Ich Liebherr dich

Living in a rented apartment for over a year was for us quite an experience after being used to the comforts of a very own flat. There are lots of things that aren’t “just right” in Dutch small rented places and lots of them, including the one we are currently occupying, only have a small fridge without a freezer. After some back and forth thoughts, we finally decided to get a standalone freezer for cheap so that we wouldn’t regret too much in case of moving to the next place. And here it appeared, an old and rusty GP1466 Liebherr for ‘only’ 160€ and almost in our city. Who could stand this offer? I couldn’t and it seemed like a great idea at the moment.

With a friend’s help we got the thing delivered and brought up through the narrow and curly Dutch stairs to our apartment. After turning on it did start cooling (that fact I also had a chance to check beforehand) and we happily made a ride to Ikea for who-knows-how-many kilos of frozen meatballs. But after our return it only showed -3 on the display. The night did not change a thing and the compressor was hot and still running, but it did freeze a cup of water pretty rapidly so we decided to buy a thermometer to double-check. And according to the thermometer it was less than -25 inside. Ok, so all of the refrigeration components were working properly and the problem turned out to be in electronics — either the thermostat module or the temperature sensor.

The only thing that I could get out easily was the module:

When debugging electronics the first thing you do is checking the power supply. This was an instant Bingo because it was clear that the supply voltage drops for a half of a volt as the blinking digit on the display turns on.

The circuit uses a capacitive power supply which, in basic words, makes use of capacitor impedance as a kind of ‘AC resistance’ to drop all the unneeded difference between 220V input and 5V output. The particular part used here is of self-healing X2 class kind. Those ‘safe-healing’ caps are known for a very specific failure mode: when subjected to voltage peaks (like EMI from lightning), they do not usually fail completely, but rather burn-out partially without visual consequences. The only observable difference this failure causes is drop of the capacitance.

After desoldering the cap to check it, it appeared that this was exactly the case, 0.19uF instead of 0.68:

I could, in principle, wait till Monday and buy a new cap in a components shop, but I do enjoy solving problems with what I have at hand at the moment. Couple of months ago I found a (broken?) printer on a street and I knew that there should be a similar cap at its power supply’s filter, so here it goes:

Using the old cap and the one from the printer in parallel has almost done the trick and the supply voltage stabilised, but.. it did not solve the temperature indication problem.

Then it must be a temperature sensor, right? Those freezers use resistive NTC sensors, which means that their resistance increases as the temperature drops. I’ve even found a table os resistance-temperature correspondence for the Liebherr sensors. But according to the measurements it seemed kind of in the proper range:

So I spent quite a while trying to figure out what does the thermostat module show depending on the voltage present at the sensor input. Those experiments turned out to be unproductive because the board does not indicate what it senses as the current temperature, but rather shows a filtered (something like 1 hour — average) value. Yet in the process I did discover that when working in the actual circuit the sensor behaves non-linearly: like its resistance drops from the measured values when the working current flows through it. That was a cunning failure — the one that does not show itself during measurements. Most probably it can be explained by water condensed in the sensor through a crack in insulation.

Now I was in need of a new sensor. The Liebherr parts page did not prove helpful, as well as their tech support. Basically the only piece of information I’ve got was that the sensor replacement kit costs 80€ and I’ll be lucky to even get one because that’s an old model and blah-blah. I was not going to pay half-the-price of this freezer for a thermistor and this time I reached the shop to get just a generic one for some cents.

The tricky bit was to find a way to install the new sensor. The sensor wire was going somewhere to the internals through layers of foam insulation and the rest of the freezer’s body was non-disassemblable. Luckily, Italian guys from this forum were extremely helpful and dug out scans of the replacement instructions.

Now, what I needed to do to replace the sensor was to locate a specific spot inside the freezer and cut out an opening with a knife (yep, not something I’d describe as repair-friendly):

Now as the sensor’s wire is reached, it’s easy to cut it and solder in a new thermistor.

Heat-shrink tube stuffed with hot-glue seals it from moisture and the sensor may be placed back under that piece of plastic.

And it works!

The calibration is somewhat off, but for my target of -20 it is just a couple of degrees. If someone wonders, the thermistor I’ve used for replacement.

This may be not the prettiest reparation out there, but now we are celebrating our happy freezing and do not need to bring more weights up-and-down the stairs. And the old thing was saved from the dump, which is good for the environment, right?