Luc Gommans/ blog

Is my no-frost freezer malfunctioning or just defrosting? How cold is too warm?

Written on 2020-09-22

It's all over the internet that your freezer should be set to -18°C / 0°F. But then the no-frost kicks in and it heats up to -8°C for 30 minutes. Is that okay?

Very few websites give answer that question. To make matters worse, people on forums would have you believe that, indeed, water flows only above zero and so defrost means heating periodically till above zero. This is nonsense. Your freezer products should remain near -18°C, and in order for that to happen for every kind of product, the surrounding air may not get significantly warmer. It's the cooling elements in another part of the freezer that are thawed in a no-frost freezer.

Table of contents

Is periodic heating to -8°C okay for a no-frost freezer?

The answer is no. Figuring that out was a huge rabbit hole involving papers from before the Internet, lots of people that are wrong on the internet, and a mechanic coming by to check if we have a warranty case or are just being paranoid. A German website[3] refers to a paper from 1964, Wikipedia[4] bases the same claims on papers from 1953 and 1969, but none of these appear to exist online. According to Liebherr, 1989 EU legislation is based on the 1964 paper. The legislation indeed claims[1] that "at -18 °C all microbiological activity likely to impair the quality of a foodstuff is suspended". Quality, here, seems to refer to safety specifically because USA's FDA specifies[2]:

Though food will be safe indefinitely at [-18°C], quality will decrease the longer the food is in the freezer. Tenderness, flavor, aroma, juiciness, and color can all be affected.

This would lead me to believe that anything above -18°C insufficiently suspends microbiological activity and would, even if only after multiple years, lead to spoilage. But then the EU legislation counters:

[it is] necessary to maintain at least that temperature [of -18°C], subject to a certain technically inevitable tolerance, during the storage and distribution of quick-frozen foodstuffs before their sale to the ultimate consumer [...] these tolerances shall not exceed 3 °C

Oh...kay so is -15°C the actual temperature below which all is fine and -18°C includes 3°C of buffer? Since I haven't been able to procure the original research these laws are based on (and they would be behind Elsevier paywalls, because why would citizens need to read what laws are based on anyway?), and no site goes into that much detail, I can't say for sure. My best interpretation is that if the outer edges of animal products reach -15°C during defrost, it'll still be fine for practically enough time that it's of no real consequence.

According to the mechanic that was over, combined with online research, the defrost cycle works like this:

  1. There is a fan circulating air from the cold coils into the freezer compartment (you can hear this run most of the time)
  2. The fan is turned off when defrost turns on
  3. The defrost heats the evaporator coil to +5°C
  4. The cycle ends, heating turns off again
  5. After cooling down, the fan turns back on

During all this, the air temperature in the freezer compartment should (due to the fan not replacing the cold air with warmer air) not rise noticeably, perhaps 1-2 degrees, if that.

Furthermore, the mechanic said that he would not be surprised if the display did not indicate a correct temperature during this time in order not to unduly worry customers. If the freezer is significantly warmer than it should be for a long time, though, our model should have started beeping and not stopped until we dismiss the alarm. This was apparently broken for us, too, as it was way too warm without ever sounding the alarm, so don't rely on that.

If you notice something strange with your freezer's temperatures and find forum threads or blog posts, you'll find lots of strangers on the internet arguing that "of course the air temperature fluctuates in a defrosting freezer". If you point out that this seems questionable from a food safety perspective, you'll find jokes about solid water not flowing very well and that you should stop measuring air temperatures. This is not directly or indirectly supported by any research that I have found. Heating the shelves or inner walls (and, by extension, your food) to a temperature at which water flows is absolutely not safe if done with any sort of regularity—as a no-frost cycle obviously would.

Given the EU regulation linked above, I expect that you also have a good warranty case if you notice your freezer reaches -14°C or above when set to -18°C. This is assuming that you did not significantly change its contents for a few days and have not opened the door for a few hours, just to be sure. My manufacturer happened not to be difficult about this, but that leads me to another problem...

Repairability of your freezer

I typically read up on a topic, select my requirements, and pick the cheapest product fulfilling those requirements. The requirements can include things like no child labor, durability, etc. (so I don't merely look at "does it work" + price, at least when I feel like I can reasonably know about whether one manufacturer would use child labor over another). In the case of a freezer, what fulfilled my requirements was a Koenic kfz45211a2nf of €500, but it turns out I haven't selected all the correct requirements.

The Koenic brand is from the MediaMarkt itself and the units are made by some Chinese company, optimizing for price. This results in a manufacturing process that makes disassembly and repairing complex and thus often more expensive than buying a new one.

Today, my less-than-two-year-old freezer was degraded to scrap metal for a simple mechanical defect. I feel bad about it and would happily pay for the few hours of extra labor it involves if that's what it takes for them to repair my unit, but it appears that the only option is to pay the full €600-700 of repairs or receive a new one for free. Well... in that case, I think it has a bigger impact for me to spend 600 euros on combating climate change in a more direct way. And on a better freezer when I next need to make that choice. And spread the word.

The mechanic mentioned Siemens and Bosch specifically as brands which would repair rather than replace. Though I do wonder if the costs might be similar and the only reason not to replace when facing a ~€650 repair is that the units are more expensive. That's a question for when buying my next freezer.

Is the environment suitable for my freezer?

If the room that your freezer is in is too cold then, counter-intuitively, your freezer can't keep cold. I've seen at least three different reasons given for this on various websites, but the more plausible two are: (1) that the refrigerant might not be able to condensate at too low temperatures (it's meant to evaporate when cold, to cool down the freezer, and needs to be able to condensate to dispel heat), or (2) that the oil becomes too viscous and the compressor can't work well anymore or even get damaged[4].

Refer to your operating manual before choosing a place to put your freezer. A garage without heating might not be suitable in winter.

How do I check my freezer's temperature?

I first noticed the issue when slides of bread thawed completely within five minutes. I like to spread chocolate paste when the bread is solid so you don't mush or tear the bread while spreading, but about halfway through the first slice that started to fail. Monday morning me was not amused. Similarly, in hindsight, the nice and easy ice cream scooping of late might not have been a consequence of getting a different brand of ice cream but of the higher temperature. (Indeed, next time the ice scream was just as hard to scoop as the old brand :(.)

Once you notice such a warning sign, start measuring. Don't put a digital thermometer in because the LCD (liquid crystal display) might not enjoy that (I don't know if it just doesn't work or if you'll permanently break it), but most analog ones that you might also hang up outside should go until appropriate temperatures. Reading the temperature (opening the freezer door) every hour will impact your readings, but not by a lot. My freezer was capable of cooling by 4°C between two openings 50 minutes apart, so opening no more than once an hour should still allow it to at least keep a steady temperature so long as you don't expect it to additionally cool new items down.

Another option is to put a Raspberry Pi inside, if you have one. The board is rated for -40°C and while the USB and network parts are rated only until 0°C, logging to sd card (the microsd itself is rated until -40°C as well) worked even at around -22°C. The CPU at that point was around -10°C. You can measure the CPU temperature and, while it obviously won't show you what the environmental temperature is in absolute terms, it should show you what kind of cycle is going on and what effect things like opening the door have. At room temperatures, I could map the environmental temperature 1:1 to the CPU temperature, assuming the OS that you installed is not doing any periodic tasks. You also shouldn't log very frequently to avoid that heating up. I found that every 30 seconds does not cause any measurable heat. I ran the power cord outside to avoid battery trouble and didn't notice that the hole in the seal around the cable had any measurable impact.

An infrared thermometer is another option, but note that most models are fairly inaccurate (give or take 1-3°C), get more inaccurate the further you go from room temperature, and that different kinds of objects radiate differently (paper and plastic of the same temperature may show up differently on the infrared spectrum). If this is the only or best option for you, you should pick one object that will be in your freezer throughout the period, does not move around, and does not touch other (newer, hotter) items, and consistently measure that within 5 seconds of opening the door. Even an analog thermometer, I can see warming up by the seconds after picking it up by the wooden frame, so infrared (surface measurements) are probably even more prone to that.

The door display, if you have one, is of course another option and may very well be reasonably accurate, but if you suspect something is wrong with your freezer... don't trust that without at least checking its readings for plausibility.

What are my warranty rights in the EU?

The merchant (seller) is the person that is legally liable for giving your warranty. They may send you to the manufacturer and you can go along for cordial reasons and see if the manufacturer is willing to work with you without requiring the merchant to forward messages back and forth, but you're not required to. The manufacturer is not a party to your contract of sale (koopovereenkomst), also not if the merchant's terms of service said so.

In the EU, new electronics have 2 years of warranty or more. The "or more" part comes from your member state's specific implementation.

In the Netherlands, the law actually just says that you have a right to a sound product, so if you find that a freezer should last just as long as your grandparents' 50 year old freezer did, then perhaps you can make a case in front of the judge. In practice, most merchants in the Netherlands claim that 2 years is what you get and will require nagging and good reasoning before you get more. In my experience though, many merchants are reasonable if you ask professionally and give logical reasons, optionally backed up by references to websites such as ConsuWijzer (made by a government agency).

Be sure to indicate that there is an issue in a timely manner. After 6 months, the burden of proof is on you, though in practice this rarely changes anything. However, if the manufacturer can prove that you should have known of an issue sooner (if you claim a missing button on a computer mouse after nine months, let's say), they can also refuse warranty even if it's within the regular warranty period. If your freezer has a mechanical issue that becomes worse and worse, it might be clear that it has existed for a long time while you kept using the product (getting use out of it) and this may reduce or nullify your claim. (Dutch: Burgerlijk wetboek, boek 7, artikel 23, lid 1.)

Should I just have gone for a simple old freezer instead of this no-frost stuff my grandkids are on about?

Nah, there is no evidence for that either. It's more convenient and should also be more energy efficient in the common case. Freezers without no-frost (no-no-frost freezers?) will build up ice and become less effective, needing more power to maintain the temperature until you do another manual defrost.

For what it's worth, the semi-broken freezer I had was still keeping within the indicated power consumption. Extrapolated from 17 days, it would slightly undershoot the yearly quota at an environmental temperature of 19°C. Turning on the ultra freeze function to make it keep food safe until the replacement freezer arrived made it use significantly more, of course. If this is something you really care about and you're the kind of person that would do regular defrosts without having food spillage because of it, just compare the energy labels for models designed to work within the same climate classes: so far as I can tell they're reasonably accurate.

As for things like freezer burn, this is caused by temperature fluctuations which a correctly functioning no-frost freezer should not really do. The main cause for temperature fluctuations in both kinds of freezer are from opening and loading.


If you still have warranty on your freezer, check it today. Put a regular analog thermometer inside or read the door display. Check it a few times at random intervals. Does it ever go above -15°C / 5°F? Something is probably wrong.

Bonus: A Python script to calculate an average of a time series when the data was collected neither regularly nor at random intervals.

P.S. If you are able to find, or performed, original research on relevant microbial activity at domestic freezer temperatures, I'd love to hear of it. It's still unclear what the actual limit for food safety is despite multiple evenings of research and some more evenings arguing with people on the internet.

Update 12 hours later: I found this paper with the following diagram:

blurry diagram from page 326

So -10°C is said to be enough, and it includes references from around the same time as the ones I was looking for! Yay... except that these are also nowhere to be found. Google Scholar is only able to come up with citations, not actual papers. This conclusion is rather different from what the EU and USA adopted.

If anyone has more insights or can track these down, I'd be interested to hear of it!