D1 doesn’t look quite well soldered next to pin 8. Looks skewed off the pad.
Might not be the caps making it do that, but they ARE backwards. And it does look like the diode might not be making contact on the MCU side. They tend to want to wiggle around when the solder melts if you have a little too much solder, and it’s really easy to get a little too much with that tiny thing!
Your 1912 should be a 223 according to the way Wight designed it. I’ve been using the 223 and they work beautifully. Have to change ADC_LOW to 124 and ADC_CRIT to 112 for a 3V warning and 2.8V shut down in the Low Voltage Protection. 
It also looks like the 7135 above Alex’s name is blown.
DBCstm, if those are backwards so are the ones pictured in the OP? I actually haven’t run that driver, so maybe they are…
That said, I wouldn’t expect the driver to function at all with a 1uF decoupling cap. Likewise if the diode was not making good contact, I’d expect no operation, not 0.38A operation.
I’d say that it’s time to get the DMM out. Get this driver wired up on the bench turn it on and switch to “high” mode and we’ll do some checking.
- Check for voltage on the Vdd pin of each 7135.
- Check for voltage on Pin 5 and Pin 6 of the MCU.
- Take Vbat (solder on an extra little wire and hand-hold this) and directly apply it to the non-functioning 7135’s Vdd pins.
The Vdd pins are the ones wired back to the MCU’s PWM outputs (Pin5/6). Refer to this datasheet’s first page if you have any confusion, the pin is clearly marked: http://lib.chipdip.ru/709/DOC000709557.pdf
The caps are correct, bigger one is C1, which is “on top” or on the outside edge, the smaller one is OTC which is in the middle.
Man, can’t trust my recall at all anymore!
And like you say, if the diode wasn’t making contact it wouldn’t work at all.
Sounds like the firmware has something commented out that isn’t supposed to be, only allowing the single 7135 chip to work.
Edit: The PZL, the layout is under the MCU. OTC is 1uF, C1 is 10uF. C1 is largest.
So yes, the caps are wrong.
I thought that was odd, seeing it like that, but I’ve only built like 6 or 7 of em. Easily confused, I am! 
I lay out my components and leave the C1 cap in the tape. I put the OTC on the board before C1, then open the tape and put that one on the board. I always leave that 10uF in the tape til all other components are populated. Then it’s not a matter of me identifying 2 caps that are close in size, it’s ID’d by being in the tape, cut off the string of 100 from the package and identified one at a time as each component was laid out.
Incorrect. I just double checked. My 10uF caps are physically smaller than my 1uF caps. I also went over here and looked at bdiddle’s caps (#46). Bdiddle’s C1 is also physically smaller than OTC.
Grantman321’s caps are not backwards.
My 10uF’s are larger, visibly larger and measure larger, than the 1’s.
I recently bought TDK brand from Mouser
The 10uF measures 0.05x0.05x0.08
The 1 uF measures 0.05x0.05x0.07
The end caps are thicker, giving the center brown portion less area, on the 1uF. That’s the only real visible difference I get.
That’s why I keep one in the tape, so it’s not about me identifying the caps.
Beats me, I’m just a hack who assembles off a list, I don’t know what all this carp is….
My OTC are absurdly larger than C1’s. It’s the tallest component on my drivers.
There’s got to be a reason, different voltage, different percentage, something.
No big surprise really. It probably depends on brand, product line, and specification. More stable stuff may require more space or it may require less space (but more money).
The main thing to pay attention to is the fact that grantman321’s MCU actually runs. I don’t think that would happen with a 1uF decoupling cap.
I checked and the ones I have now from mouser aren’t longer, but they are more square and wider/taller. I think the last OTC’s I got from Richard were smaller.
Then it almost has to come back to the firmware.
Grantman, can you post a link to your C file so it can be looked and and checked for continuity? (AS IF that would be me doing the checking! lol)
Maybe the main PWM pin on the mcu isn’t connected? Just the alt pin?
Or the FW is configured to just use the alt output?
Looking that over, I don’t see the changes made that Grantman says he made…
“dual_PWM uncommented, dual_PWM_start uncommented and left at 8, turbo rampdown basically stock, high is commented out still. After looking at the firmware over and over, I’m relatively convinced it isn’t the problem as I didn’t change much besides uncommenting the obvious dual_PWM related stuff.”
I know it’s easy to mess something up by now catching a related component later on in the program. Could it be something like that? Something that was changed in one place and not matched in another?
If it’s hardware then one 7135 pwm pin is shorted to Vcc. Simpler to check than firmware and pretty much has to be the solo chip. If it’s firmware sayonara sahib, adios bwana, later gator.
Unless I misunderstood, grantman321 gets modes of up to 0.38A, so nothing is shorted to Vcc. It’s possible that one of the 7x group of 7135’s has it’s PWM pin shorted to GND, but not Vcc.
That is correct. Modes are working as they should - just seems to be running off of one 7135 only as it’s pulling 0.38A on the turbo mode.
I’ll be able to take some measurements with my DMM on the driver later on tonight. In the middle of this my trusty Stahl soldering iron died. Got a new iron yesterday so I’m back in business.
Ah, get it now. Wouldn’t that indicate that the single chip pwm channel was functioning normally and the multi chip channel was not? I suppose it’s possible to have bad connections on all but one chip but it seems more likely that either the other pwm pin is not connected or the firmware isn’t generating the signal on the pin. Sorry, if I’m not helping I can just butt out.
I agree, my first assumption would be that it’s a firmware problem. At a glance, the FW looks OK… but I did not scrutinize it closely.
In order for the primary channel (the 7x channel) to stop working only one 7135 needs to have it’s PWM pin shorted to GND. Then they’ll all be shorted to GND, together.