Mac Mini 2012 i7 - worth redoing the heatsink compound?

[lightsout]

Hey All,

Using iStat Pro, I can see my CPU idles around 40 degrees C. When I play a game, it can often go over 90 degrees C. This is a bit hotter than I'd like. If anyone else is higher/lower, I'd be interested to see what other people have.

I'm about to put a SSD in it - is it worth putting some arctic silver heatsink compound on it instead of the stock Apple stuff whilst I am taking it apart anyway?

Thanks!

 

[Macman45]

I guess if you're in there, but truthfully? I'd leave it alone..Those temps. are normal.

 

[philipma1957]

I guess if you're in there, but truthfully? I'd leave it alone..Those temps. are normal.

THIS IS CORRECT. ^^^



If you drive the machine hard and it gets to 90C the easiest way to keep it cooler is a side mount stand


http://eshop.macsales.com/item/NewerTech/NUSTALYMINI/

In fact it is on sale for 17.95 and 3 bucks for shipping via the post office total of 20.94 I ordered one 5 minutes ago. I never wanted to pay 29 for it plus 3 to ship total 32 but 21 it is worth it.

 

[lightsout]

I did put some laptop feet on it, so it is slightly higher at the back (figured that'd help the hot air rise to the back a bit). However, those stands do look like a nice option too.

 

[KrisLord]

I'd leave it alone. Whilst installing a drive requires dismantling, it's just a few screws and basically an expensive jigsaw. Removing the heatsinks and cleaning the logic board is quite a bit more risky.

 

[scenemissing]

Depends on how much of a tinkerer you are and how much you value warranties. On a new machine I would lean towards leaving it alone, but I understand the need to mess with stuff. I have an older 13" white plastic Macbook that I'm about to take apart to re-paste this weekend and make it into a server.


Get the Arctic Silver package that comes with the paste removal chemicals, that stuff is magic for cutting through the old toothpaste style crap.

 

[jeffsaha]

I would also suggest leaving it as is, but if you have the tinker bug as I have had with plenty of machines, you could try that and while you are at it lap the heatsink also. I found that to be fun when lapping old copper water blocks for past PC builds, but I like shiny things and that water block had a mirror finish to it in the end ...but, again, I would just leave it as is bc I'm sure popping the heat sink off would not bode well warranty-wise.

 

[chickenwingfly]

What I always say is... yeah.. leave it alone.
The reason being stock hardware is tested by the manufacturer to function normally, so unless you were over clocking (I don't even know if that's possible on Apple's hardware) stock head sink and thermal compound are supposed to work.

You could bring conspiracy theorists here to tell you how it's not and you should change stock stuff since Apple created it so that it would survive up till guarantee period and then it would fail forcing you to buy new hardware. (There you go, I just troll-proofed my post).

 

[randy98mtu]

THIS IS CORRECT. ^^^



If you drive the machine hard and it gets to 90C the easiest way to keep it cooler is a side mount stand


http://eshop.macsales.com/item/NewerTech/NUSTALYMINI/

In fact it is on sale for 17.95 and 3 bucks for shipping via the post office total of 20.94 I ordered one 5 minutes ago. I never wanted to pay 29 for it plus 3 to ship total 32 but 21 it is worth it.

Good to hear. I'm waiting for the Twelve South Backpack to come, which is supposed to have posts and support a mini behind the TBD on its side. I'm tempted to then add a screen as seen in other threads here and an external fan blowing at the bottom of the machine. I just started doing ATV3 Handbrake encodes last night with this machine. I force the fan to about 5000 rpm before I start and the temps get to about 92 and stay there the whole time. If I don't turn the fan up manually, they don't even start to ramp until it hits 100C, then they ramp it so slowly it still reaches 105 before the fan starts going any good. I'm not a huge fan of that. Also once it's reached 105, it's harder to get it back to the low 90's. The normal software seems to let the thing run at 100C, as it will run the fans to about 4200-4600 and the temp will stay in the 98-102 range. I prefer running the fan to 5000 and keeping the temp 10C cooler.

I was tempted to do the thermal paste on my MBP after the first few Lightroom exports I did, but I ended up leaving it alone. I think I will be leaving the mini alone for now as well. Though I did buy the stuff for the MBP so I already have it. lol

 

[JamSandwich]

I'd leave it alone as well. I don't really care for my Mac running in the 90s, but I know odds are it should survive at this temperature, and 90% of what I do is a lot lower-load. I'm not encoding video most of the time, and gaming is maybe 10% of my work... in those cases I use fan control software to boost the fans early, which I find helps at least delay it from hitting peak temps.

Although, I do wonder about the engineering process at Apple. I kinda picture this scenario:

"So you can see, with a decent and properly applied paste, this thermal design falls perfectly into spec. However, when we slather on the goop we're using in our factories in the quantities they use, you see we still need to shave five degrees off our operating temperatures."

 

[SoCalReviews]

I would recommend leaving the stock heat sink compound alone as well. The problem with using replacement CPU heat sink compounds like Arctic Silver is that the metallic silver based formulas can act as a micro electrical conductor and if you don't apply them perfectly (by applying an excessive amount) or if a small amount accidentally gets into a surface mounted chip on the main board then you can risk shorting out the CPU or the main board. When I re-applied new CPU heat sink compound on my PC laptop years ago I used a non-current-conducting ceramic formula as opposed to a metallic sliver based one. Another problem with the replacement heat sink compounds vs. the stock paste is that the replacement compounds are usually thinner and more fluid. Although they can help reduce CPU temperatures after an initial application they also have a tendency to dry out and lose their gap filling ability over time. This means you may need to re-apply them regularly in the future in order to maintain their effectiveness.

 

[eyepea]

I am a bit late on this post but I thought I would add my 2 cents worth. It goes against what most posters say but hey....

I think it is a good idea to replace the thermal past. I would never trust factory compound. I use Arctic Silver 5 (AS5). I does not run. I have used it for a long time on MM's and my MacBook Airs. It has never run for me. See also their website on this comment.

Also it is not as conductive as people say. There are postings on the internet I have seen that say this is a rumour. I have seen some posts on doing ohms testing on AS5 with the result being zero. Right or wrong I don't really know as I am not technically minded in that regard but the AS5 website says it is not conductive. Having said that I would not want it touching anything other than the edge of the contact surfaces anyway.

A point to note on the late 2012 i7 2.6ghz (bought in March 2013) though which I noticed is that there is a very fine gap (super fine) in between the heat sink and the surface of the 'square' processor (I assume GPU), while the 'rectangle' processor (I assume CPU) sits almost flush against the sink. I noticed this when I put AS5 on and had a look with a small led pin light under the sink and noticed I have put too little paste on the square GPU processor. You could see that it had not spread to the edge of the square processor surface. You cannot bend the sink or do anything because there are little stainless steel spacers that keep the sink at the exact level so it does not appear to be a flaw. The spring loaded screws keep the sink sitting on the spacers perfectly. What that means is that you seem to need a little more paste on the GPU square processor to account for the slight space. I put about two rice (short grain) beads of AS5, while I put a thin strip (1.5mm wide) of AS5 on the rectangle CPU. After close inspection with the second go the coverage is perfect. I note AS5 is put on differently (not spread) to some pastes which you spread flat. See the AS5 website for instructions.

Having said all that you can use your choice of paste. All I am saying is have a close look first to see if you have the space like me.

I think the space exist due to simple geometry in that you have about 4 or 5 spacers keeping the sink at an exact level and two processors underneath with the chance of the sink sitting exactly on all 6-7 points being minimal in chance.

Also I use the exact stand philipma is referring to above. It works wonders and takes up less space (and keeps the bottom black cover scratch free.

Having said all that, the OP's temps are fine. I had the last mm 2011 2.0ghz quad and it ran those higher temps when I pushed it with say handbrake (which I use for a few days straight many times). It is still sweet two years on.

Watch this space... Sorry bad attempt at a joke.

IP

 

[benwiggy]

I think it is a good idea to replace the thermal past. I would never trust factory compound.

What, really? EVERY computer assembled in a factory uses sub-standard paste that causes a risk of overheating unless you fix it yourself with your "better" paste?

Having said all that, the OP's temps are fine.

And there we have the most important fact.

 

[eyepea]

What, really? EVERY computer assembled in a factory uses sub-standard paste that causes a risk of overheating unless you fix it yourself with your "better" paste?


And there we have the most important fact.

I don't recall to many posts claiming that Apple uses top quality thermal paste? On the contrary? I am not sure why claiming factory thermal paste is so inappropriate? Have I stumbled into an Apple Thermal Paste suppliers thread or something? Anyway, each to their own. I just prefer to tinker my my MM and upgrade what I can. The temp difference is only about 5'c on average for my Macs so it is certainly not worth the effort if you don't like rebuilding Macs. If replacing the wifi card with titanium screws gave me 20% more bandwidth then I would probably do that too. Anyway, stay cool.

IP

 

[benwiggy]

I don't recall to many posts claiming that Apple uses top quality thermal paste? On the contrary? I am not sure why claiming factory thermal paste is so inappropriate? Have I stumbled into an Apple Thermal Paste suppliers thread or something?

There's a world of difference between saying "there is thermal paste with better characteristics than the factory uses" and "I would never trust factory paste", which is quite a bold statement, and suggests that Macs are coming out of the factory in a sub-standard or defective condition.

Frankly, I can't believe that there are many discussions about the qualities of paste, either way. I certainly don't intend to have many.

Is the factory state of a Mini good enough for 99% of uses? Yes.
Have you improved the cooling by 5° ? Quite probably.
Has this actually achieved anything in itself? Probably not.

20% more bandwidth is useful. A slight reduction of temperature that is within acceptable range either way: not so much. If you're really worried about heat, you can probably get the same reduction by mucking about with the fans.

 

[SR45]

Anyone have a heat problem on their Mac's ? I never have.

 

[LaunchpadBS]

+1 for SMC fan control

 

[niteflyr]

Does anybody have any before and after test data to add to this debate? How much cooler does Arctic run than factory paste? Is there really any substantial difference?

 

[philipma1957]

Does anybody have any before and after test data to add to this debate? How much cooler does Arctic run than factory paste? Is there really any substantial difference?

0 to 5c difference / yeah why the swing some times you get a good factory heat sink and a good cpu chip. when this happens any old paste will work . if you got a wavy heat sink or a wavy cpu chip better paste would be good to use.

It is easy to tell how well the factory mating was done. Just get a temp program like stat nano


http://www.apple.com/downloads/dashboard/status/istatnano.html

free

Then clock your temps while pushing the mini. If you don't get over 100c no matter how hard you try don't bother doing anything at all.

If you get over 100c you can get istats menu 4 which has really good fan options. you can try some external cooling options. you can buy applecare say f it the mini will break and if I am lucky it breaks in 32 months. you can change the paste.

 

[eyepea]

Does anybody have any before and after test data to add to this debate? How much cooler does Arctic run than factory paste? Is there really any substantial difference?

I don't have image shots of the data so my word is only as good as the 1 cent worth of Internet to send this post but...

I have put AS5 on 3 mac minis (2011 i5, 2011 quad and the 2012 i7 2.6) and 2 MacBook airs (2010 i5 and 2011 i5).

This is what I found:

The MBAs run a lot cooler at idle. Approx 10'c but I think the fact that I cleaned the heat sink channels (super small on the MBA) help the decrease in temp. MBAs run about 3-5' c in moderate to heavy use but that is on average. I don't think the difference is anything special though. The main diff is the idle temp. They both used to idle in the late 40s but now I see them in the 30s. Note sure how much the cleaning contributes but after a year the temps are still down.

On my my MMs it is similar but I noticed the temp in super heavy use does not change much when you use AS5. I have two i7 quad cores with turbo boost 2.0. I think they have something to do with it in that the TB only kicks in if the temp is ok, so even if AS5 cools it down say 5'c then the TB will push harder as it is within limits, thus keeping the temp at its limit. That said you should get more processing power for the temp as well. I have now way to benchmark the turbo boost (nor can I do a before or after now) so it is just a theory so please rebutt and rebuke away as much as you like

In a nutshell, with AS5 I find a substantial difference to idle and low end temps on all my macs. I find minimal difference at peak load with say handbrake. Maybe 5-3c on the 'non quad cores). But I find both my quads always hover around 90-98c on peak processing (running HD coding in handbrake while running Parallels/win8 and various program's in OSX. By 'always' I mean before and after using AS5.

I think the turbo boost will always keep your CPU running at peak temps anyway,but very hard to tell if you get more or less TB?

Rebutt/rebuke away I am up for it

 

[MJL]

Replacing the thermal paste resulted in less variation in temperature and a slightly lower idle temperature. I prefered to use Tuniq 4 since the heatsink surface was not totally flat (seen after cleaning) and Tuniq 4 is not critical with a wide gap. Some reckon it is better than Artic silver (which I used on my Thinkpads and prefer when the heatsink is a good fit).

Tuniq 4 is also pretty close in color to the Apple thermal paste color and it is not electrically conductive, both advantages imho.

However if your temprature in general stays below 65C then there is nothing to worry about - increasing the idle speed of the fan has a decent impact on keeping the temperature down and negates the need for changing the thermal compound. I did change it with the mid 2010 model but not with the base 2011 model or the 2010 server.

 

[SoCalReviews]

I don't like what I am reading about these peak Mini CPU temps...90-98C??? I hope my late-2012 i7 2.3Ghz CPU temp never gets anywhere near that high. I almost never hear the fan spin up on my base mid-2011 (only a few times...while installing Windows 8 Pro as a VM, one time gaming in Windows 8 and while upgrading to ML) so that means it probably stays on average below 47C right?

I get worried when my old Win XP machine with AMD CPU goes over 128F/54C...which is about as hot as I ever see it get. It's got a large copper CPU heatsink (AS-5 paste I replaced every 2-3 yrs.) with a 3400 RPM fan that's been running 24/7 for almost 12 years without having to be replaced... which should run for another 12 years 24/7 if I keep the dust cleaned off of it. I have the BIOS set to shut that machine down if it goes over 142F/61C. I like my old desktop tower PC CPUs to idle between 107F/42C and 114F/45C -> which is what it is at right now as I type this.

 

[eyepea]

I don't like what I am reading about these peak Mini temps...90-98C??? I hope my late-2012 i7 2.3Ghz CPU temp never gets anywhere near that high. I almost never hear the fan spin up on my base mid-2011 (only a few times...while installing Windows 8, one time gaming in Windows 8 and while upgrading to ML) so that means it probably stays on average below 47C right?

I get worried when my old Win XP machine with AMD CPU goes over 128F/54C...which is about as hot as I ever see it get. It's got a large copper CPU heatsink with a 3400RPM fan that's been running 24/7 for almost 12 years without having to be replaced. It should run for another 12 years 24/7 if I keep the dust cleaned off of it. I like that desktop tower PC's to CPU idle between 107F/42C and 114F/45C -> which is what it is at right now as I type this.

I have 5 Macs that all run around 90 when the CPU is given a task of heavy duty processing. Handbrake is a very common app that will push your CPU 100%. If I use my macs for web and word processing then they run at your preferred temps. It depends on the CPU load, and not necessarily on how many apps you have open.

If you like 45c then this comment will make you feel faint... I have coded video on a MM 2ghz Quad core server 2011 for a few years. Sometimes I run it for days at peak CPU load to code hours of HD video. Yes... 90C for about 100 hours straight . The MM Server is now retired in the back paddock were it gets to laze about in the green pastures of my wife browsing the Internet. That is the extent of its CPU load now. Sad really. It was once a thoroughbred and now it is just treated like a pony.

If you are a religious person you may wish to pray for my new 2012 i7 2.6ghz.

 

[SoCalReviews]

If you are a religious person you may wish to pray for my new 2012 i7 2.6ghz.

I will see what I can do. I hope your new 2012 i7 2.6Ghz Mac is Christian. If not I have some Buddhist meditation techniques I can try. Since we can assume it was born in China we might be able to get it started off with the right Karma and covert it later.

 

[valdigre]

Could I ask those of you who reapplied thermal compounds on their 2012 i7 minis to report their load temperatures (Boinc, handbrakebetc)?

I put Prolinatech pk-2 to my 2.6 i7, I have around 40-44 Celsius when idle (room 20 degrees) but under load, handbrake or Boinc, I get to around 98-105 !!!

I thought changing paste would lower load temp but either it does not or I applied too little paste?