So there's this thing called Peltier Cooling.
What gives? Why are the big companies avoiding using this tech?
>slap it onto a water cooling system
>place it at the end of a phat fin and tube heat sink
>avoid the frost like a boss
Tell me someone uses it in their product line.
1. they're very inefficient
2. they're only relevant for sub-ambient temperature, something you do not need
Jews aka the same guys that hid cancer cures and free energy machines
surely they could be throttled with some sort of metering to keep temps below 40c
it would make far more sense to use a larger heatsink/fan instead of pouring watts into a peltier unit
watt-hours, i should say
but that temperature control by juicing sounds nice
why not? given theres a way around the frost in the design
>companies working for a decade to have energy-efficient chips, now at 65-85W TDP
>'hey guise let's cool them using a component that requires 110W to produce a decent temp difference'
then what about them older CPU's that run hot as hell?
a 5 dollar addition maybe makes the difference in a budget build?
if you're that tight on money, you should be looking into something that won't shift the cost difference onto your power bill
Clarification for OP
One side is cold, but the other is hot.
Guess which one is producing bigger temperature? (Spoiler: Peltier produce more heat than cold, they easily burn themselves without sufficient cooling on the hot side)
exactly
where do you think that 100W you're putting into the unit is going?
if your HSF is already having hard time with the cpu alone, it's not going to do any better with a peltier adding even more heat into it
hence why it only makes sense with sub-ambient requirements, anything above ambient is better done with regular air cooling or liquid-circulated air cooling ("water cooling")
It sure looks like summer.
You'd known why if you took the time to look up the efficiency of these things.
next time: No. These things are completely useless for PC cooling since they only move the heat to the heatsink while adding more heat.
To use one for CPU cooling you'd need a much bigger heatsink on the hot side, which you could have put directly on the CPU instead and saved power.
heat hasn't been a big problem since the mid 2000s every thing has been able to reach max clocks possible on air. sure you can freaze some thing and put massive voltage in it and get it running fast but its not safe or stable and they would never release some thing like that due to warranty and stability and yer like some one pointed out even if you released a pelter cooled cpu with high clocks and high voltage it still wouldn't be stable enough and would use massive power.
max stable clocks can be done on air even watercooling has been pointless for over a decade. you can get a higher number with extra cooling but its not stable enough and would crash daily.
Again here
To be honest, I was thinking about DIY a peltier system but for laptops, but with a low power peltier and controlled by an arduino. That'd make more sense if it'd turn on and off once in a while and would be outside of the case, so heat dissipation would be a smaller issue than on desktop. But generally this is not something normies would buy like they buy better fans and install them in a desktop out of the box.
>avoid the frost
Are you retarded? Frost? What? This isnt NO2
lrn2thermodynamics. using a peltier to make the heatsink hotter makes the heatsink much better at dissipating heat
Peltier coolers are used for devices which don't need to run for extended amounts of time.
Digital camera backs sometimes use them for instance, you need that device to be on only as long as your exposure is. A peltier is fine there.
And the peltier becomes less efficient the hotter the hot side is, plus they much like CPUs can't get too hot on the hot side before they break. (100-150c)
And the maximum temperature differential a peltier can achieve is ~70c.
So... no. The heatsink will still have to be at a reasonable temperature, meaning the slightly higher temperature the heatsink can be at before the system fails will not offset the extra heat the peltier adds.
And... why? You could just add a bigger heatsink in the same space and save the power.
>add a bigger heatsink
To a laptop or a SFF?
See
>would be outside of the case
I think you assumed too much based on my posts.
Cutting a hole for a small 15x15 peltier+a fan with few copper tubers, that can be easily closed with hatch, a 3D printed part, and adding a consumer tier heatsink requiring hacking 1/3rds of the case are two different things.
they use a stupid amount of power and you require a large heatsink still relative to the size
and putting the cold side on a cpu will just create condensation and issues
theres literally no point, the heatsink and fan is just keeping the hot side relative for the cold
they're pretty much only used on those shitty can coolers you see
For a peltier based cooler to actually be effective in a CPU scenario, it would need to consume a substantial amount of power (think 100W+) & also produce a substantial amount of additional heat (like a similar 100W+). This is a hard sell that pretty much only applies to enthusiast overclocking crowds that want to use peltier chilled water cooling loops for a cheaper solution to extreme overclocking challenges than cascade phase change.
You do realise that the peltier itself _produces_ a net quantity of heat? So if you add a peltier to a laptop you would need a larger heatsink overall than that same laptop without the peltier?
>prevents battery from expanding
>saves your legs from getting too warm
i'm not op by the way, just trying to salvage his thread
You're going to need a custom heatsink either way, but if you use a peltier the heatsink will have to be bigger.
Wat. Your peltier cooled laptop will produce MORE heat than the same laptop without peltier cooling. The CPU itself might run cooler, but the total amount of heat coming out of the laptop will be higher. So your battery would be more likely to expand, your legs would get hotter and your battery life would literally half.
no i was changing the idea to put a peltier plate on the entire lower chassis of the computer. you could have cutouts for the parts that get really hot and the battery of course is mounted below.
What the fuck no it wouldn't.
and where is the giant heatsink
1 side gets very cold 1 side gets very hot
if you don't cool the hot side it burns out
I'm still I know about heat that's why I was the first person to mention it ITT. In laptops and sff and tablets and whatever the issue is to get as much heat as possible with a smallest opening. Outside I could be cooling the heatsink with a river, but the inside part of it has to be small.
Also it'd be controlled, would run only from time to time and it'd be a low power module, enough to be almost passively cooled.
>imagine that you are cooling a CPU with a power usage of 35W, using a conventional heatsink. Will the temperature drop if you add our example peltier element between CPU and heatsink? No. For a simple reason: In addition to transporting heat, peltier elements also emit considerable amounts of heat (and thus use considerable amounts of electricity
most need 6-10amps or more at 12V so they're inefficient to begin with
I worked on a couple projects using peltier. One side were way too hot, and the other side were way too cold, and the power consumption isn't good
h-heat pipes?
Jesus christ no they don't dude.
Also a peltier doesn't need to run full blast if the chip is idle.
>In laptops and sff and tablets and whatever the issue is to get as much heat as possible with a smallest opening.
No, for normal operation you need a device that is small, light, power efficient that doesn't get to hot. But you want to cool small form factor devices with a large external heatsink...
>Outside I could be cooling the heatsink with a river, but the inside part of it has to be small.
Again, why? Instead of cooling a laptop with a huge external heatsink you could just use a desktop.
>Also it'd be controlled, would run only from time to time and it'd be a low power module, enough to be almost passively cooled.
Which is just a over-complication. If you put a piece of metal where the peltier is you would get much better passive cooling.
is there an ac version of peltier? Ie what happens when you connect a peltier plate to ac (obviously I don't specifically mean mains electricity)
0/10
trust a tripfag to be a complete fucking idiot
>buying a variable power supple to limit current now so it's not "full blast"
Why are you shilling for an inefficient shit tier design when an essentially passive system (fans use next to nothing) works fine for 99% of users? A more efficient designs exist for the 1% use case. We don't use it because it's inefficient, there's your answer, you don't need it reiterated across 100 posts to get it through your head, fuck you.
Stop pulling shit out of your ass and just admit you know nothin summerfag. 100-150W per 10°C dT is good rule of thumb for big air coolers
Peltier elements consume substantially more electrical power than they transfer heat power. If you want a peltier element to remove 20W of heat from a laptop CPU, then that peltier element will consume substantially more than 20W of electrical power (which is mostly converted into heat on the hot side of the element) and the total amount of heat that you will now need to dissipate will be a lot more than the original 20W of the CPU.
So your total power consumption and total heat will substantially increase, even if the temperature of the CPU drops. This is why I said that peltier cooling for PCs only applies to niche scenarios like extreme overclocking, where total power and heat don't matter.
>Peltier elements consume substantially more electrical power than they transfer heat power
Not always
Yes, always.
>not always does the laws of thermodynamics apply
Question: Why the fuck would companies spend money on this bullshit when air and watercooling servers works perfectly fine?
notice the parts above 100%
go on explain that graph
REKT
That'd have to be a large heatsink. Your rule of thumb is useless when discussing everything from tablets to desktops and servers.
lol c'mon man its' obvious
>every thing has been able to reach max clocks possible on air
Have you not visited Jow Forums in the past year? There have been thousands of threads about Intel housefires.
>tablets, servers
energy efficiency matters for those applications. that means no peltiers because they are a dumb idea for that
Notice that there are no percentages anywhere on that graph. One axis is Coefficient of Performance and the other is current consumption. That graph actually refutes your claims.
Same reason that liquid gallium based coolers used to be a thing I presume, it allows for the failure prone parts (eg moving parts with bearings etc.) to be moved away from the components such that they can be replaced without having to power down the machine to pull it from the rack, which isn't an option for mission critical front line duties?
1. Shit efficiency, especially at high power (which is exactly where you want to use it)
2. Needs constant regulation to match dissipated CPU power (too little power to the element and your CPU overheats as the element can't transfer much heat passively; too much power and temperature falls below ambient)
COP is a measure of heat energy removed by a chiller (or introduced by a heat pump) relative to the energy consumed by it.
1.00 COP is pretty awful, it means that in order to pull 250W from the cold side, you need to supply an additional 250W to the Peltier element, and thus dissipate 500W from the hot side. And the area of the graph where COP is significantly higher is irrelevant for CPU cooling since a difference of 10 degrees is too low to be useful.
There's a reason why extreme CPU cooling solutions (eg Intel's lol28core demo) tend to use compressor-based chillers with COP of 2-4.
Combine with these three
amazon.com
amazon.com
MD-70RT pump on other side
The main problem is regardless of what you do, you have to cool the hotside.
you also can't really pipe the hot side away from it easily either, you are not going to get the fucking thing cooled well. honestly, a mini fridge with a bucket of water in it and tubes piping into it would be a better option for cooling then this is because at least there you get an overall benefit, also if you do it right, you could get a mini fridge to stash water or other 'keep this shit cold' stuff in.
peltier is interesting, but that's as far as it goes.
I almost forgot, I have a design for a 25 gallon bucket and a radless system, I uses water and bottled water frozen in ice to cool it. it would keep the computer nice and cool along with you decided how cool it needs to be, as warming 25 gallons of water up will take quite some time, and ice will cool it down faster then it warms up. would be interesting to know how big a water tank you need to go fully radless
fully radless without additional ice being put into it I mean.
If you're gonna have a pump and huge radiators, might as well just go with a vapor-compression chiller.
Put those 3 components outside ( the cpu loop is connected to the plate heat exchanger ). You should get close to outside ambient temps rather than indoor ambient temps.
You don't want sub-ambient temps anywhere in your PC due to dew.
My CPU idles around 6 celsius in summer and occasionally reaches 0 in winter. Am I fucked?
dew point or bellow temperatures.
But if the loop is linked to another loop via the plate heat exchanger, the new ambient temp is the outside temp though? So it's still above ambient?
Wat. If it's 20 degrees in your house and 5 degrees outside, you're going to risk condensation if you chill your computer to under 20 degrees. The fact it's 5 degrees outside is irrelevant.
Inefficient as fuck, you can just cool the CPU instead of the peltier module
youtube.com
This guy has his radiator outside in the snow and he doesn't have his motherboard covered up and he doesn't have a very strong pump for the loop.
I can't be bothered to read the whole thread, but here's the thing about peltiers - it takes energy to cool the one side and that energy has to go somewhere. You need a heatsink on the hot-side which can handle the heat from the CPU plus the heat the peltier generates. So unless you need your CPU below ambient temperature, it's redundant and just adds overhead.
...so he's probably insulated around the socket with dielectric grease etc. like the phase change guys do? The point is that it is not sensible in the mainstream sense to chill your CPU beneath ambient, _unless_ you go to convoluted lengths to protect against condensation - something you don't need to do if you just stick to ambient or above.
I actually tried this back in 2006 or 07 on my 939 system.
I had a pretty top end overclock for the socket on regular water, it was a dualcore Opteron 175 (2.2ghz 1.3v) oc'd to 3.2ghz 1.5v on a DFI Lanparty NF4-d.
Load temps on regular water @ 3.2ghz was 45-50c (pretty good water setup, pic related).
With a 225w peltier I could do about 3.4ghz stable with about 1.65 or 1.7v, and that loaded around 30-35c if not more, can't remember exactly.
I just remember the heat output from that setup was insane. My normal 3.2ghz setting was around 230w in cpu draw alone on full load, probably 300w+ on 3.4ghz, add to that the 225w peltier and over 250w on my overclocked x1900xtx..
Thats alot of heat.
Was a cool (not really) experiment, but just not practical.
What if you have the water coming in go through a radiator inside the house -> reservoir -> pump -> cpu waterblock? Could that serve to lower the room temp without using AC?
cpu waterblock -> rad1 -> plate heat exchanger -> rad2 -> for the cpu loop side
you can buy one on alibaba, but u wont want it
but did it run crysis?
Kek
I actually don't think I ever tried it, but I'm guessing it whould have trouble with it.
en.wikipedia.org
Seriously this is high school stuff.
surely you didn't melt the machine before crysis came out, and i find it hard to believe you never considered running at least the demo with a setup like that
The other guy didn't have his board insulated in the vid
So? One isolated example for 20 minutes of a video in a cold room with the window open and very low humidity now extrapolates to 'it must be fine always'?
I could show you a video of somebody riding a motorcycle down the freeway at 150mph with no protective gear on who doesn't die, are you going to extrapolate from that that 'it's perfectly safe to do that'?
It was my main machine until late 2009.
I was playing CS source, BF2/Vietnam and Cod4 around that time, crysis just didn't look that interesting to me and from what I heard it ran like shit if you didn't have c2d and nvidia 8800 series, which I couldn't afford.
If I tested the demo, I can't remember it.
X-ray cameras uses Peltier devices to stabilise the temperature of the capture chips.
I made a water chiller with 4 high-efficiency ones, these things work best with a high wattage capacity run at lower watts. The delta starts nose-diving if you drive them pass 70-80% capacity.
Because now you have even more heat to dissipate you dingbat
To cool some 105W furnace they will consume like 500W.
> sure you can freaze some thing and put massive voltage in it and get it running fast but its not safe or stable and they would never release some thing like that due to warranty and stability
Intel really thinks otherwise with the current housefires and planned products for this and next year, even back in the day AMD had the nuts to release a 1.48+ vcore cpu to reach 5ghz. Granted it did on air as good as in water but yeah.
>max stable clocks can be done on air even watercooling has been pointless for over a decade.
See prior response with the last 2 gens of mainstreame intel and future 10nm desings.
Shit b8, user.