Doing a factory reset would put the light back into the very short thermal position as it comes out of the box, it would be necessary to again turn off the thermal step-down or set it after a factory reset.
I performed the factory reset first, and i did the thermal protection disable just as you (and the readme) described and it has no effect.
i checked into one other bistro-MTN17DDm-NUV light i have (and the one i built for a friend) and its voltage readout looks to follow the same pattern of 1 second pauses with a dependent number of blinks after and no tenths readout. less blinks for batteries at lower charge levels.
My guess is Richard is using a modified version of bistro? this may solve the voltage readout as actually being a problem, but still doesn’t explain why it steps down no matter what after 4/5 seconds.
Edit: All flashlights i checked have the 8 setting menu and not the 2 setting menu of the A6 FW.
The MTN17DDm driver I have, have a different voltage readout, the same as described on the MTN homepage :
“Battery check indicates by a number of flashes between 1 and 8 how full the battery is, with 8 being full and 1 being empty”
Yes, the Bistro from Richard uses the 8 blinks for full indication, rather than the blinking the volts and tenths.
I thought I had heard of people having problems with high current and the attiny 25. I never understood how the MCU could be so directly affected by the current through the FET, unless it is a thermal issue.
I am using the driver in a triple X6 with Nichia NVSL219C D320 5000K 70+ emitters, I haven’t had any problems with the setup.
I just checked the one I’m having the issue with as well. I also get a weird blinky mode where I expect the voltage reading. It blinks four times, pauses about 3 seconds, then blinks fives times over and over with about 3 seconds in between. I have also set the thermal calibration and disabled it several times, and the only time it doesn’t step down after 4-5 seconds is when it is actually in thermal calibration mode.
Also of note, it does not continue to step down after the first step down at 4-5 seconds. And, the level it steps down to is actually lower than the mode below turbo. It is set to 5 modes + moonlight. It does not step down at 4-5 seconds in the mode below turbo…
Ah that explains it, i have been going by the bistro.txt from toykeepers repo.
Looks to be the same as i am seeing.
Ahhh, I didn’t know Richard changed the battcheck modus operandi. 
I don’t have one of his drivers with Bistro on it to my knowledge (meaning, if I do, I’ve forgotten)
I was also using that as my source of info. Thanks for that tip!
So, that answers the battery check issue, but not the step down….
Only thing left to do is to take mine back apart i guess to see what can be done. 
It sounds like there is definitely some sort of problem with the voltage divider. You should measure the voltage on pin 7 with the driver powered. You should be looking at a voltage of 0.6v or higher @ 3V input.
Alright, that’s a good idea/hint. i’ll try to test it out tonight.
Basically, you want to know first whether the MCU is getting the signal it needs, then go from there.
Finally had some time to take it apart and test it out last night.
Readings on pin 7 are a pain in the ass to get, what readings i got were very inconsistent, lowest i saw was about 0.56v @ ~4.6amps (other readings were around 0.68v) during turbo then dropping to a much lower level after 5 seconds to about ~0.8v with the only cell i had at the time to use at about 3.6v. i’m going to re-test tonight using my power supply.
as a test the 5 second auto drop-down did not happen with a single led XP-L mcbcb i had, i did not test it for long though since i don’t have a heatsink for it to rest on.
I would probably solder a small wire to the pin so that you don't have to try and probe in there. If the battery was 3.6v resting then under load with a triple it could easily sag down to around 3V.
it was experiencing the same behaviour as when i tested with my power supply previously so i didn’t bother trying to use a charged cell. coming home after a long day at work then trying to debug that issue i was not thinking very clearly. :weary:
i’ll use the tiny wire soldering trick, great idea i didn’t think of.
Thanks for making the preprogrammed atiny25 and the pull down resisters available, but now when are the 15mm FET+7135 pcbs going to be available?
Just finished testing it with my bench power supply. set it to 4.2 volts, on turbo the voltage on the supply drops to about 3.6v@the max 5 amps with a reading at a average 0.635v’s at that 7th pin
If it still steps down then there's obviously something wrong with the driver or MCU. Shouldn't be stepping down yet.
When I get some more of them! I only have a handful right now.