phase change parta list

[resago]

ok what do I need?

compressor, dryer(maybe), condenser, evaporator.
this is for an r12 system using propane.

I'm friends with an HVAC cert, so assembly and filling won't be a problem.

that compreesor listed in the cooling tech article, did that come with any additional components, or was that just the compressor?

I was looking at this compressor: here

would this work?
could someone who knows What I need post a partslist?

[TigerOats]

The whole trick to this phase change thing is get the right ratio between the condenser and evaporator - if you carefully examine some freezer units you will notice that the condenser is roughly a quarter the size of the evaporator. This poses problems with regard to getting the right kind of mass of refrigerant flowing with the very small size of an on-die evaporator.
The balance of the compressor/condenser/evaporator is critical to getting the right condenser pressure and therefore heat dissipation and then metering the liquid refrigerant to the evaporator to get the right kind of low pressures required together with the right kind of quantity to enable it to get sub 0 temps.
Let no one kid you that this is an easy project.

[resago]

oh, forgot to mention, I will be using it to chill water, not directly on the chip.

so evap size is flexible.

Was wondering if a small (6"x6") heatercore could be used as an evaporator. I would then seal it in a contaner with the water.

[resago]

ok, some parts:

Compressor: Granger part 6E033 $120.85
Condenser: Granger part 5E031 $265.50
Dryer: Granger part 6X580 $ 9.24
TE Valve: Granger Part 3A664 $ 30.65

Evaporator: hopefully a small heatercore will work. $20


Total: 446.24

Now, why is the condensor so damn much? where can I get one for a REASONABLE price. all it is is a radiator sealed for high pressure!!

If I could keep this project under $300 I'd do it.

[TigerOats]

To be very honest utilising an existing system is firstly a lot cheaper and secondly a lot easier. My own system cost $20 and I am getting quite good results as you can see from my sig.
If you wish to persue the do-it-yourself route I suggest a visit to this where you will find a number are doing the same thing. Ideally try to get the details of condenser size for the refrigerant to be used from the compressor manufacturer. Then plan to design an evaporator with a volume 4 to 5 times bigger. You will find some interesting facts about metering the liquid refrigerant here as well.

[resago]

what are you using?

[TigerOats]

I am using an under-counter fridge with the old style freezer compartment in the top. You can find details of how I converted it here .
Since I wrote the article I have found a number of factors that increase the cooling performance. Firstly insulation of the res is critical. Limting the heating of the coolant in the res is a major factor and secondly cooling the compressor and the condenser can have very beneficial results.

[Brad]

trust me, it'd be much easier to use a current setup, like buy a crappy old freezer with a 1/5HP or better compressor. The heatercore idea is great, have the water outlet right next to the heatcore, to keep the coolest water going to the cpu.

[superart]

but what if I want to make the evaporator directly on the die, how would I go about mounting it? I looked arround and i couldnt find this little bit of info in any article or guide.

[TigerOats]

There is not easy answer and I for one wish there were. The major problem is balancing the mass of refrigerant required to cool 150W+++ and get an evaporator that will fit on the die. I think I have discovered a way after months of research but will hold back until I can verfy that it works in practice. The most common problem is that those that have tried it either end up with good cold temps on trial but poor results because they don't have the mass required or the evaporator is not producing low temps but is cooling the proc.
Here are a couple of links to sites where they have done it 1st and 2nd .

[Brad]

you can run the evap on the cpu, but it's performance will be less, and it is much harder to configure

[resago]

OK, I just finished reading that entire thread, and all mentioned links. My head hurts.

My Hvac Cert guy can probly hook me up for free. the question really is, is it worth it?

I have to insulate the ENTIRE waterblock,hoses, etc -Pain in the arse.

Plus I need toxic chems in the water and a pump that can handle extreme temps.

I could possibly do direct die cooling with a std waterblock as the evap, and a smallish condenser, but I'd still have the insulation hassles.

also, I like to keep everything in the case, I don't like hoses and containers laying about all over my computer room. The compressor and all other pieces must FIT.

all this leaves me with little options.

but for those who want to know.

1.) Get a TEV - Thermal expansion valve, and don't worry about capilary tubes.
2.) Your evaporator should be roughy 1.5 to 2 times the volume of your condensor. - Copper-tubed oil coolers can be used as cheap condensors.
3.) Use propane instead of R-12, its cheap and works better.
4.) Make friends with an HVAC Cert.

[resago]

also, I don't think I saw this in that huge thread, but what would be a better candidate, a fridge/freezer or a window unit A/C?

[TigerOats]

I don't go along with that evap size. I have been working on this for some months now and my examination of many fridges with freezers shows the evaporator being 4-5 times the size of the condenser.
A/C units could be used but then you must go along with the freezer concept of a much bigger evaporator to get sub-0 temps. A/C units generally only cool to the 5-10C range.
The use of propane is a good choice. You will find an excellent link to converting to propane in the Project-X thread.
The use of a small condenser and a waterblock as a evaporator will probably lead you into the minefield of good temps but insufficient refrigerant mass to effect cooling. If you think about it this way - if because of space you decided that you would use a micro pump instead of a normal pump. The water would be flowing but there would not be enough flow to cool the system. The same applies in this instance - you'll get the right temps and pressures but because the refrigerant is boiling off too quickly you won't get the result you want.

[resago]

My question with a fridge compressor is this.

As far as I can see, its the same thing as putting your radiator in your freezer, which most people seem to think won't work, cause the system isn't made for constant 24/7 use.

that is why I asked about the window A/C unit, which IS designed to work 24/7.

Any thoughts on this?

[TigerOats]

If we examine the concept of using a rad in a freezer then we have the following heat pathway - coolant to rad wall to air to evaporator wall to refrigerant. The bad part is the air. Again we come back to the fact that air is a bad conductor of heat. In any cooling design the ideal scenario is to get the heat to the transfer medium as fast as possible with the minimum of thermal interfaces. If you are using a refrigerant then then the ideal is to get the heat from the source to the refrigerant as quickly as possible.
I checked a website on home appliances and the fridge is the household item with the longest lifespan with an average lifespan of 14 years. On average with normal use a compressor works about 8 hours a day. In my current system the compressor works about 12 hours a day. This would equate to knocking off roughly a quarter of its expected lifespan. Compressors in a domestic situation are expected and designed to work in some fairly harsh conditions. In this context there is little or no danger of the motor burning out when used within the specified tolerances. The additional factor is that compressors would like most systems of their type suffer their greatest wear on startup because of a lack of lubrication on the wall of the chamber. So in this context they are more likely to suffer greater wear in the normal environment than in this application because their cycle time is shorter.
A/C units are generally bigger because they have to move a greater mass of refrigerant than a fridge. They would not be more or less prone to failure in my opinion.

[superart]

TigerOats, your second link was pretty helpfull, but it was for dual CPUs, which gives you more surface area. But if I were to use single CPU, and a big Heat spreader, would that work too? That way i can mill out a "water block" and use it with refrigerant, sort of similar to what the guy in the second link did.

[TigerOats]

You would need a pretty big heat spreader anyway to avoid thermal shock. I would advise using at least 5mm thick plate for the spreader. In terms of evaporator design I would look at a chamber design that is about 50mm^3 at least. Aim at building in vertical spill over chambers inside to trap liquid refrigerant but allow the escape of vapour. If you plan to use a TEV then it will help tremendously. Here is a source of excellent info on TEV's and evaporator design and performance. You can find an excellent piece of software for design of refigeration and simulate the effect here

[|PuNiSh3R|]

I am getting into all of this phase changing crap.. Unfortunatly I cannot keep it simple and just cool my CPU and be done with it.. NNNNOOOO!!! I need to do something grand and great!

Anyway.. I'm confronted with electrical use. I am doing this entire setup to cut down on the use of electricity from my 172w peltier running off of a 24v @ 12.5amp PSU... So that's like 300w an hour .. Plus my PC.. and monitor.. hell.. I am using WAY to much electricity .. plus I want to put an A/C unit in this room and that just won't happen if I am still using more electricity then my microwave!

Anyway.. I am just curious.. I am looking at setting this stuff up but I am on a limited budget.. so um.. cheap is the essence..

I am looking at getting this fridge from Sears.

Sears

Sears Item #04691171000
Mfr. Model #91171

My so called "plan" would be to either yank everything out of this sucker .. this be hard to do?

Or just heavily insulate the freezer compartment and stick my resevior in there... Hell.. I'd probably just water proof the sucker and put cut the top of the frige open above the freezer and put a thick piece of lexan there .. so I can see down into the freezer.

Anyway.. I would just water proof the freezer compartment and put my pump and all that crap right in and pour water directly in.. then use the rest of the fride space to store like.. food or something lol.

Any idea on my water temps if I did this .. with a Tbird 1.4GHz.. OC'd to 1.6GHz at 1.75V... or 1.80V..

I am building a new case out of wood and acrylic... So what I would do is if the water was cold enough I would run it through my danger den super cube radiator and use a low CFM fan to blow through it. The radiator would be inside my case and the fan would blow off cold air. I would circulate this air around my case then to bring the entire case down to 60F or 70F.. maybe even lower? I am lining the entire case with neoprene rubber.. thin layer of it. I would like to keep everything compact and in the case or at most right outside of it.. I really do not want a whole frige sitting outside of my case!

Thanks.

[|PuNiSh3R|]

Oh, BTW...

I would probably search through peoples garbage or go to a junk yard or call local refrigeration places to find info on this first.. before I go sinking 80 dollars into a frige alone.. plus another 50 to 60 in a case.. that just would NOT work for me.

I'd really go stellar crazy if I found a place that had a nice system sittin round that would be ready to roll for cheap and work great... but that's everyones dream .

I really DO NOT want to use more then 200W an hour on my new cooling solution... I'd like to target 150W though.. that would be awesome.

Edit: I am going to be running under 1 gallon of water most likely.

[beav]

have any details on how large or what kinda design an evaporator should be if it were to be placed directly on the die?

[Cieprus]

Just how much will you be able to OC your CPU???

Isnt there a limit regardless of the temp

[TigerOats]

I completed this project in January and you can read about it here . I have just finished a burn-in on a T'bird 1.33 AXIA Y and currently running @ 1632 at 1.94vcore at 7C. Coolant temps average -12C.

[|PuNiSh3R|]

Tiger, what is your electrical use with that frige?

[TigerOats]

I am not sure what the electrical rating is but it is a 110W power rating. The compressor runs about 14 hours a day and I estimate the cost to be about £4 ($6) a month. Its difficult to say really because I can only compare electricity consumption with last year and we have seperate fridge and freezer now compared with a combo last year.