Ultra ChillTEC Performance

prime81 · Feb 26, 2008

After my aquagate pump died I decided to go back to air...of course a stupid move on my part air didn't even come close to comparing as with loads I was pushing 58c. I started researching cooling which involved other water kits, the thermalright xtreme, and of course the vapochill units.

From a price perspective I didn't wanna go over 150 so that ruled out the vapochill systems. I read reviews on the ChillTEC saying it's not to good for Quad CPUs. But I also found a few reviews that had quad setups and the temps were quite impressive. Anyway I ran across a couple on Frozencpu.com and picked one of them up.

With that in mind here is my current setup-

Q6600 G0 @ 3.15ghz 350x9 vdroop 1.192vcore idle/1.141vcore load
ASUS Commando
Ultra ChillTEC
Sapphire 2900 Pro 1gig
Coolermaster 830 Stacker

Air: 40c-41c Idle and 58c Load
Aquagate: 37c Idle and 52c Load
Ultra ChillTEC: 33c Idle and 46c Load

Thus far I'm impressed, I may even give her a shove up to 3.8ghz and see how it performs. Also please note that I did lap my CPU to a 1500 grit finish.

Automata · Feb 26, 2008

I promise those temps will be substantially higher at 3.8, the peltiers will not handle that load.

Drinkyoghurt · Feb 26, 2008

and also, 1500 grit is only for the pretty, it doesn't make heat transfer better, it might even make it worse then 600-800 grit

prime81 · Feb 26, 2008

Well 1500 is highest i could find around here... I did 220, 400, 600, and then 1500. It has a semi mirror finish and from what I can tell it did decrease the temps and increase heat transfer. I use artic silver 5... I was able to push 3.8ghz on air but was too hot.

One thing you fellas did overlook is the thermal output for the q6600.

http://processorfinder.intel.com/details.aspx?sSpec=SLACR

This cooler is not entirely based on just the peltier...the peltier only kicks in over a certain temperature. Therefore the basic concept of the Ultra ChillTEC is actually air cooling until the peltier is needed which then can max out a 50w dissipation of heat.

So depending on how well ur CPU is lapped, how well you apply the heatsink grease, and how good you setup the exhaust on the chilltec... is what makes a difference.

I saw a 4-5c difference from when I had not lapped to after I did lap. Guess the 1500 isn't so bad after all?

Automata · Feb 26, 2008

prime81 said:
One thing you fellas did overlook is the thermal output for the q6600.

I didn't overlook it, I stated in my first post that it would overload the unit. :beer:

uberwoot · Feb 26, 2008

95w lol try about 135w at 3.25ghz

oc that and watch that Q6600 beat on that little tec.

btw at 3.3ghz i got a 43c load after 24hours of prime with just water so it dosent look like that tec is helping out to much. I would like to see how this compaires to a ultra 120

prime81 · Feb 26, 2008

uberwoot said:
95w lol try about 135w at 3.25ghz

oc that and watch that Q6600 beat on that little tec.

btw at 3.3ghz i got a 43c load after 24hours of prime with just water so it dosent look like that tec is helping out to much. I would like to see how this compaires to a ultra 120

Your on water....congrats on the 43c load at 3.3ghz(366 x9). Some expensive watercooling kits or rigs I should say would do a better job.

I pushed 3.6ghz(400x9) and it was idling at 46c, load 63c.
I pushed 3.8ghz(425x9) and it was idling at 51c, load not tested.
I pushed 4ghz(445x9) and that was the breaking point...it loaded windows but was idling at 60c, I didn't push the load at that temp.

So if your saying it will overload the unit at 3.8ghz, no it doesn't. Could the temps be better? Sure they could, but I dont push those speeds for daily use, simply for benching and testing stability.

However I will probably run 9770 speeds of 400 x 8 @ 3.2ghz daily as the temps and vcore were almost identical to my current daily setup, just gotta put it through some torture and make sure it's solid.

As far as the TEC is concerned, it's pretty impressive. There is no water, no hassle of setting it up, if the peltier fails it's still air cooled. Compared to the aquagate, and artic square I had... it's like night and day.

prime81 · Feb 26, 2008

It can definitely pull it's weight, based on a calculation looks like 3.96 ghz is around 164w output, correct me if I'm wrong.

The temps are definitely high, but as you can see it does hold it's own.

Shiggity · Feb 26, 2008

prime81 said:
It can definitely pull it's weight, based on a calculation looks like 3.96 ghz is around 164w output, correct me if I'm wrong.

The temps are definitely high, but as you can see it does hold it's own.

Are those temps @ Idle? You'll shoot into the 80s on a torture test.

prime81 · Feb 26, 2008

I did not torture at this speed, as the jump will be an estimated 10-12c. In the 80s nah.... not quite.

Yes they are idle...so far my highest stable OC has been 4.05ghz. I'm just too chicken to keep her there lol

uberwoot · Feb 26, 2008

Post some screen shots of you running prime with core temp open. Just because you got low idle temps dosent mean its stable and your load temps will only go up by 10.

This heatsink was a good idea thats about it. im not tec hateing i got one right here on my desk its just so easy to overload them its better served as a paper weight.

prime81 · Feb 26, 2008

Read up on the chilltec before ya bash the TEC.... as I said in an earlier post it's a combo of air/TEC. I could turn the TEC completely off and run air with the chilltec. I could probably stand to torture at 3.8ghz, but I'm not gonna push it at close to 4.

With the decrease in die size from 65nm to 45nm, and eventually 32nm, and 22nm(i think). Less power needed for higher clocks, less thermal output = TEC friendly. Hence the 45nm q9550 is rated at 95w...technology is good!

uberwoot · Feb 26, 2008

im not bashing i know how it works nd i know how tecs work. All i asked for was some screen shots of prime 95 running and core temp because i wass see how it does.

prime81 · Feb 26, 2008

Here's some for ya....
Idle..

Load..

Hazaro · Feb 26, 2008

Give it a few more minutes

I want to see how hot it gets, because I was looking into one.

prime81 · Feb 26, 2008

From my perspective it looks like 3.5-3.6 on a quad is about the best efficiency for the chilltec. Any higher and on load it wouldn't be good... ran it for 10 mins it stuck around 63-64c. On a dual core...probably wouldn't hurt it to run 4-4.5ghz. But I would stick with a 45nm E8400 or E8500 which both have a 65w thermal output @ 3ghz.

uberwoot · Feb 26, 2008

Let it run for atleast 30min then post results. If its deacent i might pick one up to mess around with.

Edit. Witch the quickness Stratus lol

Stratus_ss · Feb 26, 2008

that isnt loaded long enough to be a real measure of tempurature. You should take a screen after 20-30 mins. I am pretty sure you will see a bump in temps by at least 5C

edit: damn you uber! you beat me by a few seconds lol

dudleycpa · Feb 26, 2008

Stratus_ss said:
that isnt loaded long enough to be a real measure of tempurature. You should take a screen after 20-30 mins. I am pretty sure you will see a bump in temps by at least 5C

edit: damn you uber! you beat me by a few seconds lol

and you beat me!

At least a half hour and if you really want to stress your system run Orthos Small FFTs and priority set at 10.

uberwoot · Feb 26, 2008

oh yeah why are 2 of your cores that much hotter? You got that heatsink mounted right?

Ultra ChillTEC Performance

Member

Destroyer of Empires and Use

Member

Member

Destroyer of Empires and Use

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Overclockix Snake Charming Senior, Alt OS Content

Member

Member

Similar threads