is it hot?

[diluzio91]

so i put together my friends sandy rig, and temps are... kinda lame.... im running an i5 2500k at all stock, with an h60 for the cooling, stock fan at the moment.

Compare to my i7 870 with a prolimatech mega. rev b

i5 - Voltage 1.176 - Idle 40 - load 70

i7 - Voltage 1.216 - Idle 26 - load 56

What the heck is going on here... the h60 is performing WORSE than the stock hsf...

[x88x]

Well, they're both 95W max TDP chips, so theoretically they should operate at similar temps with similar cooling. My first thought is to reseat the H60 on the CPU. It could not be seated properly and that's what's causing the temp problem. Also, check to make sure the pump is running. Also, what kind of fan(s) are you using on the H60? Is the ambient temp the same for both computers?

[dr.walrus]

Yeah that sounds like something's not quite right, I'm running a 1090T (125w tdp) overclocked with a h50, and I hit 58 when I'm folding @ 4Ghz.

MOST LIKELY
- pump not getting 12v
- poor ventilation
- poor seating

It's a funny thing to ask, but what's your weather like? Because if it's 35 outside...

But seriously, I'd put money on it being the pump not getting the full 12v

[mDust]

It's a funny thing to ask, but what's your weather like? Because if it's 35 outside...
Ambient temp makes a huge difference, even with watercooling...although less than straight air. My winter OCs are always much higher than in summer due to ambient temp. If one system is in an A/C'd room and the other is sitting in the summer sun next to an open window...:whistler:

Is the same software being used to report the temps?

Also, Intel is notorious for inconsistent, rippled IHSs. My q6600 looked like a stone had dropped into a pond. Lapping would definitely improve those temps if your friend doesn't feel the need for those warranties. If your friend is a wimp, I'd try reseating it with a rice-grain size dab of thermal paste in the center...sometimes spreading it out before applying the HS gets air trapped between the two.

[Fuganater]

Everyone stated the obvious already. My 2600k sits at 60-64 under full load.

[dr.walrus]

Ambient temp makes a huge difference

Absolutely - I'm sick of reviews that use straight temperature, rather than degrees above room temperature...



I'd try reseating it with a rice-grain size dab of thermal paste in the center...sometimes spreading it out before applying the HS gets air trapped between the two.
Personally, I use a considerable amount more on my h50, but that's because I leave the pump clamp in place, just loosen it, then rotate the pump into place and clamp it down. Technically, I think you're supposed to position it and then clamp it down, but because of the lazy way I do it, the paste is very smeared, and I worry about not having a clue about where I've smudged it to!

[Fuganater]

Absolutely - I'm sick of reviews that use straight temperature, rather than degrees above room temperature...

Its tough to do that unless you have a thermometer in your office.

[Twigsoffury]

Its tough to do that unless you have a thermometer in your office.

or a built in 4 way LCD Thermometer/fan controller. then you'll always know room temperatures. Hell i know the temp of the air coming into my intakes, and my exhaust temperature.



Far as the OG question. Did you use to much thermal paste? There is such a thing as using to much thermal paste (hell i've seen people use so much it ended up dripping off the side of the CPU and down onto the socket and latches) Did you spread it correctly? just putting a blob of TMI in the middle then slapping the heatsink on top of that doesn't fly with me.

Should be a layer almost see-through and supremely evenly spread across the heat spreader on top of the processor. Artic Silver recommended the depth of the thermal paste to be 0.13u or 1:16th of a millimeter. So pretty F#@king thin. a blob won't even be close to 13 micron in the middle it'll probably make more of a wedge between the heatsink and the HSA then anything else i'd assume.

I've dropped 20 something degrees before fixing my friends rendition of thermal paste on the top of a heatsink. Done similar results on other peoples computers as well. T

Really if ANY paste is being squeezed out between the middle your using way way to much thermal paste. less is really more with thermal paste. To much and your just creating a block between the CPU and the HSA.

[Fuganater]

or a built in 4 way LCD Thermometer/fan controller. then you'll always know room temperatures. Hell i know the temp of the air coming into my intakes, and my exhaust temperature.

Ok ok.. I'll give you that. Didn't really think of that when I was typing. My 4 channel fan controller tells me the incoming from the front of the case and the out going on the back and 2 top fans.

[mDust]

just putting a blob of TMI in the middle then slapping the heatsink on top of that doesn't fly with me.

a blob won't even be close to 13 micron in the middle it'll probably make more of a wedge between the heatsink and the HSA then anything else i'd assume.

less is really more with thermal paste. To much and your just creating a block between the CPU and the HSA.

The blob in middle is recommended for a lot of chips by Arctic Silver themselves (most AMD chips). A straight line across the cores is recommended for the i5 (http://www.arcticsilver.com/intel_application_method.html). It depends on which product is used and which chip it's used on. There has been no mention of what thermal compound product was used anyway. The blob method prevents any air from becoming trapped underneath and over several heat cycles, and with adequate clamping pressure, will flow out very thin...how thin depends on how much is used. It isn't possible to create a perfectly smooth surface by spreading the compound, which always carries the risk of performance reducing air pockets.

It is recommended that you use the little dab in the center without spreading it technique on CPUs with an IHS.
http://www.techpowerup.com/printarticle.php?id=134

Next, squeeze out a dot of thermal paste directly onto the center of your CPU. Your dot should be about the size of a BB (as in, what BB guns shoot), or a little smaller than a pea. Next, take your cooler and press it straight down onto the CPU so that the thermal paste spreads evenly in all directions.
http://www.maximumpc.com/article/howtos/howto_install_cpu_and_apply_thermal_paste

Proper application is critical to optimum performance. Squeeze onto the center of the CPU an amount of IC Diamond compound about the size of a pea (5.0mm to 5.5mm) on the center of the CPU where most heat is concentrated. Place the heat sink on the CPU and press it down to spread IC Diamond over the CPU's surface, (Do not remove the sink to check the spread as it will introduce air into the joint). Clamp the heat sink and power up the PC.
http://www.innovationcooling.com/applicationinstructions.htm

Less is definitely more though. That is pretty important to understand.

[diluzio91]

@dr walrus
i was worried about the pump, when the new case gets here i'll hook it in to the 12v with a molex to 3pin

also, if it was 35 outside and my i7 was idling at 27... i think i would be winning...

@mdust
Yes, both machines are using aida 64 to measure, confirmed by the bios. that proc is definatly getting the razerblade test... since he decided to let me mod his 6950 into a 6970 i think he might be okay with lapping.

@twigsoffury

i probably have too much paste on there. i'll admit i commited the first sin of modding and used stock paste... my as5 was still in the mail and i wanted to power up...

I do have an nzxt controller with temps, but the probes are shoved out of the way above the dvd drives, so i don't know how accurate they would read the actual room temp.

[dr.walrus]

Its tough to do that unless you have a thermometer in your office.
Well, if you don't, the review becomes relatively meaningless

[Twigsoffury]

ooo you should try Artic Silver Ceramique

http://www.arcticsilver.com/ceramique.htm

I'm sticking to my method of thermal pastin'.... stripes and dabs just sound strange and i'm not sure how that wouldn't do anything but air pockets? but i guess to each his own.

everything in electronics i learned about cooling recitifers always called for a smooth application and not dabs or stripes so i just carried that over to CPU/GPU cooling and sort of honed it down over time.


hell maybe i should make a youtube video about it

[diluzio91]

i think i found the issue....

http://a6.sphotos.ak.fbcdn.net/hphotos-ak-snc6/248239_10150218688848426_601993425_7237688_6917858 _n.jpg

Wheres my damn sandpaper...

[Fuganater]

Thermal paste FTL

[diluzio91]

lol... there was thermal paste on the hsf... its flat ihs FTL.... im gonna have to lap it if i cant get these temps down...

[x88x]

...holy crap... :eek: That's a pretty horrific IHS..

[dr.walrus]

aren't you having the problem on both? Is it the heatsink on the h60 that's the banana?

[diluzio91]

lol.... i replaced it with some cerimaque and temps dropped 30 idle, 50 load... i added a little extra in the "bannana" area, and that seems to have done the trick without lapping.

[Twigsoffury]

lol.... i replaced it with some cerimaque and temps dropped 30 idle, 50 load... i added a little extra in the "bannana" area, and that seems to have done the trick without lapping.

Ceramique takes some odd thermal cycles to set up all the way. can't remember how many think it was a full 24 hours worth. so i bet it drops a few more degrees then that over room temp.


Man that is a horrendous heave on the top of that spreader.

[mDust]

Man that is a horrendous heave on the top of that spreader.

Not that I don't believe Intel sculpts the Himalayas on some of their IHSs, because I know they do, but some HS manufacturers make convex mating surfaces for better clamping pressure. This looks like two convex surfaces that met only in the center. It would run a lot cooler if they were both flat and had minimal thermal paste clogging up the copper heat highway...but sometimes warranties are good to hang on to for a while.

[diluzio91]

sir. you have it backwards. the h60 was razor flat, i checked both, and the ihs is concave. lol

[mDust]

sir. you have it backwards. the h60 was razor flat, i checked both, and the ihs is concave. lol

I would expect there to be a large wad of thermal paste in the center if one was concave and more pushed out to the edges (like it appears) if one was convex. What happened to the paste in the middle?

[diluzio91]

it was stuck to the h60

____________________________________ - h60
----------------------------------------- - paste
*******|____________________|****** - IHS


In this instance i was using the stock pre applied paste so it was already on the hsf.

[mDust]

it was stuck to the h60

____________________________________ - h60
----------------------------------------- - paste
*******|____________________|****** - IHS


In this instance i was using the stock pre applied paste so it was already on the hsf.

I see. I honestly hadn't even considered that.:facepalm:

[dr.walrus]

I have a serious question here: wasn't this affecting both processors?

[diluzio91]

both? My i7 and his i5 you mean? no, my i7 has run frosty cool since i ditched the stock cooler, the ihs on the i7 is pretty flat, not perfect, but not as bad as the i5 was...