I pulled the following quotes from thread http://forums.evga.com/tm.aspx?m=442271, but thought it was worth reposting (with my results) in a new topic, both not to derail the other thread, and to address a problem that for some unknown reason seems to be more common for the H50...

Vihsadas

HalloweenWeed

mhalstern

My core temps were fine, until the weather was >85 f.

I assume TIM means thermal paste?  I used one of those metal strips.  I did a good job of aligning it.  If I make through the summer, I'll keep the air cooler.  If there are too many problems, I'll go to water cooling.  Any suggestions?

The ambient room temperature affects core temps more than I had expected.  

Metal strips? ****? Correct me if I am wrong, but I have never heard of this. You should be using nothing but a good Thermal Interface Material such as Arctic Silver 5. You can get a free sample of a very good TIM from this guy, Falcon_CMH:

Re:All 920 owners!!

Look first to understand how to apply:

YouTube: How Thermal Compound Spreads

(And I recommend looking at some of those others too.)

Your temps will rise near linearly with ambient temp (room temperature). So when your ambient rises 9*F your CPU temps will rise 5*C (with air cooling).


Also, I have had my CPU above 100C! It was due to an E-Leet malfunction. When your CPU reaches 100C the multiplier (core) drops to downclock the CPU. I seen mine drop to 10x (with Speedstep off, I was horrified). So there is an automatic protector, reaching 100C will not 'break' your CPU. However, the life expectancy may be affected.

I have never heard of higher temps requiring higher Vcore. How long has it been since you blew out the dust from your case and cooler(s)? I just blew mine out, it had only been 2 mos and there was a lot! I was amazed as how quickly the dust gathered. It really does make a big difference in cooling.

I just wanted to re-emphasize this post. I have actually been using the spread method for so long. wow. It used to be the accepted method for years...Let's change that and see if there is any effect. Not being able to raise Vcore past 1.27 is really a pain in my butt.

WOW. X Method for the win!!

I just re-seated my Corsair H50 with the "X-method" from that Youtube video using Arctic Silver MX-3 and wow!
My temperatures didnt' *drop* so much persay, but temps stay remarkably constant as I raise Vcore past 1.3, which was impossible for me to do before reseating!
Here's my before and after at 1.27 volts and +100vtt, QPI 1.35 and CPU PLL 1.8 (3.8Ghz @ 19x200)

Before:
T CPU Case = 33-34 idle, 72-73 load
T Junction (The 4 "Cores" in E-leet) = 48-52 idle, load = 80-83C

After:
T CPU Case = 30-31 idle, 66-69 load
T Junction (The 4 Cores) = 41-43 idle, 74-79 load
These are temps @ 1.265 volts.  That's an average of about 4-5C drop in temps, but...that's not the best part!

Before, when I'd try to raise my Vcore to anything past 1.29 I'd get core temps that went as high as 90-92C! Now when I raise Vcore to 1.34 in bios my temps only become:
T CPU Case = 34-36 idle, 67-72 load
T Junction = 42-45 idle, 77-82 load
I can see 4.0Ghz in my future...My D0 seems not to perform as well as others and needs a lot of voltage to approach 4Ghz.
Here's what I did exactly with the H50:
1) I cleaned both surfaces with Arctic cleaner 1+2 and let dry cover with a piece of plastic (to prevent dust) for about 5 mins
2) Then I applied an X covering the central 1.5cm of the chip (yes that's pretty big), I did not add anything to the heatsink bottom
3) I then made sure I could pull up the mounting frame of the H50 enough so it wouldn't interfere with pushing the H50 straight down on the chip.
4) I pushed the H50 straight down on the TIM X, and holding pressure 100% of the time I twisted it back and forth 3-5 times to even out the spread. I used a fair amount of pressure-->enough pressure where the blood temporarily leaves your fingertips.
5) Still holding that pressure and not letting go, I made sure the H50 base was centered and tightened the mounting bracket.
6) I tightened it a lot ---> Until I couldn't tighten it any more.

Things that I usually do, that are different than this application: I usually "prep" both surfaces by putting some TIM on them, and spreading it around, and then basically wiping it off with a cloth. This leaves a very very small film on the surface of the sink (by thin I mean like DEFINITELY not enough to make good contact.) Then I'd spread the TIM with a credit card and put the heatsink down. I'll bet there were air bubbles forming underneath the sink and that's why my 920 temps were worse than suck. Just do the X method and keep constant pressure, and H50 owners will rejoice!
(I know my temps are still kinda high, but I have other airflow problems I'm troubleshooting).
Hope this helps some other H50 owners who were ready to send their sinks back...

(double-post)

What is your ambient?
I do get much better cooling out of my H50 than you do even with the x method at 3.8  1.26v.

I don't have specific numbers till i can get home but in Real temp
But the ball park numbers=

Ambient 25-28c
Idle:   29-34c
Load: 62-69c

I dont hit 80's until 1.35-1.4 vcore

I had high temps initially but that was with one fan as rear intake

Now i run push/pull with 2 50cfm Scythe S-Flex , a shroud on the push fan and it is set up as an intake from the front of my case. It took me a fair amount of trial and error to find the sweet spot in my system .

Thanks for posting this. It could be extremely useful to customers that don't have problems per se. Quite a few get different results when reinstalling their H50 and having to use aftermarket TIM as opposed to the stock TIM.

Those are core temps not the die temps (not 100% on the lingo)


Yeah the recommended rear intake failed from the heat of my 2-260's exhaust rising right up into the H50 intake

The rear exhaust was lackluster from the system heat being sucked in

Front intake works perfectly because it s pulling in fresh air and blows over top of my ram

Yeah i can upload a pic when i get home.