AC Compressor Cooling Thread

Status
Not open for further replies.

lukem5

Distinguished
Aug 23, 2011
16
0
18,510
before I posted this thread I didn't realize there was an entire sub forum over at OC.net dedicated to phase change cooling. if anyone's interested.
this youtube https://youtu.be/4d0B0Dli-1g?t=5m9s is pretty much, my exact idea with a few changes. So it's already being done, it's a matter of perfecting the design.

This is the future of CPU/GPU cooling. It works like your AC, except the super-cooled refrigerant is piped onto your cpu. Really simple folks. And the good news is you only need a compressor a fraction of the size of your AC, fridge, etc.

yay , no more noisy fans, dust, heat and all the problems that go along with it. Plus insane overclockage.



So this is a work in progress - no progress has been made but the idea is there and damn is it good. I ain't the first to think of it and i KNOW someone has already done this just not sure where and i'm sure it wasn't in earnest or we'd see it on the market by now. I'll see you all here in the coming years as we transform PC cooling!!!
 
hurdles

1. Condensation. New case designs will be required. Think opposite of what we have now. It needs to be airtight. This will help protect your components from dust, critters, corrosion, humidity, salt etc etc and undoubtedly make them last longer (especially when they are frosty cool :)) There are methods to make a case both highly accessible and airtight (rubber gasket, hinges, etc)
As another poster has mentioned - after opening the case and resealing it, a dehumidifying function would need to be applied to remove all moisture from the air.


2. Electricity. This cooling system would use somewhere between 500-800w. This of course, is for a beast of a gaming system with a fat gfx card and CPU insanely overclocked - much smaller, less power consuming units could be developed for our beloved casuals. Depending on how much you game- the extra electricity costs from the cooling unit alone could go into the hundreds per year. However any gaming enthusiast that understands the implications knows this is a small sacrifice.

 
Are you with a company and is this considered a use case or mission statement?

You are correct, this has been done a few times and even sold online.

https://gizmodo.com/5053065/thermaltakes-xpressar-pc-case-uses-actual-fridge-compressor-for-cooling
https://www.bit-tech.net/news/tech/cooling/alphacool-eiszeit/1/

You might wish to re-focus your efforts to produce a unit that is a coolant chiller and the ability to support multiple components, rather than just cooling a CPU.

For example, I use a 5000 BTU window A/C unit that is gutted and has evaporator coil submerged in cooler full of coolant that I use to temp control my home brewery fermentation vessels. Yeast fermentation is exothermic, but nowhere on the levels a PC would produce. I can still chill the coolant below freezing if I like.

Otherwise, you're looking to develop a 'cooler' that only has specific uses and fitment, not something that can potentially fit any system or any form factor.
 
Good stuff Rubix. I was just banned at another forum for posting this so thanks for keeping me around!

I would certainly be using freon as a coolant, piped directly onto the heatsink (R12 was supposedly much better and if we could legalize it for PC enthusiasts only would be great lol - was banned for BS environmental reasons) because its gets colder and (I think) according to laws of physics is more efficient to keep the energy transfer to one liquid only instead of 2.

It may require a radical change in compressor cooling loop design - to create an even colder, more 'concentrated": refrigerant (maybe a shorter loop?) instead of the large evaporators designed for cooling air.

We are cooling a small block of metal only.

The other components (ram, HD) can be cooled passively and probably quite well from residual of the gpu/cpu heatsinks.
 
I hope I haven't stumbled onto the grail of advancing PC performance here, I don't want to change the world or nothing. But I can't see the harm in giving gamers, renderers and scientists a little more computing power.
 
Freon is banned from many countries but there are substitutes for it although not as efficient. It's possible to make a cooler/freezer using heat exchange instead of compressor. Years ago I had German "Linde" refrigerator using heater and CO2. Worked silently and for more than 30 years.
 
I find it amazing that this has been done by Thermaltake a decade ago and still isn't widespread.

A couple differences between my idea and Alphacools crap.

It don't use a sealed case. It chills a liquid first, lowering thermal efficiency and causing all the extra piping to act as a radiator. not to mention freon can get much colder giving you better performance. It also costs over a grand.

On mass scale production the price of this unit could easily match a simple window AC unit in terms of cost ( A couple hundred bucks) and potentially even less if redesigned specifically for just cooling 2 small heatsinks. Oh and i'm not a capitalist pig so knock off 50% from that price for retail markup.
That's right, a genuine compressor cooler for less than one of those prefabbed cpu standalone liquid cooling units and A LOT less than a custom cooling loop. Power to the PC people baby.

Also freon (r134) is not banned, r12 is however (and it also about 30% more efficient) and their crap excuse to replace it with potentially even more toxic r134 is it leaks from mostly car AC systems into the environment - well what if they just designed systems to not F*cking leak? Much easier for a standing PC unit than a moving, vibrating car.
 


True about the aerosol cans being the main source. But we don't need to worry about this - the system is closed and won't ever open unless deliberately.

co2 cooling looks very promising, however liquid helium seems to be the grail as far as the lowest temps go. Not sure if that could somehow be worked into a compressor cooler type situation with human resources, i'm sure it could but how bulky, energy consuming and overall practical would it be.

Surely what we'll be using in a century from now though. Actually nvm, all our pc's will be SPACE COOLED by then LOL.
 


It's not widespread, because:

It's expensive
It's not really needed
Too complex
Its loud (probably)

Lowering the temp does not automagically mean 'better performance' or higher OC.
Only if the system is actually thermal throttling.

CPU performance at 85C is exactly the same at 60C.
Now..if you're pushing the boundaries of your specific CPU, and you've reached a point in the voltage and GHz where the actual temperature is the limiting factor, then maybe.
How many people does that actually apply to? Not many at all.

Often, it is just a limitation of that specific CPU in your system. No matter how cool you make it, that chip will simply not go any higher.


That's right, a genuine compressor cooler for less than one of those prefabbed cpu standalone liquid cooling units
Under $100, silent, easy to install (30 mins), zero condensation....can you meet all those requirements?
 


Yeah under $100 is pretty out there but totally doable. The fact is this tech is easily worth $500+ and people would pay for it to have practical OC performance far greater than anything on the market now.

All the other issues you mentioned I addressed in my previous posts or are rather trivial matters TBH.

As far as thermal throttling, it seems most CPUs except the lowest binned will keep going up as long as you keep the temps down. What prevents most people from going higher is temperature due to increases in voltage..
The cooling is the key.
 
I want you all to imagine the equivalent of nitro cooling performance as a standard in cpu/gpu cooling. This is the end goal and technically it is possible AND practical - and I can guarantee your childrens children will see its equivalent in their gaming rigs lol.
 
They have done things with liquid nitrogen and/or oxygen, with CPU being in negative -150 C or so and still hit the roof...
So.. no, low temperature is not a simple way to blissful 10Ghz speeds...

Edit: Most could benefit from some extra cooling, since core of the CPU will always be hotter than the surface (And thus cooler surface leads to cooler core temps) Said cooling will still include condensation as a problem.
If it's not the CPU die, it's GPU die, if not GPU die then GPU VRM's (they run pretty hot too) and if not those, then CPU VRM's there are a lot of things to cool off once you go full blown overclocker mode.

As for the condensation, simply having the case airtight would only work until point X since any and all air has moisture in it and if you cool it, it starts to condense on cooler surfaces (usually delta T of 10 to 15C)

Since CPU is cooled, as is GPU, said silicon dies have their own heat exchangers (you circulate the stuff through them to cool them) and as such, their top surface is cooler than the bottom (the silicon is there) and as such, condensation. Yes, you could insulate your pipes and top of the cooler, only to move the point to sides of cooler.
and yes, you could also spray insulation foam around the CPU socket to keep it all fine and dandy, except that foam would also affect negatively cooling of other components under it and don't have magic below freezing liquids to cool them.

so.. yes, you would also need to possibly take into account of getting dehumidifier in the case too, to run after each opening of the case.
 
If it is more costly than $400 US, than you can expect it likely won't be purchased by many people. If you end up having a design you want to have tested and reviews, I personally do CPU cooler reviews for Tom's Hardware.

However, there is a caveat: we, as moderators, need to know if you are a legitimate vendor for this site and what vendor you represent.
 
I can't find it now but there was a thread on this at OCN. They basically used bait boxes, or built insulated enclosures that they cooled from outside. The computer was inside the sealed case. They would go to subzero temp. to freeze all the moisture out of the air. Then run the computer for long term cooling at temps approaching dry ice systems. They were circulating methanol /Glycol mix. but it could be done with the evaporator inside also like a refrigerated truck.
 


Please provide me with your vendor association so we can validate you since your user account appears to be using a personal email address.

Otherwise, please do not provide the impression that you are part of a vendor/company as doing so is grounds for forum sanction.
 


Is this an idea, or is this a product you are building/selling?

You've been asked more than once. If you wish to keep this thread alive, please respond.
Either in a PM to a Moderator participating in this thread, or directly in here.

Otherwise, this will simply be deleted.
 
Dear mods,
You are turning this into a negative debate thread. My intention from the beginning was to create an idealization thread of a practical cooling solution that will undoubtedly be used in the near future.

I know people hate change, and scoff, and mock, and attack it until its inevitable settling into normal, everyday use. It reminds me of that famous quote. Truth always goes through 3 stages. First, it is ignored. Second, it is violently opposed (seems like we are starting to enter this stage). Third, its accepted as evident.

And so I cannot hate you all for showing human nature, but I do ask that you kindly consider your words before posting them.

I do thank everyone that posted pertinent links and suggestion to the subject (unfortunately the only useful, reasonable posters 🙁 ) and I do hope this thread can continue into the future.
 
The fact that 2 and 2 haven't been put together yet its astonishing and testament to humanities lack of creative will to change for the better. Especially in an industry that is the forefront of human innovation in many ways.
 


You are mistaking negativity with us repeatedly asking if you represent a vendor or manufacturer, or if you are a normal forum user with an idea that has not been developed, tested or proven.

There is no 'negativity' other than the fact that you have not yet answered the question around your affiliation (or lack of) and what appears to be the hope that we're being negative so you can use that to fuel your agenda.

Again, no negativity, but you have avoided several questions around specifics and design and the most important one - are you affiliated with a manufacturer or a vendor, and if so, which, and can you provide proof of such so we can grant you vendor status.

Otherwise if you are just an individual with an idea that you're working, you need to be VERY clear that this is the case rather than implying that you have a product that has gone through design testing, validation and ready for beta testers.
 
I've posted links to a thread where this idea is being explored, and your concept would be welcomed there. So I'm not negative at all.
But I also don't understand why you won't answer the moderators question. May be you don't have a commercial interest now but would like to keep that option open (like an inventor would). But playing the misunderstood genius/martyr card doesn't serve any useful purpose.
You're not the first one to think of this, neither are you the furthest along with developing it. It's a somewhat complex and difficult technique due to the issues mentioned above.
The solutions so far are labor intensive, require fabrication skills more than production techniques. Also supercomputers have used this technology in the past, and probaly are still using it. There are several DIY, and coomercial techniques to deal with condensation, humidity issues.

https://en.wikipedia.org/wiki/Superconducting_computing
 
Status
Not open for further replies.