EEZ H24005 "CH1 IOEXP failed"

[mrnuke]

I recently got my crowdfunded EEX H24005 power supply, and when I first turned it on, The LCD gave me this error message:

CH1 IOEXP failed

I tried swapping the power boards to see if the fault would move to CH2, which it did not. I plugged in a serial console, and I get the following messages:

**TRACE [2088-01-31 21:57:09]: Ch1 PWRGOOD bit changed to 0
**ERROR [2088-01-31 21:57:09]: -242,"CH1 fault detected"

My suspicion is that the fault is on the front panel board. Other than that, I'm not too sure where to start looking. Has anyone else seen this? Could it be a firmware bug instead?

[peteb2]

At this stage i would do a careful check of the interconnect looms, even going so far as buzz for continuity on each line, and then verify their connection location device (socket/plug) onto the requisite board has been soldered to the board properly and that there hasn't been any tiny 'production' intermittents like a solder bridge etc... What is handy is there are 2 channels which lends itself to making comparisons and in the practical approach something as simple as using the diode function on your DMM... I apologize if i sound like i'm teaching Gran how to suck eggs here, but i'd 100% verify the loom connects are pushed home totally and orientated correctly. This is just the usual approach for any kind of 1st stage faultfinding i guess...

I am not 100% certain but i think all the boards have been pre-tested ... but maybe i am not right on that, instead a high number, randomly selected from the production run were.

My EEZ PSU powered straight up without any dramas and once calibrated has already been put to work during a few sessions of lab/prototype work. I repair high end gear as a living as in cameras for news crews in TV and believe it or not once and a while tiny production errors don't show up until a brandnew unit has had some work in the field like a single ball of solder smaller than this ". " left from production dropping between the pins of a big FPGA and causing intermittent crazy random picture colour failures..... (that was a hard fault to find)...

It'll be neat to hear how you find and where the fault lies but i also know that annoying feeling something you've invested it aint working right can be.

[prasimix]

Hi, peteb2 suggestions are fine. So, looks for some dirt, bad connection on pin header X13 and IC19 (you can try to resolder all pins). I've tested all boards (100%!) one by one, on my test configuration and all passed basic testing, i.e. firmware selftest procedure. It's also possible that X13 is not cleaned up properly. This is definitely not a firmware bug, because all boards are preloaded with (now obsolete) v1.0 that works fine.

[mrnuke]

I've tested all boards (100%!) one by one

Looks like a faulty solder joint on IC20-14 (CH1_RESET) was missed. That's ~CH1_RESET, which wasn't pulled high, thus keeping the IO expander in reset. Took a while to find, but now everything is working fine. Thank you for the suggestions.