An A/C system for my computer case... can it be done ?

**aenigma** · 04-08-2002, 08:02 AM

Originally posted by Dan
Interesting pictures, Aenigma. 45% silver with paste flux is what I would use. Some would use Stay-brite, but I much prefer 45% silver.

Ah thanks

I dont need to solder any steel lines for my cascade chiller anymore.
I picked up an 8K a/c unit and used that compressor for the cascade chiller.

So I am getting close....
Got the beast off the table and on the floor too,so now I can get the box screwed on and everything.

**Hubris** · 04-08-2002, 04:55 PM

Aenigma - sounds like you are a lot closer to the knowledge level of the AC guys than am I. I soldered up my evaporator the other day....but was somewhat balked by the tips of the compressor...heating them I can't get them off so I don't think it's solder, but they are considerably larger than the 1/4" tubing...which actually appears to go inside the silver tip on the end of the evaporator.

I guess I'm going to have to try truck this whole rig down to an AC shop, and see if anyone is willing to touch it.

One more question....there originally was a thermostat on this system, as the compressor and condensor are from a wine cellar cooling unit. Since I need the system to run 100% duty cycle, I have removed that thermostat. Will the compressor continue to run, as long as power is applied to it - or is it somehow going to continue to expect regulation?

**aenigma** · 05-08-2002, 03:20 AM

Yep it will run fine without the thermostat.

I am not quite clear on what you are trying to unsolder?Are you using an oxy/acetylene torch?

**Hubris** · 05-08-2002, 04:18 AM

Nothing that fancy...just using a propane burner...it's plenty to melt silver solder...but I didn't want to apply heat to the tip of the compressor for too long. It does sand off very shiny silver...but it isn't silver. I suppose it could be another bonding material..

**aenigma** · 05-08-2002, 06:25 AM

Probably copper/silver.
I use silflo copper brazing rods.
They melt at about 1400f thats why your not getting thos off with a propane torch

Also if it is copper soldered to steel,which i am sure it is.
Then it is probably some other kind of solder that melts at really high temps.

**Hubris** · 05-08-2002, 06:29 PM

Obviously if it took higher heat to get it on there.....the compressor won't be damaged by the higher heat required to remove it? Other than welders, I don't know anybody with a better torch.....which is why I probably have to involve an AC shop fairly soon. I also don't have a vacuum pump to charge the system...

**Gary** · 05-08-2002, 06:53 PM

Copper to steel requires silver solder. Propane probably won't do the job. A TurboTorch using MAPP gas might do it for these small systems, or an Oxy/Acetylene set will be needed.

**aenigma** · 11-08-2002, 02:06 AM

So you having any luck with that compressor Hubris?

**Hubris** · 11-08-2002, 05:42 AM

Unfortunately I haven't had time to take it to a shop. I don't have the equipment or expertise to solder the pipes, evacuate the system, and charge with refrigerant.

The local AC shops seem to be pretty much only open during business hours....which is a little hard to accomodate, since my dayjob is keeping me more than busy at present.

Sigh...it's still exactly as it was, when last I posted. Evaporator is soldered together.....but the wet line and suction lines need to be soldered.....plus I still have no idea how much capillary tubing I'll need to use for this setup.

I need an expert...and I haven't yet found one locally who is willing to help - even for money.

**aenigma** · 11-08-2002, 08:34 AM

Well it isnt hard at all to do.As long as you have guages and a vaccum pump your set.Oh and a torch that gets hot would be nice.

Go ahead and icq me at 23619435....

**AndrOvr** · 11-08-2002, 05:01 PM

here's my cooling unid for my PC (CPU)

WEb cam sucks

**AndrOvr** · 11-08-2002, 05:03 PM

**AndrOvr** · 11-08-2002, 05:10 PM

this unid is a r409a and it gets to -50c / -51c ... if I let it runing idle for about 45mim...
It get -40c in 5mim, -45cin 10mim... and -49cin 30mim :P

BUT...
W any load at all... it wont cool!
Like... If I put the evapo. in 1/2L of 30c water... the evavpo gets 29c...at it only start to drop to 28c in 3 mim or soo...

It needs more refrigeranT?

**AndrOvr** · 11-08-2002, 05:20 PM

Could be this tube... that is too small...?
Cos it have the capilar tube in it.. and can be a botneck ?

**AndrOvr** · 11-08-2002, 05:21 PM

... what do u guys think...?

**Gary** · 11-08-2002, 08:57 PM

For a proper evaluation, the evaporator needs to be attached to the CPU under normal load. Then after it runs for awhile, we need the following information:

Type of refrigerant

Low side pressure
Suction line temperature near the evaporator
Suction line temperature near the compressor

High side pressure
Condenser air in temperature
Condenser air out temperature
Liquid line temperature near the condenser
Liquid line temperature near the cap tube

**AndrOvr** · 11-08-2002, 09:04 PM

DAMN!
Hmm... Was a A/C guy that put the R409a in it... I dont know how to mesure the presure

**Hubris** · 11-08-2002, 09:17 PM

Depending on how your friend built your system, he may or may not have installed access points where pressure can be read, using the proper type of guage. It shouldn't be too tough to get temperature readings for the various locations though.

**AndrOvr** · 11-08-2002, 09:27 PM

I can give some temps tho...

Type of refrigerant --> R409a
Low side pressure --> ??

Suction line temperature near the evaporator --> w load ( a 1/4 L of 35.c water ) it gets about 24 in 3mim The water is now 27c

Suction line temperature near the compressor --> 30 mm (3cm) from the compressor 21.6c 60mm (6cm) from the compressor.. 18.3c

High side pressure
Condenser air in temperature -->21.5c
Condenser air out temperature -->22.2c
Liquid line temperature near the condenser--> line that goes to cond. at 2 cm from it.... 33.6c
Liquid line temperature near the cap tube --> 1 cm from the deshumifider (i think is deshumifuder) 24.2c

That's it... any more to mesurE?

**Gary** · 11-08-2002, 10:22 PM

The water is not the proper load. Let's try getting the same measurements with no load. Wait until it gets as cold as possible before taking the measurements.

**abcdefg1675** · 12-08-2002, 12:57 AM

water could be a good load, if your accurate. it takes 1 caloire to heat/cool 1 liter of water 1 degree. you will need to convert BTU's to calories.

**hvac01453** · 12-08-2002, 01:50 AM

As Gary said, most computera are rated for room temperatures, why in the hell do you want to cool the computer? Cool the space, and then you eliminate the expence and possibility of condensation or moisture getting on the board. Then use an internal power fan.
Years ago My brother was doing computer rooms and they sent chilled water thru the internals of the units. The computers alone, took up about 2000 square feet of space. They were cogeneration with 4 back up ups's. All that equipment is gone now, along with 80% of the computer equipment and the cooling requirements. All PC's now....and PKG RTU's.
Things have changed..alot of money lost on computer rooms since the internet and the WWW came into the equation.

**Hubris** · 12-08-2002, 02:05 AM

The amount of resources required, plus the facilities available to us - are what prevent us from cooling our entire rooms to temps where processors can be kept at or below freezing temps.

Cooling water that is used to cool the cpu.....is fairly efficient. Having the evaporator directly contact the CPU...is fairly efficient. Even Gary's idea of mounting the guts of a computer inside a tiny commercial freezer - is fairly efficient. Cooling an entire room such that a standard heatsink, or other conventional means can keep the CPU in the intended range - is not practical. Remember, this is being done in a residence....it is not generating money - there is no value in having a room too cold to sit in, and costing $20/day in power to chill.

**aenigma** · 12-08-2002, 02:19 AM

Originally posted by hvac01453
As Gary said, most computera are rated for room temperatures, why in the hell do you want to cool the computer? Cool the space, and then you eliminate the expence and possibility of condensation or moisture getting on the board. Then use an internal power fan.
Years ago My brother was doing computer rooms and they sent chilled water thru the internals of the units. The computers alone, took up about 2000 square feet of space. They were cogeneration with 4 back up ups's. All that equipment is gone now, along with 80% of the computer equipment and the cooling requirements. All PC's now....and PKG RTU's.
Things have changed..alot of money lost on computer rooms since the internet and the WWW came into the equation.

Ever heard of overclocking?Guess not.
Computers arent rated for certain temperatures,the colder they are,the faster they run(to a certain point).

**AndrOvr** · 12-08-2002, 08:30 PM

DAMn... I dont know why... but this is taking way to long to get to -40c!

It usto to be like 5 mim to get ot -40c now it is runing for 20 mim and is at -24c

but still lowering..
WTF is rong w this unid?!

**AndrOvr** · 12-08-2002, 09:40 PM

Most of the times when I turn on the compressor runs cold... only after 30 mim it get hot.. and the line that goes to the condensator... normaly goes hot in the 1º 5mim... then.. it gets normal.. and after some time it gets REAL hot!
Is this all normal??

Is it normal... after shutting down the Compressor... it gets REAL HOT.... Like 60c or soo...

All this w out load.

**Gary** · 12-08-2002, 09:51 PM

We can't see it from here. All we can see is the description and the temperature measurements. Saying "hot" requires us to guess what you mean.

**AndrOvr** · 12-08-2002, 09:54 PM

Well.. when I say REAL HOT.. its like 55c or soo....
When I say HOT its more like 35c/50c +
Normal is around the ambient temp. NOw is 25c

**Gary** · 12-08-2002, 10:05 PM

My computer isn't doing what I think it should do. The top of the monitor is sorta warm. The screen has fingerprints on it. It doesn't make any noise when I shut it off. And my scanner isn't hooked up. WTF is wrong with my computer.

I gave you a list of temperature measurements earlier. These are the measurements that tell us what the system is doing.

**AndrOvr** · 12-08-2002, 10:36 PM

DAMN [CEnSUrED] Fu*Ing POS!
This [Censured] dont go lower than -29c now!! AGHRRRR|!
Well... the temps in the condensator are THE SAME w or W OUT LOAD! about 33c in the line that goes to the condensator and 24c that comes out...
The evapo w the load (120w TEC.. I put it when the evapo gets to -29c) gets from -29c to 56c in less than 3 mim|! and takes about 4.30mim to get from 56c to 0c w the Tec off...

The return line gets about 17c near the compressor....
and about 53c near the evapo.

<--- w LOAD

SOMEBODY HELP ME!!

**Hubris** · 13-08-2002, 12:46 AM

Are you saying that the evaporator is 53C under load, or that the return line coming out of the evaporator is 53C under load? If the temps are 53C coming out of the evaporator, and then drop to 17C before they reach the compressor....there is definately a problem.

I'm sure the experts would correct me....but if the 120W load is that drastically overheating the evaporator - but that the phase change is still dropping refrigerant temps as it moves back towards the compressor....

Either the phase change is happening too late....or there isn't enough refrigerant in the system to deal with the load of 120W.

(here's my first diagnosis....based on very little knowledge. Help me learn folks....was I even close?)

**Gary** · 13-08-2002, 05:40 AM

Sorry, but that's not a diagnosis. That's a wild guess based upon a questionable interpretation of an incoherent statement. Aside from that, it wasn't too bad for a first effort.

**abcdefg1675** · 13-08-2002, 05:42 PM

a BTU can cool 450mL (aboutone pound of water)
of water by 1 degree in one minute. My 8500 BTU airconditioner can cool 102 gallons of water by 10 degrees (celcius) in one minute.

thats one way you could check and see if your AC unit is working somewhat to its full potential.

**abcdefg1675** · 13-08-2002, 05:47 PM

Originally posted by Robearbam
Wow! Just what is it we are trying to cool down? An A/C system in a computer case? I thought the multi hole cut case with several 4" fans was a little overkill. A great site for PC cooling and over clocking is "maximumpc.com". I do believe they have a design for a heat pipe that cools the proc no matter what the freq is. Additional case cooling is supplied with fans which should be more then whats needed for today's super chips. I would personally keep any running motors far from the PC except the 12VDC fans. Were haven't reached micro wave freqs on the proc yet...I think then we will have to do something about further cooling down the PC.

ive heard and seen of a Macintosh plus that ran at 2ghz and used liquid nitrogen to cool it and the floppy drive.

and you thought multi hole cut cases were overkill? my gateway 2000 P-166mmx has some high CFM 6" brushless 120v fans getting air to circulate over the flaming hot videocard, roasting ethernet adaptor, and my toasty SB16/AWE32 soundcard. i also have a 55w peltier cooler on my PCI chipset.

**AndrOvr** · 13-08-2002, 09:32 PM

once again I ask...
Could be this tube a bit of the suction tube (5mm) that have inside the capilar tube (3mm) that is causing all this bad performance...!?

http://www.refrigeration-engineer.co...s=&postid=5515

**AndrOvr** · 13-08-2002, 10:31 PM

This is w out load... after 2h runing :P
W the evapo at -40c

Type of refrigerant --> R409a
Low side pressure --> ??

Suction line temperature near the evaporator --> -21c 1CM (10mm) from the Evaporator.
Suction line temperature near the compressor --> 20 mm (2cm) from the compressor 29.6c 60mm (6cm) from the compressor.. 19.6c

High side pressure ??
Condenser air in temperature -->25.3c
Condenser air out temperature -->25.3c
Liquid line temperature near the condenser--> line that goes to cond. at 2 cm from it.... 40.1c
Line that comes out of the condensator--> 24.2c
Liquid line temperature near the cap tube --> 1 cm from the deshumifider (i think is deshumifuder) 27.2c

This time was mesure twice

**Prof Sporlan** · 13-08-2002, 10:44 PM

a BTU can cool 450mL (aboutone pound of water) of water by 1 degree in one minute.

Your mixing energy units (Btu's) and power units (Btu/min).

A Btu will cool one pound of water one degree Fahrenheit. A Btu/min will cool one pound of water one degree Fahrenheit in one minute...

My 8500 BTU airconditioner can cool 102 gallons of water by 10 degrees (celcius) in one minute.

Mmmm... an overachieving a/c unit

102 gal * 8.33 lb/gal * 10°C * 1.8°F/°C = 15,294 Btu. If cooled in one minute: 15,294 Btu/min = 76.5 tons...

**AndrOvr** · 13-08-2002, 10:54 PM

Heres the evapo.

Compressor and Conde.

Heres the point of the suction tube is 5mm and it have the capilar in it (3mm)

**Gary** · 14-08-2002, 12:53 AM

This system doesn't have nearly enough R409A in it. It is almost empty.

**AndrOvr** · 14-08-2002, 07:14 PM

OK... Now I have to refill it

It must have some leak in the evapo

**frank** · 14-08-2002, 08:37 PM

How can you have a 16deg C differential across the discharge entering the condenser and the liquid leaving the condenser with no increase in air temperature entering and leaving?

Where is the heat energy going?

**AndrOvr** · 14-08-2002, 08:40 PM

I asked that question myself

BUT... I think that's because is a BIG radiator.. (+/-) :P and it have a BIG FAN in it!

The radiator is 29X25cm...

**Gary** · 14-08-2002, 09:35 PM

Where is the heat energy going?

There isn't enough heat energy to measure the difference in delta-T.

**aenigma** · 14-08-2002, 11:16 PM

Originally posted by AndrOvr
OK... Now I have to refill it
It must have some leak in the evapo

Oh sure you dont listen to me when I tell you its undercharged

you need to charge it with a load on the evap,like a pelt would be good.

**aenigma** · 14-08-2002, 11:21 PM

Originally posted by frank
How can you have a 16deg C differential across the discharge entering the condenser and the liquid leaving the condenser with no increase in air temperature entering and leaving?

Where is the heat energy going?

Because it is undercharged.Not enough pressure to get warm,just the discharge will be warm.

**AndrOvr** · 14-08-2002, 11:22 PM

We been chating and I already said that I needed... long before gary mention it... I was looking for confirmation...

**aenigma** · 17-08-2002, 12:25 AM

Well I got my system going finally.
Problem after problem heh

First was this:

From it being banged around from being moved up 2 stories a cap wire came loose and i didnt notice...so the compressor wasnt working and was acting as an evap

So then after getting it going I left it running overnight full load using SETI.TBird 1.4@1731 as of now
But I dont know cpu temps because the socket temp probe doesnt work too great in these conditions

So anyways the pump quit while I was sleeping so I woke up to a Halt screen,but I guess the mineral oil saved the cpu.Im not using methanol anymore.Too toxic and sucks when your trying to prime your pump and you get a mouth full of it.

so anyways the mineral oil gets really warm so I need to make a heat exchanger for it,thinking of just phase change cooling it.
Gotta insulate the res and some other things also.
got some work to do....

**Hubris** · 17-08-2002, 01:28 AM

You're crazy man....

One phase-change system to cool the evaporator of another system that actually cools the cpu. Then another phase-change system to cool the oil sitting around your motherboard....

Sigh...all I want to get going is a single system to cool my cpu in a normal case...

**aenigma** · 17-08-2002, 09:38 AM

Originally posted by Hubris
You're crazy man....

One phase-change system to cool the evaporator of another system that actually cools the cpu. Then another phase-change system to cool the oil sitting around your motherboard....

Sigh...all I want to get going is a single system to cool my cpu in a normal case...

Naw nothing is cooling the oil

I need to make a heat exchanger though....

**abcdefg1675** · 22-10-2002, 12:22 AM

Originally posted by Hubris
Generally speaking a powerful processor is going to be making something like 100W of heat, the various other components on the motherboard perhaps another 30W, the video card making perhaps 40W of heat......what exactly would it take to keep a chamber of dimensions like I listed....below freezing? While there are some who really desire extremely low temps, there are decreasing gains in performance as temps decrease. Most decent heatsinks can keep a cpu within 10-15C of ambiant temps....if it's all stored in a freezer at -10C, that should yield pretty good results.

Can I buy that in a pre-made solution...and for a reasonable price?

ive checked, and a pentium III can only make out 20-35W of heat at the VERY extreme end. its not much more for a P4.

my Gainward mobo makes only about 5W of heat, but i belive i could cook a meal for an army on my 16mb videocard.

Thread: An A/C system for my computer case... can it be done ?

Thread Tools

Display

necessity is the mother of invention

Re: necessity is the mother of invention

Similar Threads

Daikin fault codes

System Burn Out Troubleshooting

Burn out System, Clean up procedure. & Troubleshooting.

What is System balance?

Refrigeration system proposal; comments please.

Posting Permissions