Ok pretty sure my issue was a defective ADC2000 CCA. I could hear and feel its relay clicking on and off, at random, on a few occasions today, likely causing compressor stalls (compressor overload trips, to add insult to injury). In essence, this the ADC was commanding the compressor to:
“Start… no, stop… okay, start again… no, overload… wait a minute… okay, try again…”
Compressors really don’t like that.
I replaced the ADC with an EDT11 and everything seems fine for now; a few days of fault-free operation will confirm.
I’ll have have a look at the ADC2000 in a few days, try to find the root cause of its malfunction, at my leisure, on my workbench, with no time pressure, i.e., no thawing fridge to worry about.
Unfortunately, my $15 Kijiji EDT11 was defective; very disappointing. The compressor relay contact was burnt to a crisp, but oddly enough, still making contact (NC contact). However, heat damage distorted the plastic relay head in general such as to prevent the defrost connection to be made, so no output power on the defrost contact, which was incidentally not damaged. My guess is that someone used it on an fridge with an incompatible wiring harness, and without the proper adapter. I ordered a new EDT11 one from ReliableParts.
Got a Paragon mechanical timer, Frigidaire p/n 215846602. Works fine, so will leave it installed in my fridge for now. My fridge is happy. $60 CAD from a private business in Kingston, Ontario.
I also installed a new Start Device/Relay on my fridge’s compressor, as a special treat for enduring the repeated torments of multiple ADC2000 defrost controller restarts (over 100 per day some days) over the last couple of months.
On my workbench, in isolation from my fridge, I could not get to the root cause of the malfunction of the ADC2000 defrost controller, other to confirm that its relay driver is working, its relay contacts are good, both of them, Vdd power to the Microchip PIC16C54C micro is stable (5.6V), even with the relay engaging, and the ‘clock out’ from the PIC is also stable (Fosc/4=32kHz). I re-touched all the solder joints, reconnected R10 (it merely disconnects the bi-metal feedback input) and parked this CCA on my shelf for now. I supposed I could have checked the PIC’s Power-On Reset circuit; will do that later if necessary.
I had to append to this post due to new information. What I was seeing initially, and am still seeing, are compressor stalls. These occur because my thermostat is short cycling, forcing the compressor to start under high head pressure, and the compressor’s start/overload device doing what it was designed to do, stop the compressor under such high loads. Thermostat short cycling was confirmed by the an additional pilot light that I installed at the output of the thermostat and monitored by an Eve motion sensor, which has a light meter in it. So no longer simply looking at their power signature of the fridge, but looking at the output of the thermostat and capturing that with Home Assistant. This sensor is pictured below.
