PDA

View Full Version : What WC components to use?



silverphoenix
04-13-2007, 06:20 AM
I have a fairly good Idea on how to build myself an A/C unit chiller, but the question is what water cooling equipment to use to withstand the temps.

Such as waterblocks, cpu,gpu,nb etc, I hear the swiftech MCW6000 is good, but they are hard to find in stock.
Also pumps and tubing

I don't mind doing some minor modifications to already existing parts, but any help would be awesome.

Xeon th MG Pony
04-13-2007, 09:22 AM
http://www.ntcw.com/Merchant2/merchant.mv?Screen=CTGY&Store_Code=NTCW&Category_Code=STH2OWB < This is a store here in BC Canada.

Console
04-13-2007, 09:24 AM
You should look for block that are all one peice (no o-rings) so that you wont get a leak once the temp changes and the block expands a bit.

silverphoenix
04-13-2007, 10:06 AM
You should look for block that are all one peice (no o-rings) so that you wont get a leak once the temp changes and the block expands a bit.

Is the swiftech the only single piece? and what about all the other parts? I'm assuming your average tubing would also crack at super low temps.

Modzilla
04-13-2007, 10:21 AM
You should take apart the waterblock, discard o-rings, and seal with some type of industrial adhesive that is rated for low temps. I use Devcon 250. Replace plastic barbs with metal barbs, remove o-rings, and coat threads with the same adhesive. Use Tygon tubing and worm-drive clamps. The older swiftech cpu and gpu blocks work well (w/ the blue aluminum top). I haven't chilled a delrin block yet, don't know how those would hold up.

Xeon th MG Pony
04-13-2007, 10:34 AM
from what I hear crackage is what happens with the plastic ones.

silverphoenix
04-13-2007, 11:54 AM
You should take apart the waterblock, discard o-rings, and seal with some type of industrial adhesive that is rated for low temps. I use Devcon 250. Replace plastic barbs with metal barbs, remove o-rings, and coat threads with the same adhesive. Use Tygon tubing and worm-drive clamps. The older swiftech cpu and gpu blocks work well (w/ the blue aluminum top). I haven't chilled a delrin block yet, don't know how those would hold up.

Delrin can handle a minimum of -50C. and I don't really want to start another thread, but is it possible to use a TEC under an A/C chiller?

Modzilla
04-13-2007, 12:11 PM
The material data sheets I have been able to find show a minimum delrin temp of between -40 to -50 C. I will be chilling a delrin block in a week or two, so I guess I'm going to find out.

You could use a TEC with a chiller, but why would you want to? With a well-constructed a/c chiller you can get coolant temps of -20 to -30 C. It would take alot of electrical power and heat dissipation to get those temps with a tec.

silverphoenix
04-13-2007, 12:57 PM
The material data sheets I have been able to find show a minimum delrin temp of between -40 to -50 C. I will be chilling a delrin block in a week or two, so I guess I'm going to find out.

You could use a TEC with a chiller, but why would you want to? With a well-constructed a/c chiller you can get coolant temps of -20 to -30 C. It would take alot of electrical power and heat dissipation to get those temps with a tec.

Well I meant putting a TEC under the chiller block. So the chiller would be cooling the TEC which would be cooling the cpu directly. I'm thinking a 226W TEC cooling a cpu, with a chiller cooling the TEC, and a 5500BTU AC unit should be strong enough to deal with that heat load.

serialk11r
04-13-2007, 04:54 PM
Well I meant putting a TEC under the chiller block. So the chiller would be cooling the TEC which would be cooling the cpu directly. I'm thinking a 226W TEC cooling a cpu, with a chiller cooling the TEC, and a 5500BTU AC unit should be strong enough to deal with that heat load.

That will get poor efficiency because the TEC is running colder. I don't know about how low temps the chiller (really depends I think) will get but...you can only get the hot side of the TEC so cold like that...

silverphoenix
04-13-2007, 05:07 PM
That will get poor efficiency because the TEC is running colder. I don't know about how low temps the chiller (really depends I think) will get but...you can only get the hot side of the TEC so cold like that...

huh I think with all this efficiency isn't really the goal. The TEC should help bring the cold side temps even lower than what the chiller could no?

serialk11r
04-13-2007, 05:22 PM
huh I think with all this efficiency isn't really the goal. The TEC should help bring the cold side temps even lower than what the chiller could no?

What I'm thinking is, with a heat load like that of a 226W tec (550W of heat...but it probably varies) temps won't exactly get too low. If you're gonna use a compressor of that size, probably phasing the CPU would be a better idea... But I'm not sure on all this.

Xeon th MG Pony
04-13-2007, 05:54 PM
huh I think with all this efficiency isn't really the goal. The TEC should help bring the cold side temps even lower than what the chiller could no?


No, you will get worse temps with the tech and totally slaughter the systems over all efficiency.

A good chiller can easily bring you to -40 range.

silverphoenix
04-13-2007, 06:50 PM
No, you will get worse temps with the tech and totally slaughter the systems over all efficiency.

A good chiller can easily bring you to -40 range.

why? I'm trying to understand why putting a TEC under the chiller would be so bad? the chiller should be able to cool the TEC well enough, and it should produce an even lower temperature cpu side.

anyways back onto the title what would be a good pump?

serialk11r
04-13-2007, 07:09 PM
why? I'm trying to understand why putting a TEC under the chiller would be so bad? the chiller should be able to cool the TEC well enough, and it should produce an even lower temperature cpu side.

anyways back onto the title what would be a good pump?

I'm pretty sure it works like this, more heat means you can't have as low evap temps (I think), and the TEC is going to put out a lot of heat. And with a compressor like the size you are talking about, straight chilled water (with a good system) can get very good temps, and will be a lot more efficient. The huge heat output of the TEC probably ruins the low temperature potential of the chiller, and the TEC itself cannot pump heat very effectively. Now I'm not totally sure if that's the reason...so guys correct me if I said something wrong.

silverphoenix
04-13-2007, 08:23 PM
I'm pretty sure it works like this, more heat means you can't have as low evap temps (I think), and the TEC is going to put out a lot of heat. And with a compressor like the size you are talking about, straight chilled water (with a good system) can get very good temps, and will be a lot more efficient. The huge heat output of the TEC probably ruins the low temperature potential of the chiller, and the TEC itself cannot pump heat very effectively. Now I'm not totally sure if that's the reason...so guys correct me if I said something wrong.

I guess that can be it, I've heard of some people putting TEC's under LN2 though, mebe a liquid chiller just isn't powerful enough.

serialk11r
04-13-2007, 08:36 PM
I guess that can be it, I've heard of some people putting TEC's under LN2 though, mebe a liquid chiller just isn't powerful enough.

Holy F*** wow that would use a LOT of electricity!

Modzilla
04-13-2007, 09:04 PM
Well I meant putting a TEC under the chiller block. So the chiller would be cooling the TEC which would be cooling the cpu directly.

I know what you meant. I think you can get significantly lower temps with a chiller than with a tec, regardless of how well you cool the hot side of the tec.

Xeon th MG Pony
04-13-2007, 09:07 PM
I guess that can be it, I've heard of some people putting TEC's under LN2 though, mebe a liquid chiller just isn't powerful enough.

and I hear of some people who test dynamite befor using it, what can I say?

Thier power usage goes up and cooling ability goes to zero, IE they are just like a car spinning its weels, a whole lot of nothing at a whole lot of power usage. with out going into complex physics and boring you thats as clear as I can make it, ie it won't work like ya think, that simple, if you realy want it to be shot full of holes I'll drag Epiom into this, he has ALL the data to shoot down the idea.

The chiller I built for a luagh pulled down to -40 with out a hick up, It was big and mean and power hungry, I could have made it way smaller and still goten to -30 to -40C with out truble and still use less power then a tech!

serialk11r
04-13-2007, 10:01 PM
and I hear of some people who test dynamite befor using it, what can I say?

Thier power usage goes up and cooling ability goes to zero, IE they are just like a car spinning its weels, a whole lot of nothing at a whole lot of power usage. with out going into complex physics and boring you thats as clear as I can make it, ie it won't work like ya think, that simple, if you realy want it to be shot full of holes I'll drag Epiom into this, he has ALL the data to shoot down the idea.

The chiller I built for a luagh pulled down to -40 with out a hick up, It was big and mean and power hungry, I could have made it way smaller and still goten to -30 to -40C with out truble and still use less power then a tech!

The physics part...take a quick look at "carnot cycle" on wikipedia, and look at the equation...its got some calculus and crap in it but you should be able to see something...it explains why the power usage goes up super high but you're not getting a lot of cooling. I mean like, of course if you bench you want the cold, and don't give a crap about power usage, but TECs are so inefficient to start with, and like the cold temps makes it...nvm

Starkiller42
04-13-2007, 10:43 PM
Anyway, for a pump I might suggest an Iwaki MD-20. They're outrageously expensive new, but you can usually find one for ~$60-75 on ebay. They have more head than you could ever use, and separate the motor from the impeller housing, which is also seal-less (you get to avoid that lovely o-ring strip described before). If you want DC, the RD-20 or 30 are excellent, but almost impossible to find. These things were made for industrial applications, like pumping chemicals etc, so they'll survive any coolant you throw at them, and I've never heard of one failing.

I hope that helps!

serialk11r
04-13-2007, 10:50 PM
Petra is stocking RD 20s and 30s soon I think. Wonderful pumps, iwakis.

silverphoenix
04-16-2007, 03:34 PM
Will the Iwaki's plastic impellar housing be affected by the temps? what are pumps people are using right now? I don't mind doing some mild modifications.

serialk11r
04-16-2007, 03:43 PM
Will the Iwaki's plastic impellar housing be affected by the temps? what are pumps people are using right now? I don't mind doing some mild modifications.

People ARE using Iwakis in chiller's, that's what epion told you ;)
The plastic parts are rated to a low temp by the company, so they most likely can go that low, probably even lower.

Modzilla
04-17-2007, 05:19 AM
I'm using a Danner Supreme Mag-Drive, Model 3. It's an aquarium pump, got it at petsmart. Plastic impeller and impeller housing, no problems with low temps. But, you need to take off the impeller housing, discard the o-ring, and seal with a low-temp industrial adhesive. also seal all threads with the adhesive. the mounting part could be better. It's a decent pump - for the money. Next time I'll spend the money for something more "industrial."