Thermoelectric cooling: it's not great.
This is a Peltier element. It may look like a rather unassuming hunk of material, but sandwiched between these two plates are a whole bunch of semiconductor junctions that have been carefully arranged to do something remarkable: when I apply a DC voltage across these two wires, the plate becomes cold. This is called thermoelectric cooling, and it’s happening thanks to the Peltier effect. With advances in manufacturing, we are now able to produce these devices quickly and cheaply. These ones here cost less than $3 each. There’s just one little issue with this technology: it’s not very good.
Like, at all. OK, we’re gonna need some context. I’m making this video for a couple of reasons: First, you may know about my one true love, the refrigeration cycle. I adore a good heat pump and could talk your ear off about them for hours on end, but vapor-compression based refrigeration systems are bulky and mechanically complex affairs.
They also have some environmental issues stemming from the refrigerants they use. Because of those downsides, with a regularity you could practically set a watch to well-meaning people discover Peltier elements and question why we don’t use these instead. A device with no moving parts and which makes cooling happen simply by putting voltage across it does feel like the future! That’s gotta be better than all the faff we go through to build a fridge, right? Well, “better” is a slippery word. For one thing, the Peltier these are named for is Jean Charles Athanase Peltier, and he discovered this phenomenon back in 1834.
We’ve been taking cracks at this for a long time, yet it still hasn’t changed the world. It is true that this form of cooling offers some unique advantages and in very specific applications with equally specific design constraints, using a Peltier element can make some sense. However, the situations in which these are truly appropriate to use are very, very rare. You will absolutely understand why that is by the end, but for now (and this leads into the second reason I’m making this video) you should know that the majority of products on sale today which use these things are being made purely to capitalize on the fact that we just can’t stop ourselves from buying cheap crap. Speaking of cheap crap, you’ve probably seen these little personal refrigerators before. The idea is you can keep a few cans of your favorite beverage cool on your desk! You’ll be the talk of the office, for sure! All of these things use Peltier elements to cool their insides for a very simple reason: cost.
These have to be cheap enough for people to pick them up impulsively. And they are - this was priced just over $30. You can’t build a proper fridge that cheaply, but you can build a little plastic box with a cheap power supply and $2 Peltier module in it. And people will buy it. Look, I’m a people, and I bought it! But is it any good? Spoiler alert: No! Let’s turn it on and put it through its paces! It has a power switch helpfully labeled “Off” and “Cold.” That's a sign of quality, that is.
Here goes nothing! [a fan spins up] Ooh, it’s even got an interior light! How thoughtful. But, uh, that’s a lot of noise for something with no moving parts. What gives? Well, the Peltier element itself doesn’t have any moving parts but when I showed you it in the beginning, I only showed one side of it. That side was getting cold, yes, and in this... alleged fridge the cold side of the element is bonded to the rear wall
which we can see getting cold in the thermal camera. But as the peltier element generates cooling on the one side, the other side of the element gets hot. Very hot. Problematically hot. Why? Well, two reasons.
Peltier modules produce cooling by absorbing heat on the cold side and moving it through the element to the hot side. That’s how all cooling works - energy doesn’t just disappear, so to lower the amount of heat energy in one location, you have to absorb it from that location and move it somewhere else. Which is exactly what these do. They absorb heat on this side (which to us feels like coldness) and reject it on the hot side. But since the heat is only moving across a thin little barrier with very little mass, it builds up really quickly.
But that’s not all! The element itself creates heat as it runs. These here are rated 5 amps at 12 volts, so when running at full capacity they’re dissipating 60 watts of heat in a really small space, and that’s on top of the heat it’s pulling from the cold side. So, to keep it from destroying itself, you need a finned heat sink bonded to the hot side to spread that heat out, and you’ll also need a fan to force air across the heat sink to spread the heat even more quicklier. And as we can see here, the itty bitty fridge has a generously sized heat sink hanging off the back with a fan blowing right at it.
A fan and heat sink pair isn’t complicated or expensive, but it means this is noisier than you might expect and, ya know, adds a moving part. But noise is forgivable, the real question is how well does it work? And the answer to that question is: not! And by every conceivable metric you can imagine. For a start, you may have noticed this has no temperature settings.
It’s just cold or off. That’s because it only promises to lower the internal temperature by up to 30 degrees below ambient temperature. That’s it. I appreciate its honesty, but if you know anything about food safety, you’re gonna be getting the willies right now. Because in an office environment during the summertime where, say, the room temperature might be 75 degrees Fahrenheit, this thing will - at best - get the interior down to 45 degrees Fahrenheit which by US standards is not food safe.
So… strike one. But ok, let’s say you just want to cool some beverages and you don’t care about food. Well, to see how well it manages that I loaded it up fully with 6 room temperature beverage cans then shoved one of my temperature data loggers in the void space between them. After running for two hours, it hadn’t even gotten below 60 degrees inside here. After 12 hours it had only managed to drop to 48 degrees Fahrenheit, and after 24 hours the temperature had plateaued at about 46 degrees. Not very impressive.
Somehow it got a bit of a second wind right near the 24 hour mark, which I believe happened just because the room it was in had cooled off a little, and we ended up bottoming out at 45.5 degrees. Which, in fairness, was about 30 degrees below the room temperature. So it did what it said it would, but it sure took ages to get there.
However, I was measuring the air temperature in the center of the fridge and since the cans are all touching the sides, I found that the actual temperature of the liquid in those cans was a slightly warmer 50 degrees. Chilled, I suppose, but I wouldn’t want a beer that warm. So… strike two. But okay, so it’s not technically a refrigerator and it takes forever to cool down and it doesn’t keep things that cold and it’s louder than the box implies but that doesn’t mean it’s not useful! And to that I’d say, you’re not technically wrong! Some people use these things to keep certain cosmetic items in a cool, dry place, and from what I can tell they’re happy to have these! And if 50 degrees is cold enough for you, and you’re diligent about rotating cans in and out, you might actually enjoy this as a beverage cooler. But here comes strike three, and it’s a doozy (and also why Peltier devices are generally terrible).
If you’re at all concerned with how much energy this thing will use to do its job not very well, you should know that this thing consumes about 55 watts of power all the time. That’s not a ton of power, no, but guess what uses less power than that? Well, lots of things but most relevant to this discussion - an actual refrigerator. This here is what I call the standard cube fridge, the most basic mini-fridge you can buy. They’re generally around $100 and while they are incredibly basic they are also actual refrigerators. We have a lil’ baby compressor back here pumping refrigerant through an honest-to-goodness refrigeration circuit.
Now, it did cost significantly more money to purchase than the blue... thing, but it can also hold a lot more stuff. Including the blue fridge.
By the internationally recognized metric of “how many cans of La Croix can you shove in there per dollar” the real fridge wins by a lot. It’ll hold 33 cans easily, and 36 if you’re a little bit creative, which works out to $2.77 per can. And the blue fridge, despite only costing $30, only holds six cans so that’s $5 per can. But that’s not the real issue.
Remember, the blue fridge uses 55 watts all the time. Guess how much power this thing uses? Once the refrigeration circuit has stabilized, it pulls between 45 and 50 watts from the wall. I’ve consulted with some math scholars who have confirmed for me that that figure is less than the 55 watts the stupid blue fridge pulls. But that’s not all! This, because it’s an actual refrigerator, has a thermostat to maintain food safe temperatures inside! And that means it doesn’t run all the time. I monitored its energy use when loaded up with 15 already-chilled beverage cans, and to keep them cool over three hours, it consumed just 65 watt-hours representing an average draw of 21.7 watts. And, by the way, the room it was in during that test was a pretty warm 78 degrees.
Now, electricity isn’t free. So let’s do a little cost comparison. The blue fridge, with its constant 55 watt draw, will consume 39.6 kilowatt-hours per month. Going by the average cost of electricity in the US of $0.14 per kilowatt-hour, it will cost about $5.50 per month to use continuously. So about $66 a year. The much larger mini fridge, on the other hand, drawing 22 watts continuously will consume 15.8 kilowatt-hours in a month,
which costs $2.21 per month, or $27 bucks a year. That’s $39 less than the blue fridge per year, so it doesn’t even take two years of operation for the extra $70 you spend on the real fridge to pay for itself. Seems like this blue fridge might be kind of bad! Now I could stop here, but I’m not gonna because this thing's awful efficiency just gets funnier the more context you have.
You remember this goofy thing. It’s moved into my office and has an actual temperature controller now which fixes its major flaw. It’s still by no means a good refrigerator - stick a twelve pack of room temperature cans in there and it runs nonstop for almost four hours. It’s really not built for making warm things cold. But as far as its ability to maintain food safe temperatures, it does the job just fine.
And how much energy does it use? Well, after adding 6 room temperature beverage cans, I measured its energy use over 24 hours and it needed only 970 watt-hours, meaning on average it only pulled 40.4 watts. Which, believe it or not, is less than 55. And that’s despite the rather pronounced size difference between these two machines. I mean… just look at them! It’s frankly ridiculous that such a tiny little “fridge” is using more energy than the comparatively gigantic one next to it.
In an earlier 24 hour test when the weather was a little cooler, the red fridge consumed just 890 watt-hours in 24 hours while the stupid little blue one needed 1,290 watt hours over the same period. Because it’s always pulling 55 watts whenever it’s switched on. That’s 45% more energy spent only kind of cooling six cans. And ya know how many cans this sucker will hold? If you fill the crisper drawer, all the door shelves, and double stack on the top shelf, it’ll hold 135 cans.
It’ll take it forever to make them all cold but once they get there the fridge will use roughly the same 900 watt-hours per day keeping them all ice cold. Are you starting to understand why I keep calling this stupid junk? And let’s not forget the red fridge has a freezer compartment for ice cream and pizza and stuff! Now, as I’ve covered before, mini-fridges are actually some of the least energy-efficient refrigerators out there thanks to their minimal insulation. And by the way, I need to clear something up for the Europeans who keep badgering me about this: I never said the red fridge was a mini-fridge! A bunch of you heard me say that somewhere but I would never consider a fridge which goes up to my chest to be a mini-fridge.
I called it little, and I called it a bit on the small side but I also said, and I quote, “it’s much bigger than a typical mini-fridge and served me well as a decently competent refrigerator.” At the very end of that video I did lump it into the category of mini-fridges because of its thin walls, but that’s as close as I ever got to calling it a mini-fridge. Methinks your confirmation bias kicked in a little early and a little strong.
Maybe those 230 volts are gettin’ to your head. But speaking of Americans and their giant American fridges, [banjo music starts] at home I’ve got a GIANT AMERICAN FRIDGE! [said in a thick Southern accent] My 360 pound behemoth (that’s 163 kilograms for you commies) features automatic defrost for maintenance-free operation, is plumbed to a water line for its built-in ice maker because we want ice in our drinks, goshdarnit, and it has 16.35 All-American Cubic Feet of Fridge Capacity and a 5.59 cubic foot freezer compartment. That’s 461 liters in the fridge and 158 litres in the freezer. And guess which of those two fridges uses more energy? THIS STUPID THING! Yes, according to government testing (I didn’t do my own, sorry about that) my bottom-freezer, French door fridge (which are two dings against its energy efficiency), despite being able to hold over 50 times as much stuff before we even count the frozen food, and which features heated defrost uses slightly less energy over a year than it takes to run this toy.
It needs 1,270 watt-hours per day to attain actual refrigeration compared to the 1,290 this stupid thing used to keep six cans of la croix only kind of chilled. I ask again, are you starting to understand why I think these things are stupid junk? Side note, I used the Canadian energy stats because as I found out the last go-round, the American EnergyGuide label is very very pessimistic thanks to the way the tests are run. It includes a lot more food browsing in hot weather than I think is typical, and the energy guide label for the red fridge was about 20% higher than my own stress testing.
But even if we go with the 633 kWh annual figure, then my actual fridge uses only 35 percent more energy for approximately 7,000 percent more refrigerated space with a freezer on top of that. Or, below it, actually. So, why is this blue lump of sadness so much less efficient than any of the refrigerators we’ve discussed? Because all the rest have… heat pumps! Now, very out of character for me, I’m going to keep my explanation of the refrigeration cycle pretty brief.
Kind of brief. Brief-ish. But here’s why it’s such a big deal: a vapor-compression heat pump like the one we find in even this very cheap and basic mini-fridge can move more heat energy out of the fridge's interior than it takes to run the heat pump. That may sound impossible, but the compressor is the only thing in here doing any work, and all it’s doing is using an electric motor to spin some pumpy parts inside this enclosure which pressurizes a gas. In the case of this fridge, that’s literally the only thing it’s doing.
All the rest of its parts are just pipes and tubes which are exposed to the air inside and out. The gas it’s compressing is called a refrigerant. Refrigerants are a category of gasses that we’ve discovered (or sometimes engineered) that have really useful relationships between their pressure and their boiling point.
This fridge, like most on the market today, uses isobutane which at atmospheric pressure is a gas. But if you pressurize isobutane to about 80 PSI, its boiling point shoots up to 110 degrees Fahrenheit. That’s much hotter than typical room temperatures, so if you feed that high pressure gas through some tubes that are exposed to ambient air, the air will cool it down to the point that it can’t be a gas anymore and it will spontaneously condense into a liquid.
That condensing action releases a ton of heat energy because of a thing called the latent heat of vaporization which to save time I am not getting into right now, but you can see this heat being released in the thermal camera. You can even make out the tubes the refrigerant is traveling through just below the surface of the fridge’s outer skin. Thanks to the pumping action of the compressor, the refrigerant is actively pushed through those tubes.
At the entrance it’s purely hot gas (and it’s hot because the compressor just compressed it), but as it makes its way through it sheds heat energy and begins to liquify. By the time it makes it to the end of the tube, a slug of liquid refrigerant will have bunched up. Then, that liquid refrigerant gets sent through a restriction which limits the volume of fluid flow. Because of the restriction and the pumping action of the compressor, a pressure imbalance is maintained on either side. And once the liquid makes it through it finds itself inside the evaporator where the pressure is very low, in fact so low that the boiling point of isobutane plummets to something like -20 degrees Fahrenheit. Even the air inside a freezer is warmer than that, which means there’s energy available to boil the refrigerant.
That causes it to absorb latent heat energy which means it gets very, very cold. Once the refrigerant has completely boiled away and becomes a gas once more, it finds itself back at the compressor and the whole thing starts over again. The gas it’s compressing is called a refrigerant.
Refrigerants are a category of gasses that we’ve discovered (or sometimes engineered) that — okay that’s enough of that. Key to understanding this process is that the refrigerant, through boiling away, absorbs a lot of heat energy. That heat energy is then stored in the refrigerant.
But, when we compress it and it condenses back into a liquid, it releases the energy it had just absorbed. Since we’re in control of where that happens, we can move heat energy from one place to another. In the case of a fridge, we’re pumping the heat out of the fridge’s interior which cools it down and then releasing that heat to the outside air.
And the only thing we’re doing to make that happen is compress a gas and pushing it through some pipes. The refrigerant does the heat transfer stuff all on its own. That is how a vapor-compression heat pump is able to move more energy than the machine itself consumes. The heat moving capacity compared to its power draw is a metric known as the coefficient of performance. Now, not many people calculate the COP of a fridge, but if I assume it has a COP of 3 then we get a threefold increase in cooling power compared to input power. In other words, the 45 watts of power this fridge pulls from the wall when it’s running generates 135 watts of cooling power in the evaporator.
That is why we bother with a compressor and all this piping and junk, it’s just really, really efficient. And that’s also why heat pumps for heating our homes are becoming such a big deal: just as there’s heat to be pulled out of a freezer, there’s heat to be grabbed from the air outside. And so long as the COP of an operating heat pump is over 1, we're getting some free heat compared to running an electric heating element. But anyway, this video isn't about that.
Peltier elements, because they move heat from one side to the other, are technically heat pumps. And that means we can measure their coefficient of performance. But that’s tricky to do because unlike a vapor-compression heat pump, their efficiency will vary wildly depending on the temperature difference it’s fighting as well as how much current you’re shoving through it. Remember, this is really thin. And the materials used to construct this conduct heat even when it’s not running at all, so the higher the temperature difference between the hot and cold sides, the more heat energy leaks through the element itself and the less efficient it becomes. And the effect there is dramatic.
On top of that, the more current you try to push through this, the more heat it generates inside itself which also makes things worse. In theory, if everything is perfect and you can run these at very low current, you can get a COP between 1 and 2. But that will hardly generate any cooling at all, so in practice, the COP of a device using a Peltier element for cooling is somewhere between zero and very bad. And that’s why this stupid little blue fridge is using so much more energy than this actual fridge. Or that actual fridge.
Or THAT actual fridge. Now, in fairness, there is more going on here. Its reliance on the Peltier effect is of course the major factor in this fridge’s terrible efficiency, but it’s also got a lot to do with its design. Any actual fridge is subject to regulations which mean it has to have enough insulation in its walls to pass performance tests.
The blue toy is a toy, and honestly I could barely even detect any insulation at all! The walls all feel hollow. So I drilled some exploratory holes and determined that it does have what appears to be Styrofoam in the side walls but the door is in fact hollow. That’s great.
If this had more insulation it could probably get its insides cooler using less energy, but it would then also need some sort of temperature control. I think it’s actually relying on heat intrusion through the door to keep things from freezing. But any time you’re using a Peltier element, even the best designs with lots of thermal insulation are going to run into the fundamental problem of a terrible coefficient of performance. These are just not very good at what they’re supposed to do and will always lose a fight with a grown-up heat pump.
And, not for nothing, while people fixate on the harm refrigerants cause (because they have been truly awful to the environment and many in use today still are in one way or another) isobutane is not really a problem at all. It’s flammable which presents some practical challenges, especially when the systems require service, but it has a global warming potential of only 3.3 and its other environmental effects are negligible. Plus, there’s hardly any of it in a fridge. Even my big one at home has a refrigerant charge of just 55 grams. So when it comes to domestic refrigerators, you honestly don’t need to be worried about that anymore.
You may remember, though, that I did say Peltier elements have some very specific advantages in very specific applications. I’ve already covered the main one - if you need some cooling as cheaply as possible, they are an option. You just have to be aware of how energy-intensive they are and decide whether that trade-off is worth it. And sometimes it might be! If, for example, you’re only using something occasionally, the poor efficiency of these things might not matter that much. One example would be a portable cooler you can bring with you in the car.
If it’s only ever going to be used on occasional car rides, and it’s got the energy source of a car to supply it, its energy efficiency isn’t that critical. And portability is another area where Peltier elements can shine. The elements themselves weigh almost nothing, and even finished products are quite light.
This little thing doesn’t even weigh 4 pounds (it’s 1.68 kilograms to be precise in metric fashion). Meanwhile the mini-fridge, despite having the most adorable little compressor I’ve ever seen, weighs a much heftier 33 pounds (15 kilograms). In fairness the bulk of its weight isn’t from the refrigeration circuit, just the metal parts that make up its body and frame and the door, but it’s still a pretty huge difference. And when it comes to portability, vapor-compression based systems have an Achille's heel: they are orientation-sensitive. The compressor that’s inside this little black ball is made of metal parts which rub against each other and thus need lubrication, and to provide that lubrication (and also cooling for the windings of the electric motor) the system has oil in it. The compressor housing forms an oil sump, and we rely on gravity to keep that oil pooled in the bottom of the sump so that the compressor has the lubrication and cooling that it requires.
And Earth’s gravity, famously, only ever pulls things towards the center of the Earth, so you have to have this upright for it to work without destroying itself. And, if you move anything with a refrigeration compressor in it and you aren’t 100% sure it stayed upright as you moved it, you need to let these sit upright for at least a few hours before you switch them on so any oil that might have gone somewhere it shouldn’t can drain back to the sump (many manufacturers will tell you to wait a full 24 hours before switching them on after they've been moved). Peltier elements, because they don't have any mechanical parts, don’t care about that! You can operate this fridge however you like, though of course you need to make sure the cooling vents aren’t blocked, and you'll never have to wait to switch it on after you move it. It’s always ready.
And Peltier elements also have one last trick up their sleeves: when you reverse the polarity, they move heat in the opposite direction! That means you can offer a device which both heats and cools quite trivially. In fact, heated and cooled cupholders in cars is one application for this in real life. You can bond a Peltier element to a thermally conductive ring surrounding a cup holder (and use some sort of heat spreader for the other side of the element) and you can keep hot drinks hot or cold drinks cold just by flipping a switch.
But, in all honesty, that’s mostly a gimmick. I saw it at the auto show when I was maybe 12 and it hasn’t taken the world by storm yet. But back to the present. When the design requirements of very cheap to build, easily portable, and only occasional use all collide, using Peltier elements for cooling makes some sense.
But if this is going into something that’s going to be in anything close to continuous use and energy efficiency matters either for cost or scarcity reasons, it just doesn’t make any sense at all. There is in fact a very good reason refrigeration is still happening the way it always has. And if you’re looking at a product like this and thinking that might be useful, I would highly suggest you think long and hard about whether one of these cube fridges might fit your needs. No, they’re not going to fit on your desk (at least not easily) but they’ll do a lot more for you using a lot less energy. And these, too, can be had in all sorts of goofy colors if you look in the right places. That said… I would be remiss if I didn’t mention the big downside to some of these, including this model: ice builds up on the evaporator over time and so this needs occasional defrosting.
Depending on what you’re using it for that might be very rare but it’s something to keep in mind. However, there are also plenty of these out there which have rear-mounted, vertical evaporators and skip the freezer compartment entirely, and those don’t need defrosting. I would honestly seek one of those out if you’re looking for a mini-fridge like this. Honestly the freezer compartment in here is pretty useless and just eats up space. The only reason I grabbed this one was because it was on sale at Menards. Only $80 and you know I mailed in that 11% rebate! And speaking of cheap things from Menards, I should say this particular model has a few annoyances.
Its compressor is oddly buzzy, which is disappointing. The red fridge is actually quieter than the thermoelectric fridge, but this baby fridge is louder. There’s a good chance I just drew the short straw there and most of these have quieter compressors, and besides I don’t think it would really be very annoying under a desk. But… something to mention. What actually worries me about this model is the fact that the thermostat runs very, very short cycles.
Like, once it’s down to temp it runs for 2 or 3 minutes and then stops for 3 or 4. That’s usually not the healthiest thing for a refrigeration compressor. But in this case, because of the design here, I actually think that’s on purpose.
Since the interior is so tiny, you’re going to end up with stuff right against the evaporator, so if it were to run for too long at once it would likely freeze some stuff on the top shelf. But anyway, now we’re just shopping for mini-fridges so I think I should wrap it up. The bottom line of this video is that Peltier elements are interesting and cool devices but they're just not good for general purpose cooling. And since we’ve adopted isobutane as the standard refrigerant in domestic refrigerators, to be honest you can stop feeling guilty about bringing another fridge into the world. I mean, you probably shouldn't buy one unless you actually need it (and I’m 50/50 on whether I’m gonna keep this for storing my color film or donate it so someone else can make better use of it). But really this is just a cooler which happens to have a little heat pump in it.
The steel and copper in here is the only special sauce. Ooh, I could keep sauce in here! ♫ inefficiently smooth jazz ♫ It may look like the teleprompter hasn’t started but that’s cuz it hasn’t. Which to us feels like coldness, and reject ahh..
[clears throat] ehbutehbuDEDABAH Because all the rest have… HEAT PUUUuuUuuUUUumps. Woah that was not good. Even if we go with the 633 kilowattanggguhbuhdeblledyaghbleurf Even when it's not running at all. So the higher temperature difference between the cot and… dah. ..american energyguide label is very very pessimistic thanks to the way the depart oh… debeduh debeduh debbity duh It’s flammable which presents some practical challenges especially when ssssssssssssssssshhhwhww welp.
So I think we can sum this up pretty well: Peltier elements? Not cool. Oh except they are... on the one side. But not cool as in, like, good. Or hip. Or rizz. Is that what the kids say now? oh no I looked it up and that is NOT RIGHT skibidi heat pump
2024-09-23 14:32
