Summer CPU overheating and Zinc Oxide "white stuff"

[eccerr0r]

Okay, it's almost the end of the summer already, so this is kind of dumb. My i7 Socket1155 (GOD I HATE THESE "SINGLE WIPE" SOCKETS!) was overheating as it was getting about and over 93°C (die temperature). As my m/b beeps when this happens I quickly try to stop whatever I'm doing and a few moments it drops back down to about 10°C over ambient. It got so bad that I get beeps running 3 of the 4 cores.

This is the stock low profile copper core, aluminum fin heatsink.

I don't know what the original grayish heat sink compound was (it was "pre applied") but it was flaking all over the place when I removed the heatsink/fan. What a mess, it was crumbling all over the place. I air compressor blew out the dust from the heatsink (what a mess 2) and tried to put it back together...

As I don't have any other other heatsink compound other than some zinc oxide, I had to use it. On goes the white stuff and put back together.

Well at least I got my temperatures down to 84°C while using all 4 cores at the same time once more (where it got to 93°C using 3 of 4 cores), so I guess that's good.

I thought it was my case finally catching up. The case I have is a coolermaster tower/desktop case with PSU in front. So in tower configuration, the PSU is sucking air in the wrong direction (though it does have a large 12mm fan sucking in the "right" direction.) I wonder if it's worth to turn that PSU fan around...

[kaboom]

Re: Summer CPU overheating and Zinc Oxide "white stuff"

Make sure "Q-fan" or any similar quiet fan control is disabled in BIOS.

What's the airflow situation? Try two 120mm fans, one near the top/rear of the system above/behind the MB, another in the front for makeup. This is a minimum, you also need makeup air for the power supply.

Be sure to use wire fan grilles for the two 120mm units- backpressure from stamped vents leverages the fans' static pressure capabilities instead of airflow. We want the latter here, and the easiest way to get it without lots of noise is to run quieter fans at minimal backpressure.

Arctic F12s would work well.

Not familiar with that case, but just giving general advice...

Make sure the bottom of the heatsink is flat too! Can't speak for intel, but some of the AMD 'sinks are bowed to the point they'll overheat something like an FX-8350- with auto fan control not just disabled, but the fan's thermistor actually bypassed!

[eccerr0r]

Re: Summer CPU overheating and Zinc Oxide "white stuff"

I think the cooling situation is OK for now. It's not great but my Core2Quad is cooler as its die temperature seems to be only a few degrees Celsius over ambient. Not sure if the Core-i7 should be the same or not, alas they both run different software (both Linux). I just wonder how much of an improvement a metal TIC would be compared to the ceramic.

But anyway, that's what's messed up with the coolermaster, I don't know why they designed it such that the PSU is up in front, probably to reduce the overall profile (as a desktop) or reduce height (as a tower) or something, as the power supply jutting up is a limit to the height of how small you can make the full sized ATX case. Anyway it's sitting in front with the fan blowing air outwards (ATX 2.03+ spec?) to the front/bottom when it's sitting upright... which is wrong since hot air rises.

If I reverse the flow of that fan, it would suck cool air inwards and up towards the top exhaust which seems more natural instead of what it is currently having both sucking air out. I suppose in a desktop flat configuration it's not that bad, but as a tower this seems wrong.

The heat producing objects are mostly the GPU (RadeonHD) and the CPU of course. When using the case as a desktop, the PSU exhaust is warmer than the 120mm fan exhaust but both aren't that bad. As a tower, both exhausts are warm...

[stj]

Re: Summer CPU overheating and Zinc Oxide "white stuff"

it would help if you have a thermocouple probe for your multimeter,
then you could compare the bios reading to the actual heatsink temperature to try to judge how well the heat is transfering.

[jiroy]

Re: Summer CPU overheating and Zinc Oxide "white stuff"

I was doing similar researches to solve the heating problem of an HP DV6. I Found this tested on some site (I forgot , lol ) :


Bad Thermal Compound
Tital Nano Blue 59
Panasonic 58.5
KPT-8 (reference) 56.6

Decent Thermal Compound
Data Cooler 57
Titan TTG S-104 , S-103 56.7
Pasta Siliconowa 56.6
Zalman CSL 850 56.5
Noctua 56.5
Stars Silver 56.5
Stars 700/Aero 700 56
GeIL GL-TCP1b 55.6
Thermopox 55.6
(BeO) 55.51
Sil more 55.5
Shin-Etsu(white) 55.5
W.P. 55.5
STARS (white, soft pack) 55.5
AKT-842 55
Fanner 420 55
Koolance 55

Good Thermal Compound
Arctic Alumina 55.5
Arctic Silver-3 54.6
AOS 54.5
DC-340 54.5
Asetek 54.5
Arctic Silver-5 53.5
Arctic Ceramique 53.5

Great Thermal Compound
Apus TMG 301 52.5
Gigabyte 52.5
Titan Nano Grease TTG-G30010 52.5
GFC-M1 D90T8-010 52.5
Shin-Etsu MicroSi G-751 52

Best Thermal Compound
Arctic Cooling MX-1 51.5
Shin-Etsu MicroSi MPU-3.7 50.5
Coollaboratory Liquid Pro 50

[diif]

Re: Summer CPU overheating and Zinc Oxide "white stuff"

That list is from at least 2007 and there is no MX-1 any more.

[jiroy]

Re: Summer CPU overheating and Zinc Oxide "white stuff"

That list is from at least 2007 and there is no MX-1 any more.

Yes , it is old , but what's interesting is the bad ranking of Arctic Silver 5 , lol . I'm definitely gonna seek this Coollaboratory Liquid Pro 50 .

[diif]

Re: Summer CPU overheating and Zinc Oxide "white stuff"

Yes , it is old , but what's interesting is the bad ranking of Arctic Silver 5 , lol . I'm definitely gonna seek this Coollaboratory Liquid Pro 50 .

Don't get it near anything aluminium.

[eccerr0r]

Re: Summer CPU overheating and Zinc Oxide "white stuff"

I suppose the "white" stuff is zinc oxide like what I have, I wonder what others are zinc oxide... I suppose it only needs to be "good enough".

Why do they have heat sink compounds that are bad with aluminum, a lot of heatsinks are aluminum?

[diif]

Re: Summer CPU overheating and Zinc Oxide "white stuff"

I suppose the "white" stuff is zinc oxide like what I have, I wonder what others are zinc oxide... I suppose it only needs to be "good enough".

Why do they have heat sink compounds that are bad with aluminum, a lot of heatsinks are aluminum?

Cheap ones are but the people concerned with heat aren't going to using aluminium anyway. The more expensive heatsinks are copper or nickel coated copper as they have better thermal conductivity.
It might also have industrial applications using other metals.

[jiroy]

Re: Summer CPU overheating and Zinc Oxide "white stuff"

Don't get it near anything aluminium.

Very true .. A good advice here is to read the specifications and manuals of any material prior to use it . Neglecting basics can lead to disastrous things .

[jiroy]

Re: Summer CPU overheating and Zinc Oxide "white stuff"

I suppose the "white" stuff is zinc oxide like what I have, I wonder what others are zinc oxide... I suppose it only needs to be "good enough".

Why do they have heat sink compounds that are bad with aluminum, a lot of heatsinks are aluminum?

I always kept aside , for personal use , Copper heatsinks , mostly from old Dell Optiplex , they're better by far than Aluminium ones .
Most of the white compounds are ceramic base , cheap and good , but not the best .

[momaka]

Re: Summer CPU overheating and Zinc Oxide "white stuff"

it would help if you have a thermocouple probe for your multimeter,
then you could compare the bios reading to the actual heatsink temperature to try to judge how well the heat is transfering.

I've tried that before, and it was a worthless experiment because:

#1: The airflow of the fan will offset any thermal conduction to the probe.... so unless the probe is stuck between the heatsink and the CPU, you won't get accurate readings

#2: Even if you manage to avoid #1 above, you will still read much lower temperature, again due to ambient air making the probe cooler. So on average, you may see 5 to 10C lower temperatures with your probe than what they really are.

#3: The thermal performance of "good" and "bad" thermal compounds on low/medium TDP CPUs with heatspreaders is only about 3-5C apart at most, if even that (if applied properly, of course).

So it's worthless doing any thermocouple readings.

That said, the biggest problem with the O/Ps rig is the stock heatsink - those low-profile Intel heatsinks are completely inadequate for cooling an i5/i7 and maybe just enough for a Celeron and i3. At work, I saw a few dead 6th and 7th gen i7's - all with those thin coolers. I presume most likely because they overheated, as a few of them had the same flaking thermal compound like in your case.

Once the CPU overheats, that essentially begins to ruin the thermal compound between the CPU core(s) and the metal heat spreader as well. With Intel using non-soldered heat spreaders for their 6th and 7th gen chips, they really are asking for trouble. If that's what you have, then it's time to delid the chip and change the thermal compound. Better yet, use thermal epoxy for the heat spreader. Once cured, thermal epoxy actually has better heat conduction than most thermal compounds.

I think the cooling situation is OK for now. It's not great but my Core2Quad is cooler as its die temperature seems to be only a few degrees Celsius over ambient.

Unless the CPU is dissipating less than 10W in idle, its idle temps should not report just a few degrees above ambient. Most likely the CPU is under-reporting the CPU temperature then. Socket AM2 and newer CPUs (but not the new AM4/Ryzen) are even worse in that regard, showing lower than ambient temperatures - which is impossible, unless your heatsink is a Peltier cooler, lol.

That said, if the CPU temperatures even with the new thermal compound are still instantly hitting over 75C under load but going to a few degrees above ambient in idle, then you most definitely have bad thermal coupling between the CPU core and CPU heat spreader.

[Th3_uN1Qu3]

Re: Summer CPU overheating and Zinc Oxide "white stuff"

MX-2 and MX-4 is all i use. Cheap white goop only works for things running under 50-60C, otherwise you'll be replacing it every few months.

Either way, for that CPU you should be considering a heatsink upgrade.