• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Glaciator results not what I expected

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Twofan

Registered
Joined
Jan 15, 2001
After much anticipation I have to say I am not seeing what I had hoped. Was running an OCZ Gladiator with the standard 32cfm fan. MBM5 results were:

Idle = 23/41/31
Load = 23/46/31 with 100% cpu useage

Installed Glaciator with AS and a non-conductive spacer. Results are:

Idle = 23/49/31
Load = 23/51/31

Thouht maybe didn't get it seated properly so reinstalled, same results. I really am at a loss. I don't know whether to trust people saying MBM5 gives errant results, or trust MBM5 and go back to the old hsf, or figure the temps aren't that bad ( they look awfully high to an ex-intel guy) and run it like it is.
 
Update, setting here watching it rip a movie and temps get even higher, 23/54/33. Looks like I have a $45.00 paperweight.
 
twofan

the claim seems to be that the glaciator gets higher In socket thermistor temps than other HS.

just the opposite case of the Thermoengine which gives lower in socket thermistor temps than other HS.

be that as it may, it seems to me that the gladiator is an excellent HS. probably the only one clearly better would be the swiftech.

it seems to me that the cak (i have no experience w/ this one) the SK6 (i have this one) the Gladiator (i have this one) and the Glaciator really are neck and neck. all very good hs.

and despite the fact that i have 2 good copper hs, it's the WBK 38 sittin on top of my hs right now. Why you ask? to be honest unless ur running a lotta air over the hs (delta, noisy) the copper doesn't seem as effective as aluminum. plus the wbk is pretty large, allowing me to put a 40 cfm 80 fan on top of it (no need for an expensive adapter) and voila good results low noise.

too bad about the glaciator, i'm just glad i didn't spend the money on one. i wasted enough money on my two copper hs's.
 
Did you have the non-conductive shim under the OCZ Gladiator before the Glaciator? That would represent an imbalanced test, I would think, if you never spec'd the impact that shim has upon temps. As much as I hate to damage a CPU just as much as you or anyone else, I do not use shims. I don't trust their thickness compliance. All it takes is the shim to be as thick as the height of the core and you will have competition for the seating pressure on the core. Now that you have a feel for puttin on and taking off the Glaciator, try it without the shim and for goodness sake, let it settle in overnight.

Hoot
 
You may be right Hoot. I just got back on. Had to change back to Gladiator. Temp got past my 55c alarm and shut everything down. I will try again with Glaciator, maybe lapping it and NOT using the shim. I know I liked the noise level much better. So many people have had such good results, I don't understand what happened. maybe some others have an idea. I hate to not get to use it.
 
Remember MBM temps are effected by the Glaciator. There was a write up done on this on the front page. The Glaciator actually produced a lower temp, but MBM (via the insocket thermistor reading) showed a higher temp. Very bizaar, but I assume it's due to air currents.
 
I have read the articles and all the posts the last few weeks. But I think the +7c difference that I got was too much. 55c is when MBM shut it down and I just did not want to trust the theory to let it go any higher. I may try again without the shim and see if I get any different results. If I don't hello Ebay
 
I also just got my glaciator this weekend, i was highly anticipating its arrival, but have since been disappointed. but i do figure i did a poor job w/ the artic silver II job because i'm getting Load temps over around 64C!! which is awful. I'm running a GA-7DXR w/ 1.33 AYHJA Y wk20 and 256mb XMS2400 Corsair. I am also using a non conductive shim. This was my first heatsink, so i dont have anything to compare it to. My mobo does have a temp sensor beneath the chip, so it's supposed to read higher than in-socker sensors, but all the same 64+ is too high, way too high. I would try it w/o the shim but i'm just too worried about the whole "core cracking" thing.

I'm gonna play w/ it more tonight and see what happens, but so far i'm not impressed.
 
Boyermd

there is a good chance your shim is causing the high temps. mb lap the shim and thin it down just a bit, but that's risky too, cause if you lap it uneven then u can still get cracked core :(.

that's one of the reasons a lotta ppl here don't like using the shim.
 
I just got home and it is too late to start reinstalling hsf units. I am really interested in seeing IF the shim could have caused the higher temps. I don't think it could have been installation as I have done this alot and I think I know how to put on a hsf and apply thermal compound (don't we all though, hehehe). But who knows? I really was gonna pickup some sandpaper and lap the Glac before putting it back on, but I think that in the interest of O/C'n science that maybe I will reinstall the Glac without the shim first. Run some tests and then lap that sucker and then run some more tests. That should tell me something and anyone else thats interested, so I will post what I find out.
 
Plat (Jun 27, 2001 01:23 p.m.):
Boyermd

there is a good chance your shim is causing the high temps. mb lap the shim and thin it down just a bit, but that's risky too, cause if you lap it uneven then u can still get cracked core :(.

that's one of the reasons a lotta ppl here don't like using the shim.

That was my first thought after reading your thread. Try not using the shim and see if it works fine.

Aaron
 
Boy, I wish I had a dollar for every technical call or email that I have received about high temperatures that ended up being a shim problem.

If a shim is bent just 0.003" it can hold the heatsink off of the chip even though the bend is almost impossible to see. The Glaciator may be more susceptible to bent shim problems as it's clip does not provide the high attachment pressure of some other clips.

Before you use a shim, lay it on a piece of flat glass or put a razor blade across it to verify that it is perfectly flat. If you drop a metal shim onto a hard surface, throw it away, you will never get it flat enough to safely use.

I am sorry if I seem to be venting here, but I see this problem 2 or 3 times a day. People tend to get Arctic Silver II and a shim at the same time. When their temperatures go up, they call or email me complaining about the AS2. Inevitably, once I get them to remove the shim, the temps drop down to a comfortable level.

Nevin House
Arctic Silver
 
thx for the info. my shim is actually non-flat. i thought it would be alright since the heatsink would pressure-flatten it, but i guess that was a wrong assumption. i could also tell it wasnt flat when i removed the heatsink and found the ASII was not evenly distributed across the underbelly of the hs.

i contacted overclockershideout.com and requested a new one, they said they'll exchanged after i send this one back. not impressed w/ them.

a side rant about them, they charged me $8 S&H for a $8 shim, because, "only through ups could they get delivery confirmation". BS, i received another package the same week through USPS and they also have tracking/delivery confirmation. grrr.
 
Plat (Jun 26, 2001 10:12 p.m.):
twofan

the claim seems to be that the glaciator gets higher In socket thermistor temps than other HS.

just the opposite case of the Thermoengine which gives lower in socket thermistor temps than other HS.

be that as it may, it seems to me that the gladiator is an excellent HS. probably the only one clearly better would be the swiftech.

it seems to me that the cak (i have no experience w/ this one) the SK6 (i have this one) the Gladiator (i have this one) and the Glaciator really are neck and neck. all very good hs.

and despite the fact that i have 2 good copper hs, it's the WBK 38 sittin on top of my hs right now. Why you ask? to be honest unless ur running a lotta air over the hs (delta, noisy) the copper doesn't seem as effective as aluminum. plus the wbk is pretty large, allowing me to put a 40 cfm 80 fan on top of it (no need for an expensive adapter) and voila good results low noise.

too bad about the glaciator, i'm just glad i didn't spend the money on one. i wasted enough money on my two copper hs's.

Hey any info on how you modded the WBK would be great I was thinking of doing something like that because the delta is too fecking loud!
 
I fi can recall right (dont trust me) but the WBK uses a clip like the FOP sereis so mounting a 80mm is very easy... just use a 80 instead of a 60mm fan and off you go :)
 
yup

wbk uses wire clips to hold the delta fan to hs. u can use those clips to hold the 80 mm fan onto hs. very nice. easy.
 
Well here are the reuslts of my small adition to the Glaciator debate. Did an idle with the Gladiator and then full load with Sandra benchmarks looping. Results were
Idle = 22/39/27
Load = 22/43/27
Removed the Gladiator and installed the Glaciator WITHOUT the shim.
Idle = 22/45/28 (here is where I thought Oh Sh**, here we go again)
Load = 23/51/29
Remembering how the MBM figure can be +4c, maybe not that bad. I then removed the duct feeding air direct to hsf from 80mm fan in side. I thought maybe the way the Glac radiates this might be keeping air around the hsf.
Load = 23/52/29
I now know the duct is a help. Hey 1c is 1c :)
Now to take the Glac off and get to the lapping. I hate this part because in my CasEdge case with where the cpu socket is I have to either remove the m/b or the psu to get my fingers in to either hook or unhook the Glac's clip. After about 40 minutes working on it (it definately was NOT flat) reinstalled and here's the results.
Idle = 22/48/28
Load = 22/49/29
So the lapping gained another 2c and if it is -4c for MBM then it's running at 45c under full load of Sandra. I am starting to believe, maybe. Now for a real test. What killed it the other night was a movie conversion, so fired up TMPGEnc and started an .avi to .mpg conversion.
Load = 23/51/30 I believe this is REALLY 100% cpu useage. This same thing shut it down at 55c and still climbing before. I would say the removal of the shim helped but the lapping helped more. Now for another test. I was running bios voltage at 1.85 but MBM said it was 1.95 to 1.97 depending on the load. Changed the bios voltage to 1.80 and now MBM reports 1.91 Tried the same movie conversion and now results with the lowewr voltage are
Load = 23/49/31 and still stable and much quieter. The longer I have ran the higher the Northbridge temp has gotten. Maybe need to look into a Vantec for it. Temps as I type this LONG post (sorry) are 23/46/33
Conclusion I have come to is don't use shims, and always lap a heatsink, no matter how much it costs. And if the 4 degree difference is really true (which I cannot dispute) The Glac does a great job, now. Was it worth the $45 to replace the Gladiator? How can you put a dollar figure on noise difference. But the wife does like the noise level of the Glac much better and that has to count for something
 
I had read it awhile back. Now that the shim is gone and the surface is flat it seems to do a fine job. i plan to leave it in. I might be able to get a higher o/c but I cannot complain about what I have. Lower voltage and less noise at the same speed. I am much happier tonite than I was initially.
 
Back