Extreme Water Chiller?

[kingsack]

Extreme Water Chilling?
I have aquired 2 low temperature baths like these:
http://www.ftssystems.com/multicoolbath.htm
Mine are an older model (analog controls) but are essentially the same as the model 480 shown.
One of these units has a bad temperature controller and the other may have a bad compressor but I hope to combine parts from the 2 to make a fully operational unit. (The one unit will already operate flat out continuously).
These units were used to cool laboratory equipment with methanol, but I don't see any problem using water/glycol.
I am currently running an AXP 2500+ (199*12) and a 9700 Pro (385/339) cooled as follows:
Innovatec Tank-O-Matic-Large -> Eheim 1250 -> Swiftech MCW5000-A -> Swiftech 80mm Rad -> Swiftech MCW50-T -> BIX (External with 2 * 120mm L1A). I intend to replace the 5000-A with an MCW462-UHT.
Do any of you forsee any problems with this plan?
Any suggestions? What kinds of temps should I be able to get?

Thanks for any help you can provide.

 

[Korndog]

might want to consider condensation, you'll need to insulate those parts.
btw, nice aquirement

 

[Lazlo Panaflex]

With -80C you will be recieving very good temps. You will not be needing a rad, or a res for a waterchiller. You simpily poor the water in the unit and put the pump in there and it will act as you rad/res.

Make sure you PM Nealric with any questions you have, he is our resident waterchiiller

 

[vonkaar]

Insulate the tubing, you'll lose a lot of heat due to the amount of distance you'll have to travel considering the large footprint of the chillers.

If your coolant temperature exceeds ambient temperature pre-chiller, you could enhance your performance by incorporating a small radiator. This would be counterproductive if your water temp never reaches room temperature. My water-temp varies according to my needs... I swap between heatercores-only; heatercore+chiller; chiller-only depending on the current need.

Be wary of the power consumption. My chiller alone pulls about 600w out of the wall. Condsidering the fact that I also run 2 21" monitors and 4 other computers, current draw could be a problem.

 

[kingsack]

Update:

I was able to get one of these chillers up and running by swapping the control circuitry from one over to the other which has an operational compressor. I have tested it and it quickly cooled to -50C (The windshield washer fluid I was using began to
get very slushy so I turned it off at that point). Now I just need to build a stable stand for it, drill some holes in the wall, run the plumbing (I think I am going to go with copper tubing and quick disconects from the chiller to my case). I also need to seal the chiller resevoir better to avoid excess losses due to evaporation.
The it will be time to condensation proof and insulate all the internals.

The only decision I really need to make here is where to locate my pump. Shoul I submerge it in the chiller resevoir or should I leave it mounted ontop of my Lian Li like it is now? Would the extremely low temperature throughout be better than the temperature differential that would occur from running it external?

It seems to me that the relative height of the chiller outlet out need to be slighty greater than the pump if I don't submerge it to avoid loss of pump prime due to siphoning action. Perhaps vacuum lock would prevent this but I don't want to count on it.
Of course this needs to be kept to a minimum to avoid an unacceptable loss in flow rate.

 

[kingsack]

UPDATE

This weekend I had a problem with a water chilling project.
The system is the one in my SIG an I was replacing my dual radiators with this uber water chiller http://www.ftssystems.com/multicoolbath.htm
and I was going to eliminate my res as well since the chiller has a 1.5 gal res and I was installing my Eheim 1250 inside the chiller res.

Everything started out OK. I plumbed my loop, filled the chiller with the uber Prestone windshield washer fluid good to -40 and ran it for 24 hours with no leaks with power turned off on my rig. I didn't have the chiller going at the time since I wasn't insulated against condensation yet (possibly a fatal error!).

Next I removed my mother board and covered all the chips, slots, connectors, jumpers and the CPU socket pin holes with masking tape.
I then proceeded to spray down the board front and back with KonForm SR conformal coating (following the directions on the can - 3 light coatings on each side rotating the board 90° each time) and watited for the coating to dry. I removed all of the tape and filled the CPU socket pin holes and the void in the center with dielectric grease.

Next I reinstalled my motherboard. I covered all of the tubing with pipe insulation and sealed the ends with RTV silicone. I filled the CPU socket pin holes with dielectric grease as well as the void in the center. At this point I decided to power it up and test it before covering the socket area with foram which would make it difficult to see anything. I turned on the chiller and pump and the the PC. I wasn't getting any power to the MB and it took me a while to figure out that the ATX power jumper was off by on pin. Once this was solved I was able to power up and noticed a slight leak from the CPU block. The PC itself started momentarily and then shut down. It appeared that most of the washer fluid had landed on my video card.

I shut everything off an went to work with a hairdryer. I fixed the leak by coating the tubing with dielectric grease where it entered the cpu block and did this on all the other swiftech quick connect fitting thinking that the rubber seal might be shrinking or getting brittle with the cold temp. No more leaks after that. Unfortunately it will now power up (power led lights up on the MB and the NIC) and the fans and everything start up but I can't hear it post and ther is no video. The voice reporter doesn't make any sound. It does so hear noise from my floppy so it may be seeking.

At this point I had my brother bring over his rig which I built for him.
A7N8X-E Deluxe, 3200+ Barton, etc.. and we swapped in my video card, CPU and RAM with all operating normally. We did try his video card (a 9800XT and an old nVidia GeF 4 MX400 PCI I had laing around and still no video).

I have also tryed resetting CMOS and switching the FSB jumper to 100.
Unfortunately I gave away my system with a TBird 1.2 so I don't have an old 100 MHz FSB chip to use.

Any suggestions? Or should I just assume that my MB is toast? I don't think an RMA is an option, ethics aside, since I have coated the board and have replaced the NB cooling and have epoxied on SB cooling! This is very bad timming since I have just gone through a series of financial setbacks. If I don't have to replace my MB I would really prefer not to!
Also my wife is giving me endless grief. You know the tale .. Why did you have to go that extra mile? Your system was super fast already and worked without any problems ... blah blah blah ...

I have read the stickies on BIOS death which do bear some similarities to my symptoms but I'm not really sure that's whats going on here. Once again, any suggestion that you might have would be greatly appreciated.

 

[solo]

im guessing you will have to deal with a heatload of around 150 watts

according to you, you idle at -50C with no heatload

you can expect temps on your CPu to be around -30C? or maybe
-40C

anyways thats still great

sorry to hear about your mobo, i really don't know what is wrong with it

 

[Muku]

I know someone will say it soon.
Use antifreeze, not washer fluid. Much more effective for corrosion and stopping ice crystals from forming.

 

[kingsack]

I know someone will say it soon.
Use antifreeze, not washer fluid. Much more effective for corrosion and stopping ice crystals from forming.

Actually a 50:50 mixture of water and antifeeze will begin to gel at -35 and the viscosity will already be quite high at that point. Windshield washer fluid actually has a lower vicsosity than water and methanol has a better heat capacity than ethylene or propylene glycol.

Edit: On the other hand the increase in viscosity and surface tension might help to prevent leaks. I'm not sure it would be worth the rather severe effect on flow rate though...

 

[Muku]

thanks for the correction, also remember that methanol may eat at your seals no?

 

[squeakygeek]

Any guesses on what the temps will be are pure speculation. Unless you guys saw something I didn't, there was no indication of the cooling capacity of that unit.

Edit: currently reading the specs... guess I should have scrolled down

 

[kingsack]

OK, DFI Lanparty NFII Ultra B now on order from newegg.
Looking forward to some 250+ FSB goodness at subzero temps.

 

[ZachM]

Did you get those chillers for free? Hopefully your new mobo works out. Make sure to condensation proof everything, and double check for leaks before mounting it all up this time around. I would make sure the motherboard isn't DOA first before you start water-proofing it (that's a no brainer, but just making sure you wouldn't space it out). Post pics and your temps/OC afterwards. It sounds like it's going to be an awesome setup.

 

[solo]

Any guesses on what the temps will be are pure speculation. Unless you guys saw something I didn't, there was no indication of the cooling capacity of that unit.

Edit: currently reading the specs... guess I should have scrolled down

 

[ShaDrag]

Does methanol have the corrosion problem with mixed metals (al and cu) that water does? Not as an additive, I'm refering to running no water and 100% methanol.

EDIT: found this in an overclockers article, although its not conclusive for what I'm looking for

"It has no detectable Ph Balance which means that it's non-corrosive, but it should be mentioned that it's not dielectric either, so you don't want to spill it on your board. "

DOUBLE EDIT: found in the SAME article

"Not only is it toxic but it is also highly corrosive. Not good to have in your cooling system on a long term basis. "

 

[SniperXX]

Wow this looks cool, too bad to hear about your mobo that died. I cant wait to see what this setup will do, sounds AWESOME!

Also, im curious how much do those coolers run? They look expensive.

 

[kingsack]

Well, I have my new DFI Lan Party but haven't gotten my MCX159-A yet so I'm holding off until that arrives. In the meantime I managed to cobble together a retro rig from spare parts (Celeron 300A @ 450 MHz + GeF2 Ultra LOL). Anyway the chiller is about $7,000 new. It may be back together this weekend but I'm not sure due to my coaching responsibilities at IKWF and junior high wrestling tournaments. I also intend to perform extensive leak testing this time, probably at least 48 hours at temperature before I even think about powering up the system!

 

[kingsack]

OK, I have my new MB installed. I currently have the chiller maintaining 10°C and I'm showing 23°C for my CPU MBM5-Diode.
Unfortunatley Windows XP was totally trashed and I was unable to repair it so I had to reformat and start from scratch. I will be running it stock for now until I get everything reinstalled and stable. I will begin dropping the temperature while being on the lookout for condensation. The good news is I had no leaks after 72 hours @ -40°C although I did have some condensation at first and frost buildup on every surface that was even slightly exposed later so I guess I still need to work on insulating.

Note: I didn't use any conformal coating this time although I really don't think that was the problem. In fact I believe that it may have been a lack of conformal coating around the NB that killed my board. The NB is located between the CPU amd AGP slot on that board and I had a blue orb on it that prevented the conformal coating from reaching the area around the NB. When I removed it to transfer it to the SB on my new board I found that the AS3 under it was black! I have never seen this before so I think that my NB may have been shorted.