Last week, we took our two Duracell 29HM deep cycle marine batteries and the MicroSolar 24v inverter from our hurricane ground solar experiments and added an AC charger, essentially building an oversized UPS for our freezer. We’ll talk about that system more in detail once we’ve run it for a while, but this morning it appeared that this system had failed. Although this turned out to be a false alarm, the steps we went through to verify the continued operation of the freezer and its power system are good to know.
The first indication of trouble was that the thermostat tripped about 8:30 AM, with a loud click. This is unusual for this freezer, since we keep it at such a low setting that its thermostat hardly ever trips, especially in the semi-heated room it is in. Even when the thermostat does trip, it doesn’t stay off for long, maybe a half hour. Around noon, it still hadn’t turned back on, so we opened the panel and checked for anything obviously wrong. Nothing noted.
Now, when a freezer or refrigerator goes bad, especially when running from an off-grid inverter, the first suspect is the compressor motor. That is a bad thing, as repairing it usually means buying a new freezer (with an exception we’ll show in a future project). A burned-out compressor would be a nasty surprise, especially since we had only recently raved about the quality of the inverter’s AC waveform.
The other major suspect is the thermostat itself. See the picture to the left for a view of the thermostat, including the sensor tube leading to the freezer internals (click the photo for a better view). While this wouldn’t be nearly as a bad as a compressor failure, and we could replace its functionality with a Raspberry Pi and a temperature sensor, replacing the thermostat as-is would be a painstaking task.
As you can see from the blown-up photo, the gray pressure tube leads to a larger copper tube which then runs into the internals of the freezer. Rather than sensing temperature directly, the thermostat senses the pressure of the gas in the tube, which is related to the temperature of the tube. If we were lucky, the copper tube would be a sleeve that the sensor tube passes through like an oil dipstick. If we were not so lucky, the sensor tube would be a pressurized extension of the copper tube, and replacing it would not be trivial. We wouldn’t know for sure until we cut through the red plastic cap protecting the junction.
Before cutting through anything, we decided to isolate the system components and see where the failure was, and the best place to start is the compressor. If that was out, it would have been time to take the truck to a big box store and get another freezer. First, we unplugged the freezer, then unplugged the black and yellow AC lead wires from the thermostat. Then, as shown below, we connected an ammeter across the AC leadwires, and set it to the milliamp scale.
We plugged in the power cord and success! The compressor started running, and the ammeter showed 880 milliamps, or about 106 watts. This is about a typical startup current until the plumbing stabilizes. Plus, after a couple of minutes, a thermograph of the side of the freezer showed the sides heating, indicating that the compressor was doing its job.
Next to check the thermostat. To do this, we unplugged the freezer to let it start to warm. We could just have checked the thermostat every now and then to see if it had closed again, but more fun would be to let it tell us. We had an unused security system siren lying around, so we wired that and a 12v UPS battery across the thermostat terminals, as shown below:
Knowing that the compressor and the thermostat were fine, and that the charger-battery-inverter UPS hadn’t damaged anything, we put it all back together again, and restored power with the freezer UPS system.
So what did happen? Mother Nature. Last night was cool, but not cold or hot (a rarity for this area), and the same for this morning and afternoon, and so we didn’t need to run heat, nor was the outside warm, although both indoors and out were a little bit chilly. The decay curve of the freezer’s thermal mass, combined with the thermostat hysteresis, meant that the freezer could legitimately stay off for many hours and not activate the thermostat the entire time.
Lesson learned, and we now have more confidence in the inverter system, although around noon it was looking really touchy.