iMX8MP not powering up

[ninevoltz]

I just got my iMX8MP and EVB yesterday. I plugged it in this morning with a 5V 2A center positive wall wart. The power LED took a bit of time to light and came on dimly at first. The LED is bright now, but I get no power to the CPU. There is 5 volts at the test point on the SOM board, and 1.8 volts on NVCC_SNVS_1V8, but no power anywhere else. I've tried pressing the power button, and nothing happens. Any troubleshooting ideas? I can only assume that the strange behavior of the power LED points to there having been a short-circuit somewhere on the boards.  Measuring from VDD_SOC1 test point to ground I'm reading 0.164 ohms. Is that normal?

[LubOlimex]

Each SOM and each EVB board are tested after manufacturing, the test includes Linux boot.

First inspect if they are properly connected and aligned together. It is possible to misalign the connectors, if you decide to disconnect them be vary careful and don't disconnect the two boards in one go (this will probably rip connectors). Then tell me:

1. Did you prepare an SD card with the official build root image?

2. How do you access the board (via USB-serial cable or how)?

[ninevoltz]

Yes, I checked the header alignment and it was always correct. I have an SD card with the Olimex image written to it. The PMIC is trying to start up but it looks like it's faulting out while trying to start Buck1. Should I be seeing 0.1 ohms from VDD_SOC to ground? That doesn't seem right since 800mV into 100 milliohms would be 8 amps. I suspect one of the dozens of bypass capacitors may have cracked and has an internal short, or something got between some of the BGA balls. I'm not sure how I'm going to narrow down where the short is without doing further damage.

[ninevoltz]

I fixed it. I removed the inductor L3 from BUCK3 on the PMIC and it's working now. I don't know if I'll have problems later, but if I put that inductor back in circuit I get a dead short again. It's bizarre because there's no short between BUCK1 and BUCK3 or between BUCK3 and ground UNTIL I insert the inductor. I can't make sense of it.

[ninevoltz]

I take that back, there is a short between BUCK3 output and ground. Either the PCA9450 failed internally, or there's a short in the board around there. Next time I order from Digikey I'll buy a replacement for the PCA9450 and see if it fixes it. I'll let you know if there's any further developments.

[LubOlimex]

Probably there was short circuit between LX3 #31 and #32 and then when you removed L3 it no longer affects the board, but now you are left with one DCDC instead of two, so less current available.

See VVD_SOC_S uses two DCDCs for more current.

[LubOlimex]

I have no idea what went wrong but this shouldn't happen on first power up. My advice is to double check your setup, consider the power adapter might have failed somehow, exclude static electricity, and conductive surfaces causing short. Else this might happen again.

Contact me over support@olimex.com to discuss the option for return.