[Mr. Carlson] Fixes A Fridge

A dead refrigerator is an occurrence determined to frustrate any homeowner. First there’s the discovery of hundreds of dollars in spoiled food, and then the cost of a repair call and the delay of the inevitable wait for parts. It’s clear to see why a hacker like [Mr. Carlson] would seek another way.

Now, normally a fridge repair video would by unlikely fodder for a Hackaday article. After all, there’s generally not much to a fridge, and even with the newer microprocessor-controlled units, diagnosis and repair are usually at the board-level. But [Mr. Carlson] has had this fridge since 2007, and he’s got some history with it. An earlier failure was caused by the incandescent interior lights welding relay contacts closed thanks to huge inrush currents when starting the cold filaments. That left the light on all the time, heating the interior. His fix was a custom solid-state relay using zero-crossing opto-isolators to turn the bulbs on or off only when the AC power was at a minimum.

That repair kept things going for years, but when the latest issue occurred, [Mr. Carlson] took a different tack. He assumed that a board that has been powered 24-7 for the last twelve years is likely to have a bad capacitor or two. He replaced all the caps, threw in a few new relays to be on the safe side, and powered the fridge back up. It whirred back to life, ready for another decade or so of service.

Kudos to [Mr. Carlson] for his great repair tips and his refusal to surrender. The same thing happened when his solder sucker started to give up the ghost and he fixed it by adding a variable-frequency drive.

27 thoughts on “[Mr. Carlson] Fixes A Fridge

  1. Wonder how hard this would be to reverse engineer and replace the board. I bet the hardest aspects would be tunning defrost and cooling profiles. Cooling would be doable with some watching but defrost is more complex (and rarer AFAIK).

    I replaced all the relays on my washing machine when one got stuck.

    Then it turns out buying the cheapest relays on ebay was a bad idea and they all got stuck ruining the two main boards and elements. So I ripped the motor out and it still sitting in my shed.

    1. LOL – I work on these boards all day everyday. Most of these boards use optocouplers to power the relays already (some use darlington arrays); this guy went out of his way to not fix the original problem and work around it instead ($3 fix, btw).

      Replace the capacitors on a 12 year old board? That’s the cotdamn bare minimum. Come back to me when he figures out how to repair the SMPS after it has catastrophic failure and the parts are a Korean off-label NEC spec that you have to dig through datasheets to find a match for.

      1. Been there and I have fixed worse with only what I have on hand.

        I had to repair the HV supply on my scope once. The transistor driving the fly back was obsoleted. The replacement from that datasheet as well as its replacement were also obsoleted. It was that bad I checked the datasheet of the power transistor from a blown silver box PSU that comes with a $20 case and it was close enough.

        That got the flyback primary side sort of working, but the circuit seems to be overloaded. I thought the flyback shorted, so I disassembled and rewinded the secondary side with the slightly thicker wires that I got. My reasoning was that it was a hot day and got overheated when it died. Still no go.
        Finally found the problem was 3kV ceramic cap. The only 3kV one I got was from my dead CRT but the value was too low. The PI filter on the scope had the same value, so I swapped it with the damaged one and use the lower value for the filter.

        That fixed the scope. I bought some extra 3kV caps just in case, but it had been working for the last few years.

        1. That’s not actually the purpose. The isolation (on this board type, anyway) is meant to protect the rest of the board from flyback voltage when the relay disengages. On refrigerators, the coils are usually powered in groups by a single PNP @9-24VDC, while the individual relays are usually “grounded” through the arrays or optos. Powering a SSRelay with an optocoupler seems like overkill for a light circuit, but… y’know.

  2. 21st cty fridges seem to be made for repair techs more than consumers that they let the fridge die without letting us know except spoilt food. Design does not seem to account for large power spikes and could let us know when most of the components die although I would not want grandma ripping into the guts to replace a thermistor.. still very frustrating tech.

    1. I’ve had my two fridges for 13 and 4 years. They’ve never had a fault other than replacing the bulb.

      So I am very confused by this comment thread although I assume there is selection bias and those who are not unlucky with fridges aren’t here to complain!

        1. On closer inspection it’s 12 and 6.

          But I suspect this is a survivorship bias.

          Here is an intesting summary of a Google study (looking at hardrives in data centres).


          The utilisation AFR best shows what I mean.

          Some drives will be “duds”, some about average and will have a moderate survival bunching around the median lifespan. But the mean lifespan may be much higher.

          I.e. if a device has worked for 10 years, it’s not likely to fail in the next 1year. But one that has worked 1 year is far more likely to fail in that year!

          This means 30 years down the line we see a very different picture to when the fridges were made.

          It’s not co-incidence it feels like warrantees last right up until the appliance dies. That’s the stable median life span. The mean lifespan will be overwhelmed by the small number of devices that keep trucking.

  3. I posted a link to this video on a fridge repair forum. I got lectured by a moderator that since component-level repair doesn’t have the blessing of the manufacturer, that’s bad. Not sure why he cared about what the manufacturer thinks if it’s an outside of warranty repair. Only thing I could come up with is that it could cause some trouble for those that are “authorized repair techs.” I also learned that Whirlpool refuses to provide service manuals. Frigidaire is a more more repair-friendly and has service manuals published on their site.

    1. Oh my, that’s a forum I’d steer clear off… lecturing people who do component level repair on a Mr. Carlson level just doesn’t make you look like you know much of what you’re doing. It certainly does not make you look like any sort of expert I’d want to put my trust in… if they call themselves repair technicians, but are afraid of doing actual repairs on an out-of-warranty product, they don’t seem to be much into repairing stuff after all.

      1. Why have your client spend $12 on parts for a great repair when they could spend $120!

        I’d understand if the idea was that the difference is made up in labour but doesn’t apply here.

  4. Modern fridges are made to sell to hive-mind consumers ( lots of profitable “features” and shiny bits, but with no quaity control). So long as they properly contrast the granite countertops, it’s all good.

    I recently gave away a small Italian-made dorm fridge that I bartered in exchange for a stereo repair…in 1982. It was still running perfectly.

    My current domestic fridge failed after a few years, and when I pulled the board, half the solder joints…weren’t. Really. They were missing altogether. 20 minutes with a fine iron and it ran again, but with manufacturing now consisting of contractors assembling subcontracted components on behalf of a “distributed by” brand, some of the results have been horrific.

  5. My maternal grandparents had an ancient Amana freezer in their basement. They moved into the house in the 1950’s and the freezer was already there. It ran continuously (except for power outages) until 2005 when we cleaned it out and turned it off after my grandfather died in 2005.

    While on the outside it was larger than a modern 20+ cubic foot upright freezer, the door and walls were nearly a foot thick. The door was made like a bank vault door with the sides tapering in and fitting real close to matching tapers in the walls. Those sloping parts were black Bakelite sheets while the inner face of the door and the compartment lining were galvanized steel. All the inside was held together with a ton of slotted panhead screws.

    I dunno what’s become of it since the house was sold. Maybe it’s still down in the basement. After the back porch was converted to living space the the back stairs’ brick walls were covered I doubt the freezer could be removed without breaking it apart.

    I’d love to have that freezer, or one just like it. It would be interesting to do an efficiency test on it. Then I’d modify it to use a magnetic seal instead of the latch. Not that anyone could get locked inside due to the small space and fixed shelves.

  6. Plastic that cracks and breaks is the other big issue! We bought a name brand Korean 4 door fridge 5 years ago… middle cold-cut drawer pulls out, inverter drive compressor. Feature rich. Expensive! Functionally, we like it. But within a year the first plastic door shelf started cracking and failing. They used a HIPS material that is BRITTLE and seems similar to hobby model car plastic. Family of 5 pulling out and replacing the milk is a common activity. In Canada, that is a 1.3L milk bag in a plastic container. One of the glass shelves has this HIPS plastic in the rear… that broke apart too. No support from *@#SUNG stating that customer abuse broke these. I’ve reassembled the broken parts with a combination of plastic epoxy or reinforced with sheets of Styrene from the hobby store, cut and welded in place using MEK. No way I’m paying $120 per shelf for replacement. They have held together since. I do worry how long the electronics and compressor will last… good so far at year 5.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.