Staying cool without refrigerants: Next-generation Peltier cooling

https://news.samsung.com/global/interview-staying-cool-without-refrigerants-how-samsung-is-pioneering-next-generation-peltier-cooling

by simonebrunozzi • 1 year ago

387 315 1 year ago

news.samsung.com

WarOnPrivacy• 1 year ago

I used peltier coolers on K6 processors 1M computer years ago. They worked great and lasted about as long as the PCs did.

Not long after I bought mine, they disappeared from cooler offerings. I've wondered what became of the tech.

PaulHoule2• 1 year ago

I remember one objection to Peltier coolers in that application is that could possibly cool the CPU below the dew point and cause water condensation, something a fan can't do.

rurban2• 1 year ago

For this we calculate the dew point with 2 sensors and react accordingly in my temperature controller.

I'd really like to try out these better peltiers, our current ones suck. And the fans to remove the heat are huge and loud.

marcusb• 1 year ago

A friend had a Peltier cooler on a late 90s/early 2000s CPU and had the condensation problem. The cooler (or maybe CPU, or both; I can't remember which) had an algae-like growth all over it.

gsf_emergency_22• 1 year ago

Anybody who points to dew point issues would be fascinated by

https://youtu.be/qAZ-q3KmDHM

duskwuff• 1 year ago

I've wondered what became of the tech.

Remember that Peltier coolers don't make heat disappear - they just move it from one side of the cooler to the other, and produce a lot of additional waste heat in the process. There are better ways of transferring heat from a hot IC to a heat sink nowadays - like liquid cooling for really high-performance systems, or capillary-action heat pipes for more typical needs.

delusional• 1 year ago

There are better ways of transferring heat from a hot IC to a heat sink nowadays

Peltiers were always a bad way to move the heat. What they offered was the ability to go below ambient, which at the time could improve overclocking. Peltiers of course lacked the capacity to actually take you there with any decent load, but in theory it could.

It never really made much sense for consumers, and once consumers realized that, the market went away.

s0rce3• 1 year ago

They are inefficient and dump a bunch of additional heat into the heatsink that you need to get rid of.

Aurornis• 1 year ago

Your K6 processor has a TDP somewhere around 20-30W depending on the model.

Modern performance CPUs have TDPs in the 100-200W range.

Peltier cooling generates a lot of additional heat. It doesn’t scale well to the higher loads. That’s why you don’t see them any more.

hypercube33• 1 year ago

I'd consider performance processors something like a threadripper - 300W unlockable to 1000W. I haven't kept up with intel but I think they were indeed riding 200 to 300w.

rcxdude7• 1 year ago

TDP went up. TECs are extremely inefficient, and have limits on how much heat they can pump in a given area anyway. a 10W chip, you can maybe get away with, and it might be a meaningful improvement compared to the other cooling tech available at the time, if you don't mind the power usage. with a 100W chip, you have no chance.

(You can see a demo here where LTT try it, and they, after dumping 500W into the cooler, can get the CPU to a vaguely reasonable temperature, until they actually load it up: https://www.youtube.com/watch?v=sWrqyQWfhrs)

burnt-resistor• 1 year ago

They made both the problem of power consumption and removing waste heat infinitely worse. Instead, I used liquid cooling with a copper block on one side and a heater core from an RV on the other. It was so efficient, it didn't even need a fan so long as the radiator fins remained somewhat vertical. I recall it used a pink additive to distilled water. That was bank in the day when you had to make double sure the very large PVC tubing was kink free, clamped properly, and oh the joys of removing air from the system.

mousethatroared8• 1 year ago

They weren't worth it. Their only advantage is they can cool below room T, but then you have condensation problems.

anticensor• 1 year ago

Did they solve the inefficiency problem of peltiers?

cosmotic• 1 year ago

The article says they have made peltiers 75% more efficient than existing ones.

nine_k2• 1 year ago

OK, a typical Peltier device has 3.5% coefficient of performance, that is, it produces 35 W of cooling per 1 kW consumed.

Fine, let's expect that the new tech doubles the efficiency, to 7%. Still, to my mind, pretty wasteful, on par with a steam railway engine. A Peltier element is good in cases where you can afford a large heat removal device, but need precise temperature control and no moving parts. For a home fridge, I'll take the sound of the compressor and the temperature fluctuations of a 400% efficient compressor-based heat pump over a Peltier element any day.

bluGill5• 1 year ago

They claim 75% better. I doubt this is better than compressors in general but they seem to be going for a niche where a high power compressor does the high load part but when only a litte cooling is needed the compressor is now very inefficient and so the peltier is better. Normal fridges just let the temperature have a wider swing which is good enough for most needs.

mort962• 1 year ago

Normal fridges just let the temperature have a wider swing which is good enough for most needs.

These wide swings annoy me. You hear that you shouldn't let your fridge go above 4°C, because that's dangerous. And you obviously don't want your fridge to go below 0°C. But finding a setting where the hottest part of the fridge doesn't go above 4°C (or even 5°C or 6°C) during the hottest part of the cycle and the coldest part of the fridge doesn't go near 0°C during the coldest part of the cycle is really pretty difficult, in my experience.

creshal• 1 year ago

75% better than "utterly awful" is an improvement, but you're still going to be searching for applicable niches with a microscope.

veunes5• 1 year ago

Yet I'’d love to see real-world power consumption numbers compared to traditional fridges

Cthulhu_3• 1 year ago

Would a peltier element allow for a more constant temperature? Or can you turn a peltier up and down easily? With a compressor based system it's always been "on or off", something that can ease up and down would be nice.

lhmiles• 1 year ago

What exactly does "75% better" mean? 75% more cooling per watt? 75% less distance from theoretical ideal?

ethan_smith• 1 year ago

Traditional Peltier devices operate at ~10% efficiency (COP of 0.5-0.7) compared to vapor-compression systems (COP of 2-4), but recent advances in thermoelectric materials like bismuth telluride alloys and segmented elements have pushed lab efficiencies to ~15-20%.

HPsquared2• 1 year ago

It's clearer to think in terms of "efficiency relative to ideal Carnot efficiency".

Compressor systems use twice as much energy as an ideal system, while Peltier systems use about 10x as much.

est• 1 year ago

In the Bespoke AI Hybrid Refrigerator Samsung launched in 2024, the compressor operates under normal conditions such as routine storage and retrieval, while the Peltier device activates alongside the compressor during high-load situations — like when storing large amounts of groceries or placing hot food inside — thereby enhancing both cooling performance and energy efficiency

I'd choose a fridge with a larger compressor.

baby_souffle• 1 year ago

Larger compressor would probably mean short cycling though. That has its own downsides.

BearOso2• 1 year ago

AI has literally nothing to do with this. Why do they feel the need to sprinkle the phrase everywhere? AI inverter/compressor? Come on, have some sense of shame, please.

savrajsingh• 1 year ago

Because it works wonders on non-technical people

rcxdude7• 1 year ago

It's the name of the game, alas. In the area I work in, I've seen many companies get many multiples of the investment we've gotten because it's "quantum", even though they're trying to do the same thing we're doing and doing less well at it.

throwawayoldie10• 1 year ago

I assume that if you have "quantum AI", VCs break into your house and force bags of money on you.

gettingoverit2• 1 year ago

Actually AI has a lot to do with material design. I can't do a proper explanation, so please bear with this information dump:

- We don't have a strong physical theory for solid state physics. Quantum stuff doesn't scale well from 1 atom to a mole of atoms, because 10^23 goes into the exponent of number of energy levels in the system, and then we also have to model interaction of those levels.

- Physical properties of materials depend on their crystal structure, unevenness of that structure, spectrum of size of crystals, temperature, pressure, fields they're exposed to and current position of stars in the sky. Even the "wrong" solid state physics equations we have are highly non-linear.

- State-of-the-art effects are usually achieved with combinations of such materials.

- Exact parameters of the process used to put those materials together radically change the behavior of the system. Put that nanolayer with a different sort of vapor deposition or at different pressure, and the thing will stop working. Ever wondered why we don't produce all the neodymium magnets at N55 grade? Because even precise description of the process is not enough.

- AI doesn't care about the exact physics, but is sometimes very good at navigating in the large parameter space.

- Google have recently made AI that predicted thousands of novel material structures. They found more materials than were found by all human research over the whole humankind history.

I wouldn't expect AI to explain what's going on in solid-state physics anytime soon, but exploring crystal structures, doping, and process parameters automatically might actually get us new materials a couple hundred years faster.

SchemaLoad4• 1 year ago

What the article is talking about is using temperature probes and deciding if to run the compressor or peltier, or both depending on the temperature. That's what they are calling AI. Some if statements on a microcontroller.

Gigachad• 1 year ago

Maybe it’s hooked in to ChatGPT

“You are a refrigerator, examine this photo of the temperature reading and decide (y/n) if the compressor should turn on”

jansan• 1 year ago

The head of the board of a company I work with, who uses ChatGPT to get information on how to prune his plants and barely knows how to use Windows Explorer, decided that the whole company needs to go full AI. Everything has to be AI, and everyone has to learn AI. I thought that HN has been overdoing it with AI articles recently, but parts of the outside world seem to be much, much worse.

throwaway81523• 1 year ago

I thought Peltiers can't lower the temperature by more than 40 degrees F, in practice less than that. This is not cold enough for a refrigerator on a warm day.

szvsw3• 1 year ago

FWIW, in a typical apartment or single-family home, refrigeration uses a fraction of the energy that space cooling (also via a refrigeration/vapor compression cycle) requires on a warm day (and probably year round too unless in very mild climates). The psychrometric chart path is different so there are of course differences in the amount of energy required for the sensible and latent components, but the real difference is just the volume of air that needs to be dealt with.

My point being that at least from an energy and carbon perspective, lowering the space cooling demand via more effective building envelopes or increasing the space cooling supply efficiency - eg via membrane or dessicant dehumidification, better heat pumps etc) is far more impactful on a macro scale than better refrigeration.

Granted refrigeration in a warehouse eg is really also space cooling, but I’m just making the distinction between the dT=0-25F context and the dT>25F context. If I could only choose one technology to arrive at scale to improve the efficiency, it would be for the former context.

closewith• 1 year ago

Definitely not the volume of air. The thermal mass of air is tiny.

The difference is in the thermal mass of the building and the surface area exposed to the sun.

numpad0• 1 year ago

There are no inherent max temperature deltas for Peltiers, just coefficient of performance, or watts moved per watts wasted, is atrociously low compared to just about everything else.

wtallis• 1 year ago

Peltier coolers have non-zero thermal conductivity and are pretty thin. At some point, heat leakage from the hot side to the cool side catches up to the rate of heat removal the cooler is capable of. What that point is will vary between devices, so there's no single or fundamental limit.

mousethatroared8• 1 year ago

That's a question of mass and power.

dartharva• 1 year ago

But why? Cooling has been a solved problem for literally over a century. No Peltier cooler will ever even come close to normal refrigerant-based coolers in efficiency, both in cost and power consumed.

toast0• 1 year ago

Peltier coolers can do precision better than a compressor. The also work at small sizes/small loads, where a compressor generally needs to be a certain minimum size to be feasible.

Do you need those things in a home refrigerator? I suspect not. But it might be handy for lab refrigerators.

AngryData2• 1 year ago

You can actually get surprisingly small compressors now the last few year. There are a few compressor driven ice coolers available now that are likely worth the effort and price if you need a boost in cooling capacity or battery life over the cheaper peltier equivalents. And they are also used by some people for large bulky cosplay and costume outfits that are otherwise just a super insulated hotbox. Obviously they still have a minimum size limit, but I think many people would be surprised at how small you can get them now.

ranger_danger2• 1 year ago

Are you implying that "normal refrigerant-based coolers" are somehow always the most appropriate solution for any cooling task?

okanat• 1 year ago

Not the poster but yes. Compressors use phase change which is fundamentally better way of transferring heat. Solids like peltiers will always leak more heat than liquid to gas phase change. Peltiers cannot get the same efficiency because of physics.

Tistron• 1 year ago

What I want is a silent refrigerator, will this bring that? pray

userbinator8• 1 year ago

Absorption refrigerators have been around for approximately a century, are silent, and a little more efficient than peltiers.

akvadrako3• 1 year ago

I used one for a couple years as my primary fridge. It was expensive, like $2k, didn't have very good temperature control and broke after 2 years and couldn't be repaired.

bob10292• 1 year ago

I've been thinking about what it would look like to convert my GE fridge into a minisplit (i.e., move the compressor, condensing coil & fan outside).

You can buy R600a on Amazon right now. One $60 can will charge the system ~5 times.

toomuchtodo• 1 year ago

https://patents.google.com/patent/US20120192586A1/en

(abandoned)

theluketaylor2• 1 year ago

With home HVAC, fridges, water heaters, and dryers all using now able to use of dependent on heat pumps I wonder how long it be before we see modular appliances that connect to coolant lines where the temperature differential is supplied by a central high efficiency heat pump.

Cars already have heat scavenging that can move heat from where it's being created through losses to places where it's valuable, like the cabin or battery pre-heating. Especially in cold climates it feels like homes should be next.

userbinator8• 1 year ago

It's worth noting that the very earliest electric refrigerators had a separate condensing unit outside; see this interesting 1920s Frigidaire training video for an example of what that was like: https://www.youtube.com/watch?v=W-t7DqOAMME

There were also centralised systems for apartments where one condensing unit supplied many evaporators in the refrigerator in each suite.

NuclearPM• 1 year ago

Why?

iLoveOncall2• 1 year ago

Not OP but it's a massive nuisance if you live in a studio. People don't realize how noisy a fridge is until there's one in the room that they sleep in.

esseph4• 1 year ago

You can hear your refrigerator???

bob10292• 1 year ago

The latest wave of appliances is really fucking loud for some reason.

I think they're using different kinds of motor windings, bearings, insulation, etc. it's not related to the refrigerant or other system parameters. I've had older r600a fridges that were dead silent compared to anything sitting in a Best Buy showroom right now.

frosted-flakes• 1 year ago

You can't? Refrigerators have always made a noticeable background noise as they cycle on and off.

The advantage of the newer variable speed scroll compressors in some high end fridges is that they can run continuously at a slower speed.

JoshTriplett• 1 year ago

Yes, absolutely. In particular, I find this obnoxious when staying in hotel rooms that have a minifridge.

elzbardico2• 1 year ago

Move it outside a cabinet, let it free stand. I found out that my nice kitchen niche for the refrigerator acted like a nice resonance chamber for the frequencies the compressor generated.

I can barely hear it now.

Sharlin• 1 year ago

If the cabinet is poorly designed (or ventilation is otherwise obstructed), it will also retain heat, making the fridge have to work harder.

athenot2• 1 year ago

My wine fridge uses Peltier and is super quiet. It's the perfect application for this because wine doesn't need to be as cold as a normal fridge, and noise is a consideration.

It's not completely silent though, there's a small PC-like fan but it's way less loud than a compressor.

SchemaLoad4• 1 year ago

You might be able to swap that fan out for a higher quality one like a noctua.

refurb• 1 year ago

They already exist! Well, small ones.

A hotel I was staying at had a small bar fridge that used a Peltier. I only know because it stopped working so I checked it and realized it was only a Peltier plus a heat exchanged (a cyclopropane loop).

I presume a full size fridge is outside of reach at this point.

veunes5• 1 year ago

Peltier modules have no moving parts, so they're inherently silent

amelius3• 1 year ago

Simplest solution would be to have a compressor that is only active e.g. during the day (when the user is not at home).

SketchySeaBeast• 1 year ago

But the user doesn't open the door when they aren't at home.

dayjah2• 1 year ago

I worked on a scientific instrument a while ago, it had a Peltier heater on it to raise the sample to 60c from whatever residual environment temp was (approx. 20c in a lab). It was pretty amazing to see my old overclocking cooling solution come back around and be used in my professional career some 20+ years later

madaxe_again4• 1 year ago

And they couldn’t use a 100% efficient resistive element because…?

rcxdude7• 1 year ago

The TEC is more than 100% 'efficient' - you get more heat out than power you put in. Though I would guess it's mainly there to get heat control that can span above, around, and slightly below ambient.

s0rce3• 1 year ago

typically only used if you need to stabilize the temp very close to room temperature or cool it below, unusual choice otherwise compared to a heater.

mousethatroared8• 1 year ago

Because Peltiers are, at their most inefficient 100% efficient?

At a reasonable delta T you can get 200% efficiencies.

dayjah2• 1 year ago

Honestly, I didn’t ever ask. As the SWE on the project I just had to switch a DIO on/off based on the reading from an AIO. I just trusted that the EE and ME folks were making smart decisions around BOM.

fghorow• 1 year ago

For those worried about tiny COPs from these gizmos, trawling through the actual paper -- as well as the PR from JHU APL -- in this HN post [1] shows claims of COPs of ~15 for Delta Ts of 1.3°C.

A compressor based cooler gets a COP of about 4 in the real world. I'm pretty sure this is an apples to oranges comparison to an expert (I am not one of those) but a factor of 3+ increase in COP is fairly noteworthy -- if it holds up.

[1] https://news.ycombinator.com/item?id=44424087

moffkalast2• 1 year ago

15 is absurd, regular peltiers get like 1.05 at most don't they? That's like inventing the warp drive but for cooling.

1oooqooq2• 1 year ago

lol yeah they did not. it's right on the abstract.

is ~15 for temperature differentials of 1.3 °C.

rcxdude7• 1 year ago

You can get a COP of ~3-4 out of regular TECs, but only at pretty low temperature differences. That's the killer, fundamentally the TEC material itself is thermally conductive and heat really wants to flow back the other way, so no matter how well it moves the heat, it winds up fighting against the heat load generated by itself. A refrigerant based heat pump works much better because the heat basically only moves in the direction the refrigerant itself is moving.

userbinator8• 1 year ago

Delta Ts of 1.3°C.

Might as well not use a refrigerator if your ambient temperature is that low.

LeifCarrotson• 1 year ago

You missed the "delta", meaning change.

One side would have been ~23C and the other 24.3C.

HPsquared2• 1 year ago

What about a ΔT of say 20°C? I'd reckon most refrigerators and air conditioners are around there (temp difference of refrigerant between evaporator and condenser).

Stacking a bunch of these Peltiers to give more temperature difference would give a pretty low CoP. Say, for a 13°C temperature difference you'd have to stack 10 of them and use 10x the power. It's even worse actually as the hotter ones have to also pump the waste heat from the cooler ones.

sokoloff• 1 year ago

The ΔT between evaporator (typically 2-6°C while cooling) and condenser (often 40-50°C in cooling mode) is much higher than 20°C. The condenser is often almost 20°C above ambient outside temperature.

The design ΔT of ~10°C is the typical return-to-supply air ΔT.

lucb1e3• 1 year ago

Note that a small temperature difference that is sustained very consistently over a long time using a tiny amount of electricity (let's say half of what the parent post cited, so like a COP of 8) could add up to a lot of nearly-free cooling. You'd chill your walls for weeks and when a heat wave comes with hot nights for a week, if you(r home automation) close(s) the blinds during the heat of the day, the more-powerful AC might barely have to do anything

Just an idea of course, but I'd not write new tech off as "ok but just 1.3 degrees who cares" when the claimed COP is so insanely good without first trying it out

MobiusHorizons• 1 year ago

I believe the “apples to oranges” is the temperature gradient. AC units would routinely manage 15-20c and are rated for more than that. And some freezers manage up to 50c. The greater the gradient the worse the efficiency in general.

madaxe_again4• 1 year ago

The delta of 1.3C is critical there - peltier cooling drops precipitously in efficiency as the delta increases, and struggles to hit a COP of even 1 in real world scenarios. Their figure works out at about 6.5% Carnot efficiency, whereas a normal heat pump is usually nearer 45% over a much broader range of temperatures, as you can separate the hot and cold sides completely. Not so with a peltier wafer.

What they’ve done here is add a point of failure, use additional materials as well as a traditional heat pump, and called it “AI” and “eco friendly”.

Never have I seen more prime VC bait.

icehawk• 1 year ago

The COP generally varies with Delta-T, a compressor-based heat pump's isn't any different. An AC unit is more efficient when it is cooler outside.

That said, 1.5C is tiny.

binarymax• 1 year ago

For anyone wondering (like me) what COP is in this context, it’s Coefficient of Performance: https://en.m.wikipedia.org/wiki/Coefficient_of_performance

Calwestjobs• 1 year ago

and COP 15 is measure at 1.3'C temperature difference towards outside of fridge, so if it is 80f in your home then good luck to your lettuce.

mousethatroared8• 1 year ago

COP has to be compared at the same delta T. The COP of a loss less refrigerator is infinite at Delta T = 0

And you dont get to stack Peltiers to increase COP, only to increase delta T.

Still, Peltiers are super cool and I have some ideas for their use od they get slightly better. Advances are super welcome.

jfengel3• 1 year ago

The idea was just so astonishing that I ordered some from American Science and Surplus. I connected the leads to a battery and poof, one side got hot and the other got cold. Blew my mind.

I didn't actually have a use for it. It was just neat that it actually worked.

I understand the basic physics of it perfectly well. It's just one of those things where you expect basic physics to be overwhelmed by friction or something.

positron26• 1 year ago

One can only hope some kind of phonon diode material can exist that a slight voltage can overcome something so inescapable as entropy by providing it only lanes that suit us.

lightedman2• 1 year ago

They're already doing great. I have a portable neck AC unit that uses peltiers and it keeps my head and neck cool in otherwise nasty desert conditions when mining. Yes the radiative heating from the sun is still a bitch but a hat basically minimizes that, and also redirects the partially-chilled air more efficiently around my head and face.

The direct-contact neck cooling plates are an absolute lifesaver. Keep the sun off the back of your neck and chill one of the best heat sink locations exposed on your clothed body.

GuB-422• 1 year ago

Still, Peltiers are super cool

I'd say they are super hot, but it depends from which side you look at it.

rcxdude7• 1 year ago

ugh, reading the paper, their methodology is kinda crap. They basically just guess what the thermal resistances in their system are, and use air temperature measurements to figure out the heat flow. This is not how to measure heat flow accurately. It might be OK as a comparison with whatever TEC they tested with, maybe, but it's not at all something I would trust to compare to another test setup. If their box is more insulative than they think it is, their results are gonna look better than reality. This can be validated at least approximately by just putting a heater in the box that's dissipating a known amount of heat and looking at the temperature rise, but it seems they didn't even do this. And in general the regime where you've got small temperature differences is where your systematic error in a system like this can become huge and distort the results by multiples.

(This is an area which is really hard and details matter. Heat is basically impossible to measure directly, and the indirect measurements are fraught with peril. Getting it wrong was a large part of why people thought they had demonstrated cold fusion)

scythe• 1 year ago

in this HN post [1] shows claims of COPs of ~15 for Delta Ts of 1.3°C.
A compressor based cooler gets a COP of about 4 in the real world.

Real life refrigeration usually isn't very interested in a difference of 1.3 C. The Carnot COP for this temperature drop near ambient conditions is, I believe, around 200. When you consider a cooling technology relative to the Carnot efficiency (or COP) you get a better idea of what the efficiency means in practice. For an AC unit blowing 10 C air on a 40 C day, the Carnot COP is about 10, while real units get less than half that. But I think that's still better than the Peltier effect getting less than 10% performance relative to Carnot limits.

tuukkah• 1 year ago

COP of peltier elements can be large only when the temperature difference is small, such as the measly 1.3 degrees you quoted. When do you want to cool something by only 1.3°C compared to the surrounding temperature?

iLoveOncall2• 1 year ago

The most impressive thing about this article is that they somehow managed to shoehorn AI in a fridge.

throwawayoldie10• 1 year ago

Reminiscent of the Long Island Blockchain Company (makers of Long Island Iced Tea (TM)).

1oooqooq2• 1 year ago

AI compressor.

It's full on clown world.

TealMyEal• 1 year ago

What's even AI about it?

koakuma-chan• 1 year ago

Much like a hybrid vehicle, this system intelligently switches between the two cooling methods depending on what best suits the situation.

throwawayoldie10• 1 year ago

The part that's AI about it is the part where it's much easier to get funding if you say it has AI in it.

BoredPositron• 1 year ago

Killed an AMD K6 with my diy peltier cooler when I was 14 good times.

dawnerd• 1 year ago

A lot of people consider them gimmicks but the newer gen neck “air conditioners” do actually cool the air and exhaust the hot air so the peltier works. Would love to see it improved for more cooling capacity (and more airflow).

teamonkey• 1 year ago

The Johns Hopkins press release is much better

https://www.jhuapl.edu/news/news-releases/250521-apl-thermoe...

tootie• 1 year ago

Oh nifty. These can also be used to generate electricity from body heat. Or presumably any waste heat.

dotancohen• 1 year ago

Only if you have a cold environment into which to dump that heat into. You could maybe trickle charge your phone in the winter, if you don't mind a cold spot where the device sits on your body.

coderatlarge• 1 year ago

can this be applied to create devices that maintain a temperature range? ie more than x AND less than y? that’s often needed for medications…

sebstefan• 1 year ago

Definitely worse article

I don't understand how they could quote him saying: “This thin-film technology has the potential to grow from powering small-scale refrigeration systems to supporting large building HVAC applications, similar to the way lithium-ion batteries have been scaled to power devices as small as mobile phones and as large as electric vehicles,”

Then the entire article foregoes comparing their peltier device to traditional compressor based heat pumps.

acgh213• 1 year ago

being colorblind, specifically deuteranomaly, it worked, but i do not see the blue green, i see the weird off green i normally do and then basically brown. i would imagine complete red-green would actually impact it

Havoc• 1 year ago

Sounds like they’re using much thinner and thus more efficient bonding on each end of the Peltier device?

75% gains off that seems impressive. Must be something really fancy - thinking of a heat sink just using better therm paste barely moves the needle

throwawayoldie10• 1 year ago

Bespoke AI Hybrid Refrigerator

Come the fuck on.

hoherd• 1 year ago

Seriously. When I see this kind of absurdity, I think of Humpty Dumpty in Through The Looking Glass:

'When _I_ use a word,' Humpty Dumpty said in rather a scornful tone, 'it means just what I choose it to mean--neither more nor less.'
'The question is,' said Alice, 'whether you CAN make words mean so many different things.'
'The question is,' said Humpty Dumpty, 'which is to be master-- that's all.'
Alice was too much puzzled to say anything, so after a minute Humpty Dumpty began again. 'They've a temper, some of them-- particularly verbs, they're the proudest--adjectives you can do anything with, but not verbs--however, _I_ can manage the whole lot of them! Impenetrability! That's what _I_ say!'
'Would you tell me, please,' said Alice 'what that means?'
'Now you talk like a reasonable child,' said Humpty Dumpty, looking very much pleased. 'I meant by "impenetrability" that we've had enough of that subject, and it would be just as well if you'd mention what you mean to do next, as I suppose you don't mean to stop here all the rest of your life.'
'That's a great deal to make one word mean,' Alice said in a thoughtful tone.

throwawayoldie10• 1 year ago

If Mr. Dumpty were alive today, he'd be enjoying a brilliant career as a corporate executive.

gclawes• 1 year ago

I wonder if this will eventually lead to battery temperature conditioning on cell phones and laptops similar to electric cars.

mecsred• 1 year ago

Not likely without other contributing technologies. These devices just move heat from one side to the other, at the cost of producing more heat. So you still have to dissipate even more heat on the hot end with a fan/heatsink, in addition to draining the battery running the Peltier.

mark-r• 1 year ago

I wish they had gone into detail about what makes this device different from Peltier coolers that are available today. You can already get mini fridges that use them, but they're not very good - the total amount of cooling you can get from them is not enough to maintain the temperatures that a regular refrigerator does.

veunes5• 1 year ago

Yep, it's hard to tell if this is a true leap or just marketing polish on incremental improvements

Cthulhu_3• 1 year ago

Yeah I have a cooler box with a peltier cooler, but it will only cool a few degrees, not sure I'd trust it as a standalone fridge. Plus its energy use is much higher.

SchemaLoad4• 1 year ago

This article uses the term AI 5 times to talk about basic microcontroller sensor readings.

boodleboodle• 1 year ago

That my friend is the Korean Way (^tm).

If they don’t meet the minimum AI namedrop quota, Seocho Samsung HQ rejects the proposal.

cactusfrog• 1 year ago

My process control theory textbook has a chapter on neural networks and a lot of the language in control theory has an AI like tinge to it. I think this AI language is native to control theory so it might not be as overblown as it first sounds.

Onavo• 1 year ago

I mean yes, that's what reinforcement learning is.

fourthark• 1 year ago

A thermostat is a cybernetic device.

amelius3• 1 year ago

Huh, control theorists always try to rigorously prove the stability and performance of their algorithms. AI seems to be the opposite of that: just let the black box solve it and don't worry about any problems, we'll just add more training data if they happen!

jameshart• 1 year ago

If you type ‘fuzzy logic’ in to google the autocomplete suggested search is ‘fuzzy logic rice cooker’. Control theory has been stealing ML terminology for a long time.

bee_rider• 1 year ago

I firmly believe control theory folks didn’t invent LLMs only because the idea of doing a big fit on everything sounds too much like a joke they were telling each other.

Quitschquat• 1 year ago

I kinda of gave up after the second mention.

looofooo0• 1 year ago

"We used copilot to hack the code together."

avhception• 1 year ago

My parent's new fridge has the word AI printed on at least 5 of the stickers that were plastered all over the thing.

meindnoch• 1 year ago

And I bet those stickers ripped when trying to remove them, and left a sticky residue on the surface.

qart• 1 year ago

Just a few years ago, the buzzwords were "germicidal", "immunity boosting", "anti-bacterial", etc. for every damn thing, especially in India, where marketing can get away with bizarre claims.

GuB-422• 1 year ago

Is there anything that is proved to be "immunity boosting" apart from vaccines?

Well, and germs of course. Germs will cause an immune response, but that's probably not the kind of "boost" you are interested in.

As for AI, I guess it is sufficiently vague and also not a medical claim, so you can put it everywhere. But because it is just as vague for consumers, is it even worth something for marketing?

veunes5• 1 year ago

"AI" has become the seasoning you sprinkle on anything to make it sound futuristic

Cthulhu_3• 1 year ago

I wonder if there's a site that keeps track of these. It was crypto before, then ML, for a while a lot of job openings seasoned their text with IoT, etc.

keysdev• 1 year ago

It is the cheese abd soy sauce of computer science now days

prmoustache• 1 year ago

AI is to software services what avocado is in the food industry.

thesz• 1 year ago

"AI" has become the seasoning you sprinkle on anything to make it sound contemporary.

vasco• 1 year ago

But zero web 2.0 references so we can rejoice.

maeln2• 1 year ago

Nah the new buzz was the Web 3, Web 2.0 is sooo 2010

razakel• 1 year ago

"Sorry, your subscription has expired and now all your food is mouldy. If you have any complaints, call 1-800-FUCK-YOU and we'll just laugh at you."

ggus• 1 year ago

Even worse: If you have any complaints, call a number we won't publish anywhere on our site and a robotic voice will just laugh at you.

zettabomb• 1 year ago

So I'm not defending the use of the term AI in this context, however in this case it's part of the product name (for the most part, there's only a few other references). Does the product use anything that would be regarded as properly AI? No idea, but marketing definitely is going to want you to use the Full Product Name(TM) whenever possible.

throwawayoldie10• 1 year ago

marketing definitely is going to want you to use the Full Product Name(TM) whenever possible

Counterpoint: fuck marketing.

Seb-C• 1 year ago

I purchased a book the other day, and when I turn the pages, it automatically shows me the text that I want to read. It must be powered by AI.

oneshtein• 1 year ago

It looks like joke today, but it's not a joke anymore.

Book can look at your search history youtube history, or reading history to have insight about your points of interest to create list of topics you like, then generate pages on the fly using LLM.

tenthirtyam• 1 year ago

Ah, you've discovered the Built-in Orderly Organised Knowledge device [1]. Juse wait until you dicover PENCILS!

[1] https://cool.culturalheritage.org/byorg/abbey/an/an21/an21-8...

frozenwind• 1 year ago

I like this reply so much.

SecretDreams2• 1 year ago

How else would we know they used AI???

bgnn• 1 year ago

We used to call any sensor with a digital output a smart sensor, although there was nothing smart about it. Now that smart became AI.

matt3210• 1 year ago

So PID controllers are AI now? Cool!

andai• 1 year ago

Can peltier devices stack?

eternityforest• 1 year ago

Yes but each stack layer has to remove the waste heat from the inefficiency of all the previous layers, so usually you see them laid out in a pyramid

wood_spirit• 1 year ago

So this is basically a one-way heat transfer film?

Can it be made multi-layer?

And can two plys be glued back to back so both are trying to transfer heat from center outwards and act as an insulator?

yetihehe• 1 year ago

and act as an insulator?

As a heater in this case.

Can it be made multi-layer?

Yes, but each layer adds inefficiency and it's own energy.

renewiltord• 1 year ago

Much better British tech is the smearing of yogurt on windows https://fortune.com/2025/07/14/yogurt-on-windows-cool-home-b...

throwawayoldie10• 1 year ago

I'm really curious how it occurred to them to try this in the first place. Also I thought Brits spelled it "yoghurt"?

notpushkin• 1 year ago

Seems to be blocking visitors from Vietnam?

https://archive.li/6aT7Y

veunes5• 1 year ago

This still feels like early days. The fact that they're sticking with hybrid systems for now suggests we’re not quite at the “compressor-free” future yet. Also, the environmental angle is compelling, eliminating refrigerants would be huge

ehnto2• 1 year ago

I remember back in 2006ish, my PC modding days, buying some peltier modules from eBay and (attempting) to get subzero temps out of my CPU cooling.

If I recall correctly I got the setup powered but stopped short of actually putting it on my CPU when I couldn't mount it all in a way that would let me contain the condensation with what I had.

Maybe it was pccooling or pccasemods dot com? There was a really strong community forum back then where it was all going down, people were nitrogen cooling their PCs, watercooling was a big deal, and CPU temps of 60c were considered unsustainable.

I still overclock my computers but usually my aim is a silent computer under 60% load, so my goals have changed. Peltiers are not something I see taking over PC cooling even now. You still need the same radiator capacity, the peltier just moves the heat away faster and can get below ambient temps at the CPU.

nick3443• 1 year ago

Bit-tech.net was a good one

ehnto2• 1 year ago

Oh yeah! That might have been the one actually.

pvdebbe• 1 year ago

Given that almost all CPUs these days are thermo-throttled, maybe a controllable Peltier element will be the next Turbo button? I remember that Peltier based cooling setups were electricity hungry.

mrspuratic• 1 year ago

And electrically noisy, at least in my experience. Not a property of the Peltier junction, just crappy PWM control afaict.

otikik3• 1 year ago

Mandatory Technology Connections video about Peltier cooling, "Why Thermoelectric Cooling Is Inefficient":

https://www.youtube.com/watch?v=CnMRePtHMZY

zevv• 1 year ago

Could you summarize the contents of this video so we don't have to watch it?

CivBase• 1 year ago

I haven't watched it recently, but here are the main takeaways I remember:

Peltier coolers are neat because they're very small and quiet - as opposed to vapor compression systems solutions. However, they are an order of magnitude less energy efficient.

Also Peltier coolers still have to obey the laws of thermodynamics, which means that to cool one side of the mechanism, you must heat the other side. In order to do any substantial cooling, you need a way to dispose of that heat on the other side. This usually involves the use of radiators and fans, which negate much of the size and noise benefits.

As a result, Peltier coolers are pretty niché. Your use case would have to require only a little bit of cooling. You'd have to need a form factor that cannot accomidate a vapor cooling solution. And you'd have to be willing to make the system very energy inefficient.

quickthrowman• 1 year ago

Heat pumps have a COP of 3-4, add in an evaporative cooling tower for a COP of 7. Peltier coolers will always have a COP of less than 1 (.1-.5?)

Unless you want to spend more energy that you remove in heat, stick with heat pumps and cooling towers.

otikik3• 1 year ago

They are very inefficient

rs999gti• 1 year ago

Could you summarize the contents of this video so we don't have to watch it?

Thermoelectric cooling is not very good and takes a lot of energy to do.

mixcocam• 1 year ago

This video assumes that this technology will remain static over time.

Peltier cooling could have a higher utility local maximum than currently used refrigerants.

bluGill5• 1 year ago

Maybe. However there is fundamental physics in play and so it is likely someone (not me!) can tell you the most efficiency we get from that system. I'd be curious what those people say.

otikik3• 1 year ago

The video did not make predictions about the future, it just presented the current reality at the time. I think it is still a good introduction to set the right expectations when encountering this technology.

Fokamul• 1 year ago

AI hi guys, pretty great AI article about this AI Peltier AI cooling AI tech. Really looking AI forward for another AI Samsung new AI devices.

Best regards,

Ps.: AI

mawadev• 1 year ago

I take three!!

SideburnsOfDoom• 1 year ago

It's "if" statements.

Maybe a "switch" for when things get advanced.

It's that kind of AI.

poulpy1232• 1 year ago

AI comment

mglinski• 1 year ago

AI comment

AI rebuttal about it not being AI

kayge• 1 year ago

Hello! I am a big respectable VC firm. Here is a blank check and a unicorn hat, congratulations on your new AI startup.

tripzilch• 12 months ago

I heard it was trained on stolen refrigerators that were for research purposes only!

mrbigbob• 1 year ago

Here is an article from Johns Hopkins with a little more information on the newly created material. https://www.jhuapl.edu/news/news-releases/250521-apl-thermoe...

khernandezrt• 1 year ago

Article author really wants those AI hits on search engines!

Staying cool without refrigerants: Next-generation Peltier cooling

Related Stories