF (PWM) pin. I wonder why?
QX5241
1. CSN
2. DIM (PWM)
3. VIN
4. Vcc (regulated output for MCU/etc)
5. DRV (gate)
6. VSS (GND)
Mystery chip
1. DRV (gate)
2. VIN oops, GND!
3. (PWM)
4. VDD (from zener W8 by way of S4 RP protection diode)
5. ?
6. CSN
QX9920 (marked with LEDA & date code probably)
1. DRV
2. VSS (GND)
3. EN (Chip Enable)
4. VDD (needs regulated VIN)
5. TOFF (PWM)
6. CS (CSN)