Amrel/Motech PPS-2322 dual programmable power supply weird turn-on waveforms

[molasses]

I've had this power supply for a while and could never figure out why it gives such ugly turn-on waveforms (first image). Only recently I've had some time to tear it apart and see if I can fix it.

My first thought was to look at the voltage and current setpoint signals, in the second image. Top to bottom:
1. Voltage setpoint from 4066 output D
2. Current setpoint from 4066 output A
3. DAC output voltage into 4066 input
4. Voltage enable signal into 4066 enable D

The setpoint signals are generated by the DAC, routed through a 4066 and then held by a capacitor for the CV/CC control op amps. When the output is switched on the setpoint waveforms have a noticeable glitch which matches the glitch on the channel outputs. I also notice the DAC output voltage looks very messy, not sure if that's normal. It's the same for both output channels, which are fully isolated from each other, they each have their own DAC and switch.

It's weird to me is that both isolated channels would have the exact same problem. That would make sense if there were an issue with the digital board that controls both of them, but there's not much to go wrong there, at least I think so. There's just the MCU, one GAL16V8 CPLD and the optoisolators there.

So I'll keep thinking about it but I'm not sure where the problem is. Would a DAC or analog switch failure cause this? Would it be likely for those to fail on both channels independently?

It would be very helpful if anyone has one of these Amrel/Motech units and has some idea about what might be the problem here. Or if you just know these sorts of digital controlled power supplies in general.



Another user here reverse engineered the relevant parts of the DAC and 4066 switch circuit: https://www.eevblog.com/forum/testgear/motech-amrel-pps-1002-diagramschematic/msg3605273/#msg3605273

[molasses]

I have some notes that I figure I'll add here so I don't forget them myself.

The constant current limit kicks in on every startup if the voltage is too high. The start up waveform is the current controller taking over and limiting the current while charging the output capacitor, which probably should have been obvious to me. That checks out but the behaviour is still not as expected, as adjusting the current limit upwards has zero effect on this. This is because the supply first programs the setpoint voltage, and then a few cycles later the current. So while the capacitor is charging, the setpoint current hasn't been set yet and the supply is limiting the current at 0A. (clearly, some current still flows due to slight offset error).

I tried another firmware (http://www.kerrywong.com/2018/01/07/amrel-pps-2322-programmable-power-supply-firmware-backup/) for this power supply and it had the same behaviour. However, the one posted by eblc1388 is a later revision. This time it programs the current first, and then the voltage. As a result, you get clean power-on curves unless you attach a very large load capacitance or set the current limit very low (<50mA or so). This makes a lot more sense to me. Maybe they realized this problem afterwards and fixed it.

Files attached to this post:
1. Firmware of the PS BDY-1204FCE and EEPROM data