Toyota Prius +2004 phase-change cooling system?

Has anyone noticed that Toyota now uses an electric-powered air-conditioner in the Prius, from 2004 and up?

Now here's the idea that comes from this. I'm building a system with multiple boards inside a single Pelican 1780 case. There will be 5 boards in total. 1 quad-socket board and 4 dual-socket boards. The system is meant to be liquid-cooled. The quad-socket board already has it's loop completed from a previous build. But now I want to add the other 4 dual-socket boards to the system.

Since I want this system completely contained inside this case, I want to create a cooling loop that will accommodate all 5 boards. I recently came up with the idea to cool the liquid loop with an a/c evaporator. And since there won't be an engine running inside this case, it needs an electric-powered a/c compressor. Well, Toyota Prius to the rescue.

If I cool the liquid with the a/c, I will get much lower (and stable) temps than I would have with just straight radiators moving the heat out of the system. In this case, the warm liquid from the processors will pass over a heat-exchanger, giving it's heat to the evaporator. The condenser will be mounted to the case externally. So I have a few questions about what I might experience here.

1.) How low will my temps likely be with such a system?
2.) How large of a power supply will I likely need to run this?
3.) I was thinking about using a Silverstone ST1500. Does it have the amps?

I'll leave this here for now, and see what kind of response I get. I'll continue with it later.

Shingoshi

So wait.. Are you taking the engine out of the prius to put 4 computers in it... or what? Because thats a big case... even for 5 severs.

The Pelican case is indeed very large. But there will be many components going into it. I also want to completely insulate the system for temperature control and sound. And I want additional room for later expansion.

And it just occurred to me, that I can run the electric motor for the a/c compressor from a wall-adapter. So power is no longer a concern for the compressor. I only need now to focus on the efficiency of the cooling loop.

OmegaAvenger said:

So wait.. Are you taking the engine out of the prius to put 4 computers in it... or what? Because thats a big case... even for 5 severs.

Click to expand...

My largest motherboard (a Tyan S4989) is 16x13". And each of the "smaller" (S2912) boards are 12x13". When you take into account all of the plumbing, there's not really as much room as you would think.

Everyone is also welcome to see and participate in this thread as well:
http://www.hardforum.com/showthread.php?p=1034261610

Shingoshi

looks quite interesting.. let us know how it works.. so that when my mothers prius dies.. i can laugh at her for dropping 35k dollars on a useless car.. then rip the AC unit out of it..

Shingoshi said:

....

1.) How low will my temps likely be with such a system?
2.) How large of a power supply will I likely need to run this?
3.) I was thinking about using a Silverstone ST1500. Does it have the amps?

I'll leave this here for now, and see what kind of response I get. I'll continue with it later.

Shingoshi

Click to expand...

It should be similar to running a chilled water loop. I'd expect about 5-10C below ambient.

not sure if a single ac unit from a car will be beefy enough to remove the heat from 12 cpu's on its own. will it be cooling the video cards too?

Yeah, I was just trying to kid around. why not just get a small window ac unit or a really small ac unit like aparments use for individual apartments? then use the parts you need from those.
though the ac unit would probably need a lot more juice..

sirmonkey1985 said:

looks quite interesting.. let us know how it works.. so that when my mothers prius dies.. i can laugh at her for dropping 35k dollars on a useless car.. then rip the AC unit out of it..

Click to expand...

Only if she brought one after 2004! Too damn funny!

Shingoshi

shadow_419 said:

It should be similar to running a chilled water loop. I'd expect about 5-10C below ambient.

Click to expand...

That's what I was thinking. I actually investigated this over a year ago. Depending on the refrigerant used, I should be able to get even much lower temps. Finally now, I think this is feasible.

serpretetsky said:

not sure if a single ac unit from a car will be beefy enough to remove the heat from 12 cpu's on its own. will it be cooling the video cards too?

Click to expand...

The cpus won't be cooled by air. The a/c will be removing the heat from the coolant liquid. It's the temperature of the liquid that matters here, NOT the internal ambient temperatures.

OmegaAvenger said:

Yeah, I was just trying to kid around. why not just get a small window ac unit or a really small ac unit like apartments use for individual apartments? then use the parts you need from those.
though the ac unit would probably need a lot more juice..

Click to expand...

A/C units for housing installations are too fragile. And they ABSOLUTELY MUST be operated in a nearly vertical position. That's fine if the cluster never moves. But this system is intended to be transportable.

Only an A/C unit for a vehicle is designed to handle the kind of environments this cluster is destined for. Because if you tip an A/C unit for stationary installations, the fluid (lubricant in particular) will wind up where you don't want it. Automotive A/C units are position independent.

Shingoshi

subscribed to watch the failure later...


btw op, one signature per post.

See these links for a project synopsis:
http://www.teamnexgen.org/forum/har...quid-temperatureanyone-has-ever.html#post3353
http://www.hardforum.com/showthread.php?p=1034261610

This should answer many of the questions that may later come up.

Shingoshi

so ummmm... besides it not being an "air" conditioner anymore, my biggest question is: are you going to be building a new heat exchanger to reject heat from your cpu/gpu loop to the prius AC coolant loop? sounds hard, and im an engineer developing heat exchangers for a living...

Where the liquid-cooling loop would have typically been cooled by a radiator, it will be cooled instead by the evaporator from the A/C system. The A/C system will then dump the heat to the outside. Although, I'm also considering using a submerged condenser for the A/C system. Using the outbound liquid from the processors to cool the condenser. That liquid will then travel to the externally mounted radiator. The performance advantage would be the same as having a Geothermal Heat-Pump (http://en.wikipedia.org/wiki/Geothermal_heat_pump). The system will work less than it otherwise would. And since the inside of the cluster's case will be at very low temperature, Gel-Packs can be used to keep the temperature very stable. I am essentially creating a computer inside of a freezer!

Shingoshi

So you'll have 2 coolant loops. The primary coolant loop will be cooling your processors and then you will need to dump the heat from that into the secondary refrigerant loop. How do you plan to transfer the heat energy from your primary water loop to the refrigerator loop's evaporator? You have to maintain pressure in the refrigerant loop, so you want to submerge the evaporator in a reservoir of liquid from the primary loop? How will the evaporator hold up to being submerged continuously in terms of corrosion?

1.) A/C_Compressor -> Submerged_Condenser
Refrigerant from the compressor is cooled by the Outbound LC fluid from the processors. Liquification of the refrigerant gas occurs here. Highly efficient heat-exchange!

2.) A/C_Evaporator -> Heat-Exchanger (LC)
This essentially works like any freezer/ice-maker. The Inbound LC fluid being cooled by the expanding gases of the evaporator. The chilled Inbound LC coolant then cools the processors. Temperature stabilizing Gel-packs can be used here also.

3.) CPU -> Exhaust_Fluid(LC)
Heated Outbound LC fluid (still at relatively low temperatures) from the processor flows to the Submerged_A/C_Condenser, absorbing the heat from the condenser.

4.) Submerged_A/C_Condenser -> Radiator
The Outbound LC fluid from this heat-exchanger takes all of the heat generated within the system to the (large) externally mounted (automotive) radiator. Likely to be the only radiator in the entire system.

5.) The now cooled Inbound LC fluid completes the loop, moving back to the A/C Evaporator (for further cooling). Cycle is now closed.

Shingoshi

sirmonkey1985 said:

looks quite interesting.. let us know how it works.. so that when my mothers prius dies.. i can laugh at her for dropping 35k dollars on a useless car.. then rip the AC unit out of it..

Click to expand...

I don't even know what this post is about? you think prius are bad cars? umm no.

my mom's has 184k miles on it, and it runs just like new.

don't be a hater.

There are two separate heat-exchangers in this system. Each at opposing ends of the A/C cycle.
1.) Draws heat away from liquid going to the processors. Evaporator is used here.
2.) Draws heat away from the condenser gases (for liquification). Submerged Condenser is used here.

Shingoshi

I smell what your steppin in. All the refrigeration system does is move heat from the cooled side of the radiator to the hot side. Your not actually transferring all of the heat through the refrigerant loop.

The Delta-T that your refrigeration cycle will have to overcome may be pretty steep. Your taking heated fluid from the CPUs and trying to dump even more heat into it. You may not gain anything from this part of the deal. Obviously you want your evaporator cooling the fluid right before it goes to the CPUs, but you may actually improve your fluid temps going into the procs by cooling your condenser in ambient air. It all depends on how much your fluid temps rise on their way through the CPUs. If your CPU output fluid is well above ambient then you lose ground by cooling the condenser with it.

The A/C Evaporator will lower the temps of the Liquid-Cooling well below 0C. In fact, the temperatures (of the ethanol/antifreeze mixture) could rest around -40C, easily. The typical ambient temps (outside the computer case) will be nowhere near that low. So the raised temperature of the fluid leaving the processors will still be very cold. That should be more than enough to cool the gases in the condenser, since it (the condenser) typically works just fine even during summer.

I posted this link for a reason. Please read it:
http://en.wikipedia.org/wiki/Geothermal_heat_pump

Shingoshi

With a fully-configured system, having 12 processors, the following numbers are given:
Tyan S4989 = 4x75Watt processors = 300Watts
Tyan S2912 = 8x55Watt processors = 440Watts

Total = 740Watts

Even if I did use 75Watt processors throughout the entire system, the total wattage is still only 900. And if I chose to run nothing but 55Watt processors, the total load would still be only 660Watts. The A/C system can easily reduce these temperatures to be very manageable.

Shingoshi

Your looking at an insulated system which will be entirely sealed if I'm not mistaken. Any power you put in must be dissipated. The CPUs alone are not the entire heat load. That's beside the point however.

A quick google search shows that the Prius system is capable of 3.4 kW of cooling at optimum. Any refrigeration system loses effectiveness as you increase the temperature difference between it's evaporator and it's condenser. I'm not saying your idea is a bad one. I'm intrigued by it. I'm just recommending that you make your design versatile enough that you can put your condenser in ambient air or in hot-side fluid depending on which way nets the best results.

Any way you slice it you have to get your radiator and condenser above ambient temps in order to dissipate any heat.

flogge said:

The Delta-T that your refrigeration cycle will have to overcome may be pretty steep. Your taking heated fluid from the CPUs and trying to dump even more heat into it. You may not gain anything from this part of the deal. Obviously you want your evaporator cooling the fluid right before it goes to the CPUs, but you may actually improve your fluid temps going into the procs by cooling your condenser in ambient air. It all depends on how much your fluid temps rise on their way through the CPUs. If your CPU output fluid is well above ambient then you lose ground by cooling the condenser with it.

Click to expand...

This should have already been answered. So I don't know what else to say to you here.

flogge said:

Your looking at an insulated system which will be entirely sealed if I'm not mistaken. Any power you put in must be dissipated. The CPUs alone are not the entire heat load. That's beside the point however.

A quick google search shows that the Prius system is capable of 3.4 kW of cooling at optimum.

That would be almost 5 times higher than whats needed here.

Click to expand...

Any refrigeration system loses effectiveness as you increase the temperature difference between it's evaporator and it's condenser.

Please show your evidence for this analysis.

Click to expand...

Any way you slice it you have to get your radiator and condenser above ambient temps in order to dissipate any heat.

Click to expand...

I do have the (cost-prohibitive) option of running a second-stage compressor. That would definitely get the temperatures high enough. That along with all the other components to be cooled (by liquid) will remove all the heat to the outside.

Shingoshi

so... like so?

That's really a pretty good representation of what the system will look like. I was going to use a circle with a slash through it to give a not so elegant example. But this works well. The circle would have represented the liquid-cooling loop. And the slash would have represented the A/C loop. I'm glad someone actually understood what I was trying to say.

Shingoshi

Edit: I hope you don't mind. But this was so good, I had to save it.

serpretetsky said:

not sure if a single ac unit from a car will be beefy enough to remove the heat from 12 cpu's on its own. will it be cooling the video cards too?

Click to expand...

Do you still think that would be a problem with 3.4kW of cooling available to the system? I'm thinking it's more than enough to do this.

Shingoshi

Remember, this compressor is likely larger than those used in most phase-change systems. I think running a two-stage compression would be just insane! Feasible, but insane!

Edit: If this compressor turns out to be too small, I'm just wondering if another larger automotive A/C compressor could be modified to run on an electric motor as well?

Shingoshi

I was just thinking that if I really wanted to get the best performance from this system, I would have to run separate coolant lines to each and every cpu. That's really not so big of a deal. This came up while considering a two-stage system and how to configure it.

The biggest problem with a multi-stage system though, is that you wind up with temperatures that are so low, it would be difficult to find liquids that won't freeze. And that's even with using azeotropic (http://en.wikipedia.org/wiki/Azeotrope) mixtures. So what you're left with are compressed and liquefied gases.

That's when you start getting into real difficulty. If you use one refrigerant to super-cool (and liquefy) another, you're pretty much forced to use that refrigerant as the final coolant. But since that refrigerant can only be evaporated once, you would HAVE to run separate lines to EACH cpu, and use your cooling block as the evaporator. If there were no problems with the metallurgy of the block, you could do it fairly easily.

You would have to use some sort of nozzle like an expansion valve on each cpu block as well. Because the pressure in the line from the condenser to the evaporator must NOT drop. And that really drives the cost up. This whole project is already almost unrealistic for me. But then I'm crazy, and need something to drive my imagination. Maybe I could find G1/4 nozzles with pinholes to do the trick. But the hoses would become an issue then as well.

But then, it comes down to candidates for refrigerants. Propane freezes at @-187C. Neon freezes a bit lower than that. But I don't know if you can get Neon to freeze in only two stages. I don't think it would be a problem for Propane. I just don't like the idea of working with such a dangerously explosive gas.

Shingoshi

I'm pretty sure that a refrigerant in it's liquid state, can be compressed to extremely high pressures. Think of power washers. You basically wind up working with hydraulic pressures. You could then evaporate your first-stage refrigerant directly in a chamber/reservoir of another refrigerant.

I'm thinking back to my readings on dry-ice slurries in ethanol mixtures. I have to remember how this works. I think that if you maintain high pressure, the liquid won't freeze. But I'm not sure about that. I have to go back and read through my data again. It falls under the heading of super-chilled fluids. Fluids that are chilled below their freezing point.

I just keep wanting to reach that magical point of -120C. Where cmos processors double their performance and become nearly super-conductive.

Oh, I should stop dreaming, and go to bed!
Shingoshi

The compressor must work in the gas phase. You have to evaporate the refrigerant at low pressure in the evaporator in order to lower it's temp and allow it to absorb heat. Compressing in the liquid phase (or pumping) can't be done. If you really want to go for this I would recommend buying a Thermodynamics book for a college course. That way you can get an idea exactly how a basic refrigeration system must work and why forcing a change from liquid-gas and vice versa actually causes a temperature change. Just the tables at the back of the book are going to be crucial to your project.

http://www.achrnews.com/Articles/Technical/BNP_GUID_9-5-2006_A_10000000000000106551

Shingoshi said:

I'm thinking back to my readings on dry-ice slurries in ethanol mixtures. I have to remember how this works. I think that if you maintain high pressure, the liquid won't freeze. But I'm not sure about that. I have to go back and read through my data again. It falls under the heading of super-chilled fluids. Fluids that are chilled below their freezing point.
Shingoshi

Click to expand...

uhh.. as far as i can tell... ethanol has freezing temperature far below carbon dioxide's sublimation temperature. So there's nothing special needed, just drop it in and the alcohol will easly go down to -110F or so.

flogge said:

The compressor must work in the gas phase. You have to evaporate the refrigerant at low pressure in the evaporator in order to lower it's temp and allow it to absorb heat. Compressing in the liquid phase (or pumping) can't be done. If you really want to go for this I would recommend buying a Thermodynamics book for a college course. That way you can get an idea exactly how a basic refrigeration system must work and why forcing a change from liquid-gas and vice versa actually causes a temperature change. Just the tables at the back of the book are going to be crucial to your project.

http://www.achrnews.com/Articles/Technical/BNP_GUID_9-5-2006_A_10000000000000106551

Click to expand...

At least I don't think you got this. I was talking about using a separate hydraulic pump after the gas had returned to a liquid coming from the condenser. Compressing that fluid to high pressure should be no different than using oil or water for hydraulic power systems. Spraying that fluid through a nozzle and allowing it to expand will return it to a gaseous state again, and super-chilling anything in the vessel containing it.

Shingoshi

serpretetsky said:

uhh.. as far as i can tell... ethanol has freezing temperature far below carbon dioxide's sublimation temperature. So there's nothing special needed, just drop it in and the alcohol will easily go down to -110F or so.

Click to expand...

I read about this where the CO2 would be turned into microscopic snow flakes (referred to as a slurry) suspended in the liquid. Being in a super-chilled state increases the amount of heat the CO2 will absorb upon expansion. But then, we're back to needing separate lines to each cpu.

But I was mostly thinking out loud.

Shingoshi

http://www.filtersfast.com/Pentek-150235-Filter-Housing.asp
This would have to be the monster of all reservoirs. Think how much money most of you have spent for some small reservoir (by comparison). This thing will sit inside my case and be chilled by the lower temperatures inside it. And being that I would be using an automotive radiator as the only (external) cooling device for the LC loop, having something this large would be very nice. This would allow the liquid to be precooled before being deep-chilled by the evaporator.

And this brings up another point. I think we've all overlooked the fact that the entire system will be cooled by the LC loop, whether the A/C system is installed/running or not. At no point is the LC loop interrupted. It functions no differently than any other LC loop. The only difference here is that it's temperatures are forced lower by being exposed to the cold expanding gases from the evaporator. But like I said, even if the A/C weren't running, the LC loop will be sufficient for "normal" operation.

I've become rather familiar with using NPT fittings, thanks to Lowes.com. I've purchased virtually all of my plumbing for my LC loop from Lowes. Having 1 1/2" NPT is not a problem for me, now that I know how to use them. In fact, I'm using two B&M Racing 11x11" SuperCoolers (#70274) for my radiators in my existing system. And they use 3/4" NPT fittings, which work easily with 3/4" to 1/2" hose barb adapters.

Shingoshi

At least I don't think you got this. I was talking about using a separate hydraulic pump after the gas had returned to a liquid coming from the condenser.

Click to expand...

How would this serve any purpose? When the expansion valve decreases the pressure, it reduces the boiling point under the temperature of the evaporator environment, therefore forcing the liquid to evaporate at a low temperature and absorb heat. The compressor does the same thing by increasing the pressure and therefore the boiling point so that the gas must condense into liquid phase at a high temperature and expel heat. It doesn't matter how high the pressure is before the expansion valve as long as it is entirely in the liquid phase.

where in the hell are you going to dump all that heat and noise? also are you ready for a huge gain in power usage?

d50man said:

where in the hell are you going to dump all that heat and noise? also are you ready for a huge gain in power usage?

Click to expand...

As far as dumping goes (try it):
Heat = Automotive Radiator (externally mounted on case)
Noise = Fully insulated case. That includes the acoustics.
Power Usage = I'm looking to make a special power cord to run from my kitchen refrigerator power socket. This is to be used with the Silverstone ST1500 power supply.

End of story.

Shingoshi

This (use of hydraulic pressure) isn't something that's required. It was just something I was thinking about in conjunction with the overall cooling scheme. It follows that any compressed gas absorbs heat upon expansion. The higher the rate of expansion (drop in pressure), the more heat is absorbed in the process. Hydraulic pressures are typically very high.

The only thing I'm not sure about here, is if the same thing is true for the expansion of compressed fluids. I was thinking about how swamp coolers work. The same or similar thing works with Ultrasonic humidifiers.

An ultrasonic humidifier is usually silent, and also produces a cool fog.

Click to expand...

The cavitation (http://en.wikipedia.org/wiki/Cavitation) of a fluid excited by ultrasonics will cause the temperature of that fluid to drop. The point is that any time you create an aerosol, the vapor from it will absorb heat. An evaporating fluid will absorb heat, and doesn't need to be a compressed gas.

As with high pressure washers, vapor is created. Whether that vaporization would enhance cooling, I'm looking into that.

References:
http://www.wipo.int/pctdb/en/wo.jsp?wo=1993010404
http://www.wipo.int/pctdb/en/wo.jsp?IA=AU1992000619&wo=1993010404&DISPLAY=DESC

That's all for now. I'm still researching.

Shingoshi

This is a very interesting cooling solution. I've often thought about ac coolers that chill ambient liquid coolant after a radiator (my ideas always involve weak ac coolers and peltiers that wouldn't be able to handle the heat themselves...). However, i have never considered using the same coolant to cool off the hot side of the ac.

However, you keep saying the liquid cooling loop will be overall colder with AC than without, I just wanted to clarify that i do not believe the radiator side will be colder, in fact, im pretty sure it would be hotter, as it should be. This is effectively taking heat right before the cpu and dumping it right past the cpu, allowing the cpu to run at lower temps: a little bypass. However, it will also be dumping some extra waste heat as well (ineifficient AC). The radiator will have to run hotter in order to make up for that extra waste heat. All in all, for a closed system, and not having to deal with 12 different blocks designed for phase change operation, this looks pretty cool.

I'm also concerned, however, that the internal air temperature of this enclosed and insulated case will be a little high if the coolant loop is not designed with proper radiators/coolings fins inside the case. maybe that's not a problem though

edit: also, im curious do you know if the 3kW rating on the ac is total electrical power usage? they don't measure the actual heat transfer right?

Let me clarify, hopefully without further confusion...

serpretetsky said:

However, you keep saying the liquid cooling loop will be overall colder with AC than without,

See note below

Click to expand...

I just wanted to clarify that i do not believe the radiator side will be colder, in fact, im pretty sure it would be hotter, as it should be.

See note below

Click to expand...

This is effectively taking heat right before the cpu and dumping it right past the cpu, allowing the cpu to run at lower temps: a little bypass. However, it will also be dumping some extra waste heat as well (ineifficient AC). The radiator will have to run hotter in order to make up for that extra waste heat.

See note below

Click to expand...

All in all, for a closed system, and not having to deal with 12 different blocks designed for phase change operation, this looks pretty cool.

I'm also concerned, however, that the internal air temperature of this enclosed and insulated case will be a little high if the coolant loop is not designed with proper radiators/coolings fins inside the case.

See note below

Click to expand...

maybe that's not a problem though

edit: also, im curious do you know if the 3kW rating on the ac is total electrical power usage? they don't measure the actual heat transfer right?

Click to expand...

Note 1) The system will have a lower temperature WITH the A/C system running. However, an A/C system is NOT required for the operation of this project. The A/C system could completely fail, without taking down the rest of the system with it.

Note 2) Once more, you have understood what other's have not!

Note 3) It is technically correct to say there are no cold states of matter. The correct explanation is that matter is cold because of an absence of heat. In order for any matter to have it's heat content lowered, something else in it's environment must absorb heat away from it. So let me continue with a proper explanation of the terms here.

The fluid coming from the radiator will be moderately cooled (near the external ambient). It will then enter the very large reservoirs I'm choosing to use. These reservoirs will act like storage tanks, kept inside of the cluster case. It should also be noted, that the larger the reservoir, the less the temperatures will change within it from the introduction of additional fluid. Because the inside temperature of the case will be substantially lower than the external ambient temperature, the fluid inside the storage tanks (reservoirs) will be additionally cooled. The fluid will then pass on to the evaporator, where it will be aggressively cooled, to very low sub-zero temperatures.

Upon passing over the cpus, the temperature of the fluid will be moderately raised. However, the temperature will still be substantially lower than the external ambient temperature. This is where the submerged condenser will benefit. Having temperatures much lower than the external ambient, will draw heat away from the compressed gas more quickly and efficiently than would occur outside the case.

Essentially, this system will be like having an air-conditioner running in the middle of the coldest winter on earth. This is why I posted the link for Geothermal Heat Exchangers. Look above and use the link. It's important for understanding this.

Note 4) Just today I was looking at this unit from Koolance:
https://www.koolance.com/water-cooling/product_info.php?product_id=563
I had already considered using some form of internal heat-exchangers inside the case. I can still do that if I find that it would be better. But please read my initial thread for this project, as this thread is only a sub-topic of the main thread:
http://www.hardforum.com/showpost.php?p=1034246412&postcount=1

I'm also concerned, however, that the internal air temperature of this enclosed and insulated case will be a little high if the coolant loop is not designed with proper radiators/coolings fins inside the case.

Click to expand...

I've already noticed that my other computer displayed lower temperatures internally, just from the liquid-cooling blocks chilling the air inside the case. And the temps of those blocks were nowhere near as cold as this system will be. The inside of this case will get very very cold. And those low temperatures will be held down by the internal storage tanks. They will resist heating, and therefore keep the internal temperatures very low. Another benefit of having such large reservoirs, is that I can stuff them with Gel-Packs, and still have plenty of fluid capacity.

Shingoshi