Please post your Glaciator temps

[Len]

I've got a Glaciator installed on a 1.2 Tbird running at 1425 @ 1.95v. In my Antec SX830 I have 3 x 80mm lower CFM intake fans in the front and 2 medium CFM exhaust fans in the rear + the power supply fan. If I had to guess I'd say I'm probably moving about 50-75 CFM through the case. Air flow is very unrestricted and seems balanced but I might have a little negative pressure. I tried about 14 different fan combinations and this one resulted in the lowest case and CPU temps for me. However, I still feel like my CPU temps are too high. A 23C difference between the case temp and CPU load temp seems a lot for a high end cooler like the Glaciator.

My temps:
27C - System (ambient case temp I assume?)
43C - CPU idle (no cooling tweak)
50C - CPU load (CPU Stability Test program running for 15+ mins)

The Glaciator itself seems to stay VERY warm when the CPU is loaded. I'm wondering if my CPU temps would drop if I could find a way to cool the heatsink better. I tried taking the side case cover off and blowing a higher CFM 80 mm fan on the Glaciator from different directions & angles but the loaded CPU temp didn't change at all. Short of a blow hole (which I'd rather not do) does anyone have any suggestions to lower my temps further?

Also, please post your Glaciator temps. Maybe my temps are normal and I'm just worrying too much about it.

 

[boyermd]

i'm getting about 35C idle and 65C load (!!). i'm not satisfied w/ my job on the artic silver II, so i'm gonna be redoing that tonight. i also tried changing airflow in my case to see if i could improve those Loads temps, but no good. The Glaciator seems to pull air down onto the heatsink instead of off, so i changed my rear high-speed fan to blow air towards the HS instead of out of the case, didnt make a bit of difference.

 

[Hoot]

Len (Jun 27, 2001 11:01 a.m.):
I've got a Glaciator installed on a 1.2 Tbird running at 1425 @ 1.95v
-snip-
My temps:
27C - System (ambient case temp I assume?)
43C - CPU idle (no cooling tweak)
50C - CPU load (CPU Stability Test program running for 15+ mins)
-snip-

So you are running 91W of heat.

50C-27C=23C
23C/91W=.25c/w

Joe got .20c/w running at 96W, as measured with the Glaciator drilled and a thermocouple sitting on top of the core. You should be able to get a similar result. At .20c/w you should have gotten a temp of 45C, but you did not. For the chance of improving 5C, I would take off the Glaciator, degrease it and the core and try remounting it again. That, having been said, the fact that you are successfully achieving 1425 Mhz in an ambient of 27C is a pretty good achievement. I would not accept that temp, but I am more fussy than is probably defendable.

Hoot

 

[Len]

I came to exact same conclusion regarding the 0.20 C/W. I find these numbers sort of odd. Before I added more fans to my case I had a case temp about 10C higher or about 36C. At that time my CPU temps were only 4C higher or 54C, not 10C higher as you would expect with a fixed C/W value. So doing that math with my OLD fan setup I got:

54C - 36C = 18C / 91w = .20 C/W

BUT, after adding fans and dropping the case temp to 26-27C my CPU temp only dropped 4C from 54C to 50C. This part puzzles me a bit since I would expect the C/W to 'appear' constant and the CPU drop close to 10C also.

Can you help me understand why this is the case?

 

[CSaddict]

ambient 33C,full load 48C. This is way better than the stock set up I had. Night and day difference. Yes its hot in here!

 

[ripper]

T-bird 1000 @ 1400 (1.85v; 10x140)
Epox EP-8kta3

system Temp 30C
Idle temp 41C
Load Temp constant 48C running cpuburn.

 

[Len]

I did more temp testing last night and learned that my inside case temp as reported by MBM5 is at room temp or sometimes 1C above room temp. So not much room to improve there.

Without the A/C running the room/case temp was 26-27C and CPU load temp was 50C. Running the A/C dropped the room/case temp to 22C and the CPU temp dropped accordingly to 46C. So, alternating nothing by the room temp keeps my C/W fixed at around 0.25 (23-24C above ambient and 90-91 watts of heat).

I tried turning off one intake and one exhaust case fan (still had 2 on each end running) and with a room temp of 26-27C my CPU temp shot up to 53C and I crashed running the CPU Stability tester. MBM still reported my case temp to be within 1C of room temp so you can see this sort of changed my C/W value which I know is impossible. It seems these calculations are based on a FIXED case/fan configuration or maybe I'm just confused.

 

[Len]

| Nevermind, took advice to delete my cookie for this site and now I can see again!!

BTW, can everyone see this topic just by clicking on the 'Cooling' category? For some reason its invisible to me. The only way I can get to it is by clicking an email notification link or using the forum search engine to look for 'Glaciator'. Wierd. I don't have this problem with other topics I've started and/or posted to.

 

[Len]

Update: I made one simple change ... I swapped one of the medium cfm fans in the rear with one of the slower cfm fans in the front. I did this to better balance the airflow since it seems I was running with negative pressure which most agree is bad. With just that one change my CPU load temp rose 3C from +23 to +26 above ambient.

I don't know why the big difference. Maybe the Glaciator is very sensitive to air around it. Maybe its just more inaccurate MBM readings. Maybe more of the heat off the backside of my GF2 Ultra is rising to the CPU area instead of being exhausted out the rear of the case.

One thing is for certain ... this case cooling business is tricky stuff.

 

[Len]

CSaddict, you have 33C ambient and 48C load temps for 1100 Mhz at 1.80v. That's +15C cooling 64 watts for a C/W of .23 with a Glaciator.

ripper, you have 30C ambient and 48C load temps for 1400 Mhz at 1.85v. That's +18C cooling 80 watts for a C/W of .225 which is very close to CSaddict.

my current setup, I have 26C ambient and 50C load temps for 1425 Mhz at 1.95v. That's +24C cooling 90 watts for a C/W of .267 which seems high for a Glaciator.

Previously, I had 26C ambient and 46C load temps for 1377 Mhz at 1.85v. That's +20C cooling 80 watts for a C/W of .25, still a bit high.

I'm not sure why I'm doing worse than others. Maybe MBM variations? I've tried reapplying the Artic Silver once already. I put a tiny bit on the core, spread it around gently with my finger. Then I used a razor to thin it out removing the excess. I found the razor to not be perfectly smooth leaving some lines of missing AS on the core so I spread the rest around a little more with my finger. I did my best to apply as thin a coat as possible while still covering the core. Given the amount I used, I could probably do another 20+ CPUs with the amount of AS included with the Glaciator. I don't see how it could be too thick.

 

[crunch]

Specs:
T-Bird 1333/266
Millennium Glaciator
Iwill KK266 w hsf on Northbridge
Antec sx1040 w/4 80mm fans 2-in, 2 out
GF2 Ultra

All temps recorder using the questionable MBM 5.04
(if its really 4+ degrees high, all the better!)
Room temp and Case temp verified with Raytec temp gun.

Results BEFORE Glaciator using Alpha PEP66/Sunyon fan@ 1466mhz 1.7v
Room Temp 26
Case Temp 27
CPU temp 29 idle, 50 after 1hr prime95

Results After Glaciator @ 1466mhz 1.7v
Room Temp 26
Case Temp 27
cpu temp 29 idle, 46 after 1hr prime95

Results After Glaciator @ 1512mhz 1.81v
Room Temp 26
Case Temp 27
cpu temp 29 idle, 49 after 1hr prime95

Prior to Glaciator, 1466 was highest stable Oclock I could achieve.
The most noticeable difference is the recovery time
(the amount of time to return to idle temp from max)
With the PEP66, temp would drop 1 or 2 degrees per min avg.
the Glac drops 4 or 5 degrees per min so It seems to bleed off heat much faster than the PEP66.

Overall opinion is its a KEEPER!

 

[waver123]

Tbird 1466/266 (1333 AXIA)
Millenium Glaciator
1.85V 133 x 11

44C loaded 27 case (prime95)
38-39C idle.

Huge difference from my old volcanoII.