The efficiency of Peltiers in regards to heat moved per watts spent is actually kind of crap. And the hotter their hot side gets, the worse that efficiency becomes. Overclockers attempting (often in vain) to use them for sub-ambient-temperature cooling for computer components have known this for a long time.
Even a dinky compressor based refrigerant system is more efficient at moving heat than a Peltier of equivalent-ish total wattage.
That's before getting into the details specific to the crappy tiny Peltier cooler in the video, which runs at 100% of its duty cycle all the time and has functionally zero insulation, unlike an actual refrigerator.
I have a Peltier based car cooler, and that's basically the only use case for these things that makes sense: Short term storage of things where the actual temperature achieved isn't that critical, and which needs to remain light enough that you can easily carry it. Basically, just treat it as a regular cooler that if you play your cards right you won't have to add any ice to.
Overclockers attempting (often in vain) to use them for sub-ambient-temperature cooling for computer components have known this for a long time.
aren't you still limited by ambient air temp because the hot side of the Peltier needs to be cooled by air anyway?
I have a Peltier based car cooler, and that’s basically the only use case for these things that makes sense:
I don't really even see the point of that to be honest. if we're talking short periods anyway, a nicely insulated cooler with ice packs (cooled by a heat-pump freezers) is way better imo.
aren't you still limited by ambient air temp because the hot side of the Peltier needs to be cooled by air anyway?
No, that's why the baby fridge works. The peltier TEC in the fridge can produce temperatures below ambient, but generally not below freezing. Computer chip TEC coolers would go farther and push more power through the TEC and do things like water-cooling the hot side instead of a little fan.
The reason they don't get used anymore more complicated. It's my understanding that basically TECs have a sort of limit on the amount of heat they can push from the cold side to the hot side, because moving more heat means more energy used and thus more waste heat. Apparently most modern chips are past that limit. IIRC, TECs can only move something like 100w of heat - past that and they start to heat themselves up because of waste heat. Modern chips can be like 300w.
Sub ambient cooling also comes with a bunch of issues like condensation, so no one really uses it day to day. Also, chips are run so close to their limits now that cooling like that doesn't get you nearly the performance difference it once did.
aren’t you still limited by ambient air temp because the hot side of the Peltier needs to be cooled by air anyway?
You can certainly get subambient. Put some electrical current through a Peltier and one side gets cold, and the other side gets hot. Use the cold side to cool your components, and get the heat away from the hot side, and you can make it work.
It can be a bit tricky. The hot side is right next to the cold side and it gets really hot, so if you can't get the heat away it'll leak right back over. Peltiers use a lot of power so you need a beefy power supply, and that'll be another source of heat. Assuming you can figure that all out, you also have to be careful that the cold side doesn't get too cold or you get condensation. Electrical components tend to not like moisture very much.
I remember people experimented around with it back in early-mid 2000's. General consensus nowadays seems to be is that it's not terribly effective or practical and not worth the trouble.
Planing a cooler presentation when I get my Redneck Business channel going. I'm stunned by how ignorant people are on the subject. My $100 Lifetime cooler keeps cold exactly as long as a $500 RTIC or Yeti. And those are only for if you need multiple days of cool. I have a dozen coolers for various needs, and most are Igloos I got from the thrift.
That explains why that old minifridge I got for $15 on black friday was such a piece of shit that lasted less than 6 months.
These devices always remind me of these guys trying to make Solid State Refrigerant:
I suspect they fall into the same category of “neat concept but commercially useless”
Thermoelectric cooling where you have a water loop and second stage heat sink is actually very effective. The problem with that however is you can generate temperatures below freezing, which can cause condensation in your device under cooling and lead to internal short circuits. Ideally you'd feed back the temperature of the device under cooling to your second stage heat sink so it can ramp up or down heat transfer
