The steps you describe all occur. On power-up the unit fills and rinses then fills and begins freezing. The freezer plate gets plenty cold and the first bit of ice begins to form in about 5 minutes. Then the unit stops pumping and the water is drained. This is exactly as it works when operating correctly except this freeze cycle is halted prematurely. It is as if the unit has decided that it is time for ice to be harvested.
My first suspect was the evaporator thermistor. Original tested fine and new replacement did the same.
I then tried running without the thermistor connected (open) and the unit did not default to a timed cycle so I thought it was the PC board.
New PC board does the same thing.
Does the water stop pouring over the freezer plate prior to the hot gas valve coming on? In other words, after the fill from the water valve into the reservoir, the pump should be off until this fill stops, then the water pump should turn on. It should stay on for around 20 minutes. If the water stops, it should happen at the same time the gas valve comes on to warm the freezer plate. Do these two things occur together, or is there a bit of time between them?
I don't recall if they occur exactly together, but I think it is at least real close. I should also mention that this failure was not total and sudden. I noticed a reduction in icemaking and then as the failures progressed over days, I was actually losing ice. Occasionally the unit would perform perfectly for hours or sometimes days on end.
The filling seems to be OK. I have checked it pretty carefully. On this unit it appears that it is filled based on the selected ice thickness chosen in the programming. (It was like pulling teeth just to get the info on hoe to program ice thickness. This is done by the user interface and not with jumpers) It goes all the way to the top (just short of overflow for "thick" and the less for normal and thin. It appears to be regulated by a hall effect probe (touch sensor?)
I didn't find any corrosion but cleaned all connectors anyway. The strange thing is that it doesn't even default to the "timed cycle" upon disconnecting the thermistor.
The freezer plate does get VERY cold and appears to be working.
The unit has been kept very clean and dust buildup has been minimal.
After studying the wiring schematic for way too long, it seemed that the only remaining feedback to the control board was the 'overload' circuit. Would that tripping appear like a premature ice harvest or just shut down? The compressor doesn't seem to be straining and isn't making unusual sounds.
If the evaporator is getting cold, you have no issues in the compressor system. How thick of a slab does this make before the unit releases it?
If it is working correctly, about 3/8 to 1/2 inch.
With the failure that I'm having, the ice is anywhere from non-existant to paper thin. Usually it all melts with the hot gas release.
I will disconnect it again and try it but please understand that I am now into this thing well over 50 hours. This is the second thermistor and they both check out good using an ohm meter and ice bath and such.
While we were waiting for each others replies, I timed two cycles from end of fill/start of freeze - to "early harvest".
The first cycle was 7 minutes 49 seconds.
The second was 8 minutes 15 seconds.
Both had a barely detectable sheet of ice when they stopped.
I have never been able to get either PC board to go to the 25 minute cycle. They both do the same thing.
What happens if the overload relay fails? Can it look like this?
How do I contact you in a few days?
I have spent another 10 or so hours since we communicated.
I followed your advice and checked all the harnessing (even ohm'd it out and checked continuity). The connectors all looked perfect as in factory new.
I made several more attempts to get the controller to go to the timed cycle (25 minutes) with no luck. Even went to the extreme of completely removing the evap thermistor from the ice machine and locking it into a ziplock bag (not just unplugging it). That is how crazy this has become. The unit persisted in going into an early harvest anywhere from 5 to 10 minutes into the freeze cycle.
I then found a descrepancy between the two side by side wiring diagrams on the wiring data sheet.(W10163556A) on the left hand version the wiring color code for the evap thermistor is shown as RD/WH. On the right hand version it is TN/RD. I thought that the crisscross if bin and evap had to be a major clue. Again, no dice.
I went back to square one and ran the self test diagnostics again, and again everything checked out. It was then that I noticed that the test for the water level 'touch sensor' was correctly showing that the reservoir was full. In the diagnostic cycle the reservoir filled during the test. For some reason I decided to run the touch sensor test again with the reservoir empty. It still showed that the reservoir was full so I then suspected that the touch sensor may be faulty. I have had some experience with hall effect sensors and I always thought they were a little squirrelly.
Very long story way shorter, I eventually disconnected the hall effect 'touch sensor' and now everything is working perfectly.
It appears that the filling of the reservoir is handled completely by the smart valve and even compensates for the selection of cube thickness.
I was planning to order a replacement touch sensor (hell, why not? I've replaced everything else!) but it is pretty expensive and I can not figure out why it is needed. So I guess the only remaining question I have for you is: Why do they even have it in the system. What does it do? and should I replace it?
I wouldn't even ask you this except Kitchenaid and Whirlpool refuse to be of any assistance unless I am an authorized tech (and even some of the 'authorized' techs I've talked to get stonewalled). So I figured you might have access to a decent answer.
That is what I originally thought, but I'm finding out that reliable information from these companies is difficult to come by.
The reason that I asked the question is because the fill cycle is NOT time dependent with the smart valve. The fill is running about 30 seconds or so (not 2 minutes) and changes in line with the cube size selection. It is NOT going into overflow for the ice production fill.
The first fill or rinse is filling to the top and will slightly overflow. I would expect this if setting up for a complete rinse including reservoir. However when it fills for ice production, it appears to be carefully dosed so that it will not use extra energy chilling more water than is needed for the selected cube size production.
I don't have any more questions.