That’s what I’m thinking. The last step before it does the two big jumps is reading at around 380mA so it is definitely only getting to 16 steps, then jumping to half turbo and full turbo.
Using an Efest 35A that’s a bit over 4v at the moment so I don’t think there is a problem there.
Indeed there is a problem with that PWM circuit---it's shorted to ground. Some of the drivers develop a short between the center 7135 pin (ground) and the PWM via that sits directly underneath it (design error). Removing the 7135 and then cutting the center pin off fixes the problem. After I started getting a few reports of the problem I started checking every driver for the issue, and on future drivers that I build I am going to either cut that 7135 leg or insulate it better, so there shouldn't be many, if any, issues going forward.
If any of you experience this problem (I've had a few others report this same issue, but I had them send the driver back because I hadn't figured it out yet) here is the 7135 leg to cut:
You can easily diagnose it by checking for continuity between the PWM pin (right pin) and ground pin (center pin) on the spring side 7135 chips.
As usual Richard, you were 100% correct! Thanks for your help. I removed the 7135, cut off the centre leg, added a tiny piece of kapton tape to the board then bent the two remaining pins slightly down so they still had contact even with the kapton tape underneath. Battery back in and all is working perfectly! Thanks again!
Has anyone used this driver with an mtg2? If used on hi, will it regulate to a good usable level? I’m making a light for my grandfather and I don’t want him to have to worry about remembering how to deal with temperature.
From a hardware standpoint, couldn’t you add a zener to dump the voltage to the mcu and swap the protection diode with a resistor? Like a qlite driver… I might mess with the complicated firmware though?
You could zener mod it, but you'll burn up the 7135 chips, especially the single channel 7135 chip, in a hurry. Not recommended. The duty cycle would have to be greatly reduced on the single 7135 chip, which would require a firmware change.
I depends largely on the total resistance of your circuit, but with a good bypassed tailcap spring and GA I would expect around 5A. You might get 0.5A-1A more from a 30Q.
Yeah, the spring would be bypassed. Hmmm…that doesn’t seem like it would be a huge visible impact. That being said, 30Qs are cheap enough to pick up down the line and see for myself.
So how does the thermal regulation work? It steps down a full level (of the 24 programmable modes)? Or it just dynamically decreases the PWM to maintain their thermal target but can do it in a far more granular fashion than the programmable levels?