So I did some experiments and found that indeed that the problem is due to the PLA filament oozing out of the nozzle after the wipe process.
To try to stop the ooze, I did my first cold pull (extrude filament at normal print temperature then turn off the heat and wait for the hot end to get to 90C then yank the filament back out by hand) followed by feeding some ABS through to purge any left over PLA. After doing this, I had success with the auto-level process when I left the ABS in the print head and printed using a PLA profile. I noticed after the wipe that nothing came out so the hot end made a good switch contact with the washer. I let the machine cool back to room temperature and then started a normal ABS print and the auto-level worked again. This is the first time in 4 months that I’ve seen it work twice in a row.
I’m going to order some of that cleaning filament
and try a PLA print next. But, it seems that the problem is too many PLA prints in a row without a clean out or an ABS print such that the retraction can never yank all the filament particles out so they keep leaking which creates an insulator.
Update:
PLA printed fine after the ABS print. A few more PLA prints after that only had one failure that I had to restart. I think for me, my problem is a dirty nozzle interior and not the exterior. I’m hoping a new nozzle or the cleaning filament will make my printer as good as new.
Looking at the schematic of the controller board, the design of the Zmin as used in the Mini for auto-leveling is scary from an ESD (Electro Static Discharge) point of view. The nozzle is tied directly to the micro-controller pin and the washers are tied to ground. If you rub your feet against a carpet and touch the nozzle (with the power on or off) you will send a very high voltage to the micro-controller pin. It would have been nice if the head included some fast switching diodes (to +5 and ground) to help the little diodes in the pin of the micro-controller. The controller is using a weak pull-up feature (behaves like a 40K Ohm resistor to +5V) to float the head voltage to +5V so that when the nozzle hits the washer, a zero can be detected on the pin.
The last sentence means that you don’t have to trace wires and check continuity to see if the wires are broken. Just power up your mini and stick a volt meter between a washer and the nozzle. If your wires are not broken, you will read about 5V (I read 4.96V). Also, the M119 command can be issued with and without a wire shorting the nozzle to a washer. Zmin should return “open” when not shorted and “TRIGGERED” when shorted.
I think it is insane to buy a $120 controller board if you just blow the input pad to the micro-controller. It seems to me that you could just wire Zmin to Zmax (pull the Zmin connector, strip the red wire and short it to the red wire of Zmax and hack the firmware to look at Zmax whenever the code is looking for Zmin. I have not looked at the source but it seems possible. A harder hack is to connect Zmin to an unused input and change the firmware to use a different pin.