View Full Version : Swifty releases new pelted chiller module.
DaGooch
09-28-2003, 04:48 AM
The MCW-CHILL 452™ is a refrigeration unit using thermoelectric technology to chill liquids below the ambient temperature. This product is designed as a cooling module for commercial and industrial refrigeration applications such as:
Electronics
Automotive
Military
Medical
Looks nice but 1/4" ID tubing?
saaya
09-28-2003, 05:29 AM
yeah, way to small.
Belgian_Dude
09-28-2003, 05:31 AM
330 §§§§ING dollar I read :o Are they insane??? You can buy a 2nd hand prome for that amount of money :p (almost ;) )
KnightElite
09-28-2003, 06:50 AM
My friend is getting one of those as a review sample. I should be helping him review it this coming weekend.
afireinside
09-28-2003, 07:07 AM
Damn 330 thats crazy. Mine is gona be like 110 after a brand new 5200 BTU window AC, cooler and taxes...
saaya
09-28-2003, 07:58 AM
lol 330$? wtf? nobody will pay that much money for it! the 227Wtec waterblock bundle costs 110$, who will pay 330$ for a chiller if he has the chance to slap a 227W tec on his cpu for 30% of the price and get the same or better results?
Tedinde
09-29-2003, 07:15 PM
Dam thing will cost you out the XXX in electric too!!!!
DaGooch
09-29-2003, 08:01 PM
It seems the recommended setup for the dual 226W Peltiers is 2 Meanwell power supplies.
OPPAINTER
09-29-2003, 08:26 PM
Originally posted by KnightElite
My friend is getting one of those as a review sample. I should be helping him review it this coming weekend.
KnightElite,
What are you guys going to test it on? A CPU?
It will be interesting to see the review.
OPP
KnightElite
09-30-2003, 04:46 PM
Don't know if you're familiar with him or not, but my friend is pHaestus from procooling.com . We're going to be testing it on a Die Simulator, I think, possibly on a system as well. We should also be testing several TEC waterblocks.
OPPAINTER
10-01-2003, 07:15 AM
Originally posted by KnightElite
Don't know if you're familiar with him or not, but my friend is pHaestus from procooling.com . We're going to be testing it on a Die Simulator, I think, possibly on a system as well. We should also be testing several TEC waterblocks.
What I'm thinking is that the cold side loop has such small ID that it would be tough to run it through a normal water block.
Yea I know of pHaestus:)
Let me know once your guys review is complete, I'll check it out.
Thanks,
OPP
KnightElite
10-01-2003, 08:38 AM
Will do. Looks like the testing will probably be on Sunday.
gouda96
10-01-2003, 11:05 AM
seems like swiftech has come back to life. They seemed dormant for the last year and a half. Even if this isn't the best of products it could be a sign of things to come.
They have released what appears to be the best air cooled nb cooler i have seen. They also redid their waterblocks to naturally take 1/2" id tubing. I don't know how much of a difference that will make, but I love the timing...a month after I bought my mcw5000-pt, and a week after buying my mcw50-t :D
ellsworth
10-01-2003, 11:56 AM
why not put the 226w pelt directly on your cpu...it would be much more efficient. i can imagine people wanting to cool the water loop from their vid card and nb too, but still...the idea of a pelt waterchiller always has rubbed me the wrong way. a phase change chiller would be cheeper, but it would take up more space which is one of its few drawbacks. oh and both would use a large amount of power. swiftech does make some nice products, i have owned quite a few.
gouda96
10-01-2003, 12:03 PM
you could chill the pelt that you have on your cpu. I don't know what kind of a difference that would make though...
OPPAINTER
10-01-2003, 12:39 PM
Originally posted by gouda96
you could chill the pelt that you have on your cpu. I don't know what kind of a difference that would make though...
I use to run my 220 pelt on the cpu and cooled it with -15c water. It made a huge difference.
A cool pelt is a happy pelt:D
OPP
Originally posted by OPPAINTER
I use to run my 220 pelt on the cpu and cooled it with -15c water. It made a huge difference.
A cool pelt is a happy pelt:D
OPP
What idle temps and @ what vcore you were getting Opp?
I wanna go this route (chilled water to cool my pelts) and would like to konw how it works.
THX :)
OPPAINTER
10-01-2003, 01:43 PM
Jabo,
I can't remember at all.
Temps were well below zero though. But this was with a lot less Mhz cpus so temps would have ran hotter on todays clocks.
OPP
Thanks, man :)
I am getting -6C idle on straight water (with some redline additive) and water is 31.2 hot line, 29.9 cold out-of-the-rad line @22.3C ambient air temp. Those are idle temps while net surfing :)
So, if I got my water to -15C, 45 degrees drop, would I have -50 my cpu temp idle?
saaya
10-01-2003, 02:19 PM
why not put the 226w pelt directly on your cpu...it would be much more efficient. i can imagine people wanting to cool the water loop from their vid card and nb too, but still...the idea of a pelt waterchiller always has rubbed me the wrong way. a phase change chiller would be cheeper, but it would take up more space which is one of its few drawbacks. oh and both would use a large amount of power. swiftech does make some nice products, i have owned quite a few.
4x55W tec 22€ (12v)
6x30W tec 18€ (5v)
400W cheapo psu 20€
old 200W psu 5€? :D
400W tec power for 60€... still think a tec chiller is epxensive? :D
jabo, i dont think youll reach -50°C idle... maybe -35°C
OPPAINTER
10-01-2003, 02:36 PM
I would questimate in the -20s.
OPP
gouda96
10-01-2003, 07:44 PM
hm...you would have to make some sort of compartment for the water to pass through the chiller, and i wouldn't know how to do that.
Originally posted by gouda96
hm...you would have to make some sort of compartment for the water to pass through the chiller, and i wouldn't know how to do that.
That's the easiest part, mate:)
My problem is sourcing an appropriate chilling unit... something powerfull enough to be worth my time...
gouda96
10-02-2003, 05:25 PM
jabo, do you have your mcw5000-pt and mcw50-t on the same loop? How big is that saab heatercore. Right now I only have an 86 chevette heater cored, and an eheim 1250 pump for both of them. I am gonna set it up and see what the temps are like. If they are too high I think I will go grab a nice big heatercore for the cpu, and either a realy nice pump for the cpu line, or a 1048 for the gpu.
Holst
10-06-2003, 05:40 AM
Originally posted by saaya
4x55W tec 22€ (12v)
6x30W tec 18€ (5v)
400W cheapo psu 20€
old 200W psu 5€? :D
400W tec power for 60€... still think a tec chiller is epxensive? :D
jabo, i dont think youll reach -50°C idle... maybe -35°C
Pelts dont perform as well when chiled with cold water.
Most people with good chillers (-20 or less) usually find that they do better without the pelt. I think opp was right on the limit of TEC usage at -15.
I think your PSU is going to blow up saaya, no way it can take those loadings for any length of time.
Id love to see you proove me wrong :D
And I still dont think a TEC chiller will work on a CPU, maybee on a GPU, but as others have stated a 172 or 220 on the CPU/GPU will be a much better option.
saaya
10-06-2003, 06:10 AM
im not chilling the tecs, im using the tecs to chill the water and use that cold water to cool the system.
im using 2 psus for those tecs, my 200W samsung (has powered 4 30W tecs for a few weeks already) and the 400W leadman i got for 15$ to power the 55W tecs and 2 30W tecs.
KnightElite
10-10-2003, 01:21 AM
Alrighty, we got some testing done today, and here's the preliminary results thus far (see farther down the thread, big pics in the post):
http://forums.procooling.com/vbb/showthread.php?s=&threadid=8126
Originally posted by gouda96
jabo, do you have your mcw5000-pt and mcw50-t on the same loop? How big is that saab heatercore. Right now I only have an 86 chevette heater cored, and an eheim 1250 pump for both of them. I am gonna set it up and see what the temps are like. If they are too high I think I will go grab a nice big heatercore for the cpu, and either a realy nice pump for the cpu line, or a 1048 for the gpu.
Yes, both MCW5--P-T and MCW50-T are on ther same loop:
RAD -> CPU -> GPU -> Eheim1250 ->RAD
My rough calcs for my h'core is ~3.5m^2 of heat transfer area )very thin copper fins very densly packed requring LOTS of air flow). Dims of finned are are :
175mm wide, 160 high and 100-120mm deep. 7 flat pipes in two rows.
I am quite happy with my temps (MCW50-T screams for stronger TEC!)
I am going to drain my system this weekend, add more insulation and see how it effects temps.
In your case you may try to increase air flow through your rad if CPU temps are too high. I am getting such a god temps a=only because of very high air flow.... Witgh second rad you'll be better off but got to pay for it ($$)...
But then again, fans are not for free niether :)
Originally posted by Holst
Pelts dont perform as well when chiled with cold water.
Most people with good chillers (-20 or less) usually find that they do better without the pelt. I think opp was right on the limit of TEC usage at -15.
I think your PSU is going to blow up saaya, no way it can take those loadings for any length of time.
Id love to see you proove me wrong :D
And I still dont think a TEC chiller will work on a CPU, maybee on a GPU, but as others have stated a 172 or 220 on the CPU/GPU will be a much better option.
With TECs 'not performing' below ambient is just simple misconception and lack of understanding of working principles of TECs.
On manufacturer's web sites guaranteed working temperatures spread is form -50 to +200C. It mens that they will work as specified under these conditions.
They are the simplest possible het pumps, solid state devices.
When you're dropping temps of coolant below zero another factor comes into play. There's such a difference between cold side of a TEM and ambient temps that such a module s cooling mobo/vid card's PCB as well as CPU/GPU (it contains LOTS of copper!).
If you insulate exessively (compared to ambient water setup) which basically means covering as much as you can in as thick as you can get neoprene, you'll be massively suprised how low temps will drop.!!!!
No your wrong tecs efficencys (and dT's) drop exponentially with respect to hot side temperature, its a similar principle (although different physics) why people using aircon compressors rated @ 5000btu @ 45f find the capacity drops to 300btu @ -45f in there direct die setup.
The same is true with tecs, in fact if you use the kryotherm simulator @ -50c the efficency of tecs is close to 0, i.e. you get no amount heat pumped for the input power. Likewise the efficency goes up lots with hotside temp at around 100c hotside you get very close to the predicted dT and heat pump capacity.
]JR[
gouda96
10-10-2003, 08:59 AM
Come on guys...bring out some physics and chemistry equations that are way over my head!! I want physical proof right now, and I won't be sattisfied until I have spent over an hour trying to decifer the eq's...maybe soon I could join in on this conversation being that I am taking both chemistry and physics at georgia tech :D
jabo, thanks for the advice, but I ended up trading my whole water cooling system for a prommy :D. I have an eheim 1060(600gph) here, so I am just gonna go pickup a heatercore that is a good size for 2 120mm fans, and get another mcw50-t and a swifty nb waterblock.
Anybody know what heatercore is good for 2x120mm fans on the same side?
KnightElite
10-10-2003, 09:02 AM
gouda96: I have two 1973 Chevy blazer heatercores (at least I think that's what they are) and they are definitely large enough (see link I posted above for pics, or see link in my sig).
gouda96
10-10-2003, 09:07 AM
What size tubing did you use? I see you left the stock barbs on there. Any idea how much one of those would cost me at a napa or autozone?
What was the final verdict on the swifty chiller?
KnightElite
10-10-2003, 09:32 AM
Further testing is still continuing, there will be an article written up sometime in the next week or two I would imagine. See the thread at procooling for further developments.
I use 1/2" Silicon tubing, and the barbs are 5/8".
gouda96
10-10-2003, 09:41 AM
oh...5/8 that is nice. No need to bother cutting them off and sticking on 1/2" barbs. I will just grab some more clearlflex I guess...only cost me like $10 for 10ft, so no big deal.
Originally posted by ]JR[
No your wrong tecs efficencys (and dT's) drop exponentially with respect to hot side temperature, its a similar principle (although different physics) why people using aircon compressors rated @ 5000btu @ 45f find the capacity drops to 300btu @ -45f in there direct die setup.
As far as I cn understand it BTU rating for any AC unit is constant at all times.
In your example this unit 'uses' 4700btu of it's capacity to maintain -45f and tat's why you get only 300btu cooling capacity left to cool your ggpu and still maintain -45F. It doesn't just dissapear into the thin air;)
The same is true with tecs, in fact if you use the kryotherm simulator @ -50c the efficency of tecs is close to 0, i.e. you get no amount heat pumped for the input power. Likewise the efficency goes up lots with hotside temp at around 100c hotside you get very close to the predicted dT and heat pump capacity.
]JR[
Imagine running such an AC unit in an environment of ambient temp at -35C (I am mertic, sorry). What temps do you think this AC would be able to achieve @ it's evap?
THat's the case of a TEC cooled with -35C coolant. In ideal world if you could keep energy losses to 0 (perfect insulation of CPU/TEC assy - it cools ONLY CPU/TEC) a 226TEC would produce temp of -85C on it's cold side with no heat source applied - this is when there's NO heat being pumped . TECs work at highest efficiency (pump it's rated heat amount) when dT is zero , there's no difference between cold and hot sides.
If manufacturer states working paremeters of -50C to +200C it meanss that such a module WILL pump it's rated heat in this temps brackets. Below or above it iwill not work as rated.
Waht you wrote about TECs being most efficient at 100C hot side is miselading. It has nothing to do with TECs. It only applies to radiators where the bigger temps difference between coolant and ambient rate the better heat transfer.
Solid state heat pumps do not work like this.
There is only one physics ;)
If you have any test results I'd be more than grateful if you could post them, cheers mate :)
Originally posted by OPPAINTER
It will be interesting to see the review.
OPP
in magical times, long ago there was this crazy warrior who decided to make his own water-chiller using a line-up of 80W peltiers and a seperate watercooling loop to cool them down.. I can't remember his name.. maybe this picture may help to refresh someone's memory.. I just can't seem to be able to find his name ;)
:stick:
KnightElite
10-13-2003, 12:48 PM
Opp used to have something similar with 226W TECs. The thread is still in the forums, but all the pics are fux0red :(.