Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Swampus

macrumors 6502
Jun 20, 2013
396
1
Winterfell
I have some Arctic Cooling MX-4 in the mail and will apply and repeat the tests and post the results. AC MX-4 is suppose to be better than Arctic Silver 5, so I will find out.

I think MX-4 is a lot more versatile. It gives solid performance over a wide range of conditions. But if you're taking the extra steps that OP has taken to ensure that well-finished surfaces are meeting perfectly, it's still hard to beat AS-5. That said, there shouldn't much difference between compounds under those conditions. It's when dealing with various imperfections that some compounds will stand out more than others.

Shin Etsu grease is actually not that bad compared to the fancier greases you can purchase today.

You know, I really do agree with you here. Shin Etsu makes some pretty nice stuff. The syringe that you got from Apple is one of their X23 products, right? Do you know which one? X23-7783? X23-7762? X23-7921?

Is the stuff that you get in a syringe from Apple the same stuff that is used at the factory? Or is the factory stuff a cheaper version? Or is the factory stuff Shin Etsu at all? Does anyone know? It's the same color, but the texture seems odd to me.
 

ouimetnick

macrumors 68040
Aug 28, 2008
3,552
6,341
Beverly, Massachusetts
Apple uses the same thermal grease at the factory as they give with repairs (well they looks and feel the same) Anyhow Apple uses Shin Etsu X23-7783D paste.

Attached is a photo showing the syringe. Usually it is repackaged in the translucent syringe with a blue cap, but sometimes it comes in a white labeled syringe. The blue syringes have more in them because I had about 15 blue syringes and decided to put them all in two syringes instead of having a cluster mess of them. The white labeled one arrived after I made two full syringes.

I will have the results for my Arctic Cooling MX-4 tonight. Seems to be worse than the X23-7783D though. I applied it the same as I applied the Shin Etsu (just a dot and let the heatsink spread it out) I also attached photos to show how little Shin Etsu was applied to the CPU and chipset die. I applied that. You can see just how little is needed to get the best results. I also cleaned up the heatsink and polished it to a beautiful mirror shine BEFORE applying Arctic Cooling MX-4. So the AC MX-4 has two advantages, a smoother heatsink surface, and supposedly being a better compound than the X23. Sadly the results point to it being WORSE than X23, and it sucks considering I bought a 20g syringe of AC MX-4.

Maybe it works better with CPUs with an integrated heat spreader?
The first four photos are for the AC MX-4, and the last photo is just to show the syringe that Apple uses.
 

Attachments

  • photo 1.JPG
    photo 1.JPG
    2.1 MB · Views: 301
  • photo 2.JPG
    photo 2.JPG
    2 MB · Views: 319
  • photo 3.JPG
    photo 3.JPG
    2.1 MB · Views: 289
  • photo 4.JPG
    photo 4.JPG
    1.8 MB · Views: 329
  • Shin Etsu X23.JPG
    Shin Etsu X23.JPG
    1.5 MB · Views: 631

Brian Y

macrumors 68040
Oct 21, 2012
3,776
1,064
Apple uses the same thermal grease at the factory as they give with repairs (well they looks and feel the same) Anyhow Apple uses Shin Etsu X23-7783D paste.

Attached is a photo showing the syringe. Usually it is repackaged in the translucent syringe with a blue cap, but sometimes it comes in a white labeled syringe. The blue syringes have more in them because I had about 15 blue syringes and decided to put them all in two syringes instead of having a cluster mess of them. The white labeled one arrived after I made two full syringes.

I will have the results for my Arctic Cooling MX-4 tonight. Seems to be worse than the X23-7783D though. I applied it the same as I applied the Shin Etsu (just a dot and let the heatsink spread it out) I also attached photos to show how little Shin Etsu was applied to the CPU and chipset die. I applied that. You can see just how little is needed to get the best results. I also cleaned up the heatsink and polished it to a beautiful mirror shine BEFORE applying Arctic Cooling MX-4. So the AC MX-4 has two advantages, a smoother heatsink surface, and supposedly being a better compound than the X23. Sadly the results point to it being WORSE than X23, and it sucks considering I bought a 20g syringe of AC MX-4.

Maybe it works better with CPUs with an integrated heat spreader?
The first four photos are for the AC MX-4, and the last photo is just to show the syringe that Apple uses.

The problem isn't the thermal paste - it's how it's applied. I've seen Mac's come from the factory with horrifically thick thermal paste applications, and I've seen some with hardly any.

Also - there's hardly anything (if anything at all) in the ACMT training about how to apply thermal paste. Yes, you'd hope they'd have done it before, but there's always the possibility that they hadn't.
 

ouimetnick

macrumors 68040
Aug 28, 2008
3,552
6,341
Beverly, Massachusetts
The problem isn't the thermal paste - it's how it's applied. I've seen Mac's come from the factory with horrifically thick thermal paste applications, and I've seen some with hardly any.

Also - there's hardly anything (if anything at all) in the ACMT training about how to apply thermal paste. Yes, you'd hope they'd have done it before, but there's always the possibility that they hadn't.

True, but does it really matter how much you put on? Regardless of the amount (back in the day, you used a half of a syringe per chip) it all gets flattened when the heatsink is put on. The excess just goes over the green fiberglass substrate. The more you have, the more that flows over the die, but the grease is still flattened appropriately by the heatsink.

macbook-thermal-grease.jpg
 

Brian Y

macrumors 68040
Oct 21, 2012
3,776
1,064
True, but does it really matter how much you put on? Regardless of the amount (back in the day, you used a half of a syringe per chip) it all gets flattened when the heatsink is put on. The excess just goes over the green fiberglass substrate. The more you have, the more that flows over the die, but the grease is still flattened appropriately by the heatsink.

Image

The thickness isn't just about the amount - it's also about how tightly (or close to the chip) the heatsink is bolted down. If it's not tightened enough at the factory, you're going to end up with quite a wedge of paste.
 

ouimetnick

macrumors 68040
Aug 28, 2008
3,552
6,341
Beverly, Massachusetts
They bolt them down well in my opinion. My mom's computer with the stock thermal paste gets the same readings my computer does when it has the Shin Etsu. (we have identical computers)
 

Swampus

macrumors 6502
Jun 20, 2013
396
1
Winterfell
Apple uses the same thermal grease at the factory as they give with repairs (well they looks and feel the same) Anyhow Apple uses Shin Etsu X23-7783D paste.

Wow. Now, I am seriously confused. I've had my own suspicions about the heatsink as culprit in some of the outliers folks have reported. In threads where pictures were posted, the images (in my mind, under those assumptions) were consistent with the narrative. But that all assumes a poor quality compound.

How do we mathematically account for these variations against X23-7783D? It is practically immune to mount pressure, right? Its thermal resistance relative to thickness is also very impressive. I'm not talking about OP or others who have lapped. I'm talking about those of us who have improved the situation by 5 to 10 degrees at load by simply redoing the compound. How is this mathematically possible compared to X23-7783D?
 

Swampus

macrumors 6502
Jun 20, 2013
396
1
Winterfell
By the way, X23-7783D is one of the compounds studied by the United States Department of Energy (download pdf), so there is actually some well-funded research here. I'm not just basing these opinions on some overclocking website where things are tested under ideal conditions.

Immune to mount pressure:

attachment.php


Though not ideal for any compound, X23-7783D holds its own better than most as thickness increases:

attachment.php
 

Attachments

  • pressure.png
    pressure.png
    24.9 KB · Views: 1,663
  • thickness.png
    thickness.png
    43 KB · Views: 1,581
  • thickness100.png
    thickness100.png
    31.2 KB · Views: 220

ouimetnick

macrumors 68040
Aug 28, 2008
3,552
6,341
Beverly, Massachusetts
Well here are the results. Like I mentioned I applied Arctic Cooling MX-4 and when I got higher numbers, I took it back apart to make sure I had the proper amount of grease on the chips. I did, so I cleaned the chips and heatsink off and reapplied. No difference. The results speak for themselves.

I ran two terminal windows with
yes > /dev/null

after 2 minutes
201ºF/93ºC
Fans: 3563 rpm

after 10 minutes
180ºF/82ºC
Fans: 6192 rpm

This is what you get with Arctic Cooling's MX-4 grease. Other tests show this as an amazing grease, so either I got a bad tube of it, or it isn't amazing compared to Shin Etsu's X23-7783D.
 
Last edited:

ouimetnick

macrumors 68040
Aug 28, 2008
3,552
6,341
Beverly, Massachusetts
My mother purchased her MBP in April 2010. Never been opened, so no thermal paste changes or even dust cleaning. Still has the wimpy 250GB drive inside.

Idle
99ºF/37ºC
Fans: 1996 rpm
_____

Web Browsing, no flash
113ºF/45ºC
Fans: 1999 rpm
_____

yes > /dev/null in two terminal windows (100% CPU)

after 2 minutes
189ºF/87ºC
Fans: 3367 rpm

after 10 minutes
184ºF/84ºC
Fans: 3979 rpm

Who knows how much is applied on the chips inside? It could be the right amount, or a CPU sandwich, but regardless, the results are still better than AC MX-4.
 
Last edited:

Swampus

macrumors 6502
Jun 20, 2013
396
1
Winterfell
This is what you get with Arctic Cooling's MX-4 grease. Other tests show this as an amazing grease, so either I got a bad tube of it, or it isn't amazing compared to Shin Etsu's X23-7783D.

I don't know. I think X23-7783D is pretty good stuff and very well suited to attenuate for the range of manufacturing imperfections normally associated with this type of dual heatsink assembly. I'm still struggling with the notion that this is what is used at the factory.

Sorry, but I got nothing but confusion at the moment.

If you don't mind my asking, though, what was your motivation for doing this? 80ºC after 100% CPU for 20 minutes is pretty good, no?
 
Last edited:

wordoflife

macrumors 604
Jul 6, 2009
7,564
37
Very interesting thread. Enjoyed reading many of the findings especially those by Doward and iMacC2D. Though I haven't gotten through the whole thread yet.

Overall, I agree with that Apple laptops run pretty hot. While I don't have any figures for how much performance decreases when the temperature maxes out, it's not unusual to happen. Most of the time though, even when it's not maxed out, I find that the computer is hot enough to make me feel uncomfortable using it. The $400 HP laptop that I have never becomes hot enough where the question of being uncomfortable comes up.
 

ouimetnick

macrumors 68040
Aug 28, 2008
3,552
6,341
Beverly, Massachusetts
I don't know. I think X23-7783D is pretty good stuff and very well suited to attenuate for the range of manufacturing imperfections normally associated with this type of dual heatsink assembly. I'm still struggling with the notion that this is what is used at the factory.

Sorry, but I got nothing but confusion at the moment.

If you don't mind my asking, though, what was your motivation for doing this? 80ºC after 100% CPU for 20 minutes is pretty good, no?

80ºC for a CPU under full load is great. Problem is that Arctic Cooling MX-4 doesn't get you that low, but Shin Esu X23-7783D does.
 
Last edited:

rev.b

macrumors regular
May 1, 2009
232
0
Portugal
Mine is running hot, but not too hot, as it can hold turbo boost @ 3.3 Ghz indefinitly running 8 instances of glucas (Mac CPUtest). So it's running well within spec.

Late 2013 15" rMBP, 2.3 Ghz Haswell.

Could it be better? I guess so, but I think that obcession with lapping, exotic thermal paste, and so on, makes sense when discussing overclocked CPU's, that can be severely damaged by high temps. Definitly not the case with a Apple laptop.

If it's running within specs, it's OK.

My previous early '11 17" 2.2 Ghz Sandybridge could handle 2.4 Ghz under that load, even at lower temps. I guess that wasn't from poor cooling, but from vastly improved thermal and turbo boost management with Haswell.
 

Sadisterr

macrumors member
Sep 1, 2013
43
1
My mother purchased her MBP in April 2010. Never been opened, so no thermal paste changes or even dust cleaning. Still has the wimpy 250GB drive inside.

Idle
99ºF/37ºC
Fans: 1996 rpm
_____

Web Browsing, no flash
113ºF/45ºC
Fans: 1999 rpm
_____

yes > /dev/null in two terminal windows (100% CPU)

after 2 minutes
189ºF/87ºC
Fans: 3367 rpm

after 10 minutes
184ºF/84ºC
Fans: 3979 rpm

Who knows how much is applied on the chips inside? It could be the right amount, or a CPU sandwich, but regardless, the results are still better than AC MX-4.


This is actually quite interesting, I use the same paste: AC MX-4 which is supposed to be great and I have read outstanding reviews everywhere on the internet. My results are bad and people are saying that MX-4 is better than outdated AS-5. I may try to use AS-5 next time.
 

sixrom

macrumors 6502a
Nov 13, 2013
709
1
I've owned nearly every new MBP & retina equipped MBP Apple has released. Not one has failed because they run so hot. It's well known that Apple values style over function. Ultra_Thin and made from aluminum makes for a pretty computer. Shiny and skinny sells, a fact Apple capitalizes on. Heat on your lap is the proof it's genuine Apple :)
 

Swampus

macrumors 6502
Jun 20, 2013
396
1
Winterfell
This is actually quite interesting, I use the same paste: AC MX-4 which is supposed to be great and I have read outstanding reviews everywhere on the internet. My results are bad and people are saying that MX-4 is better than outdated AS-5. I may try to use AS-5 next time.

You started your own thread a few months ago didn't you? Heatsink didn't fit flat against the CPU? Wide variation from core to core? You mentioned that you were thinking of ordering a new heatsink? Did you ever get around to doing that?

----------

I've owned nearly every new MBP & retina equipped MBP Apple has released. Not one has failed because they run so hot. It's well known that Apple values style over function. Ultra_Thin and made from aluminum makes for a pretty computer. Shiny and skinny sells, a fact Apple capitalizes on. Heat on your lap is the proof it's genuine Apple :)

True, but some run a little hotter than others. For someone who has one of these and still needs it for CPU intensive tasks, investing a little time to make it run cooler can be very worthwhile. In some cases, doing so, for all intents and purposes, is performing a CPU upgrade.
 
Last edited:

Gav Mack

macrumors 68020
Jun 15, 2008
2,193
22
Sagittarius A*
Giving this a bump - I've just got myself a 17 inch late 2011 to replace my 2010 and about to give it the AS-5/tint/lap treatment to drop the idle temps down from close to 70 degrees c for the CPU/GPU. Too warm for my liking!

Anyone think that these high temperatures with the 2011 generation models could have a link to 'radeongate'; which I suspect could be connected to the banning of leaded solder and issues with manufacturers finding decent lead free solutions over the exact period these 2011 models were built. Excessive heat, expansion and contraction of solder joints, weak joints due to problems with the solder causing many GPU's to fail, many outside the warranty period. Throw excessive thermal paste applications plus rough surfaces on heatpipes into the mix..

EDIT: Post treatment I've got the CPU die on bootcamp down from 68 degrees C down to 49 degrees, it's idling in OSX at the 39/40 degree mark from 57 with the same tabs I had open before I started in Firefox, and AS-5 should get better after a couple of hundred hours. A small amount of fluff in both the heatpipe grilles, hair and dust removed from the logic board and inside the case. Polished the contact plates nice and shiny, they were very dull beforehand after cleaning. Feel much more comfortable now with the improved thermal performance and lower default temps for my new mobile work tool. For these Sandy Bridge models doing nothing is not an option if the temps are high, I have had too many reballs, a couple of successful logic board baking revivals to do otherwise!

The unibody cooling design until the retina came along is roughly the same, but the transistor count on both the sandy bridge plus 6770 is rather more than a core2duo plus nvidia 9600 for pretty much the same kind of heatpipe layout to dissipate all that extra thermal energy. Hence why through having to repair and reball these units and now being the owner of a 2011 getting the thermal transfer as optimal as possible will improve the chances of your logic board lasting as long as possible!
 
Last edited:

barnettgs

macrumors member
Dec 12, 2006
93
0
Northern Ireland
Application method

It has been few months since my last post (re-pasting): https://forums.macrumors.com/showthread.php?p=18373119#259

I thought it has been improved but by spending time with it every day, it was still running hot and has been more or the less the same before re-applying paste a few times last November. I thought it had improved but I was not convinced of my results though.

But last night, I have found out that the 'spread' method isn't suitable for Macbook Pros because their heatsink is not level and also poor machined.

Spread method would be ideal if the heatsink and CPU are perfectly aligned to each other with good machining but it doesn't seem that way with my MBP. My MBP heatsink for CPU is slightly 'dented' therefore using the 'spread' method was just a bad idea or almost useless.

So I decided to go the 'blob' route, by applying a 'rice grain' method with Active Silver 5 paste this morning. Which meant ignoring Active Silver's recommendation of using 'spread' method.

The temperatures are much better now and lower than they were used to be with spead method of Active Silver 5 or factory paste. As I am typing on this Firefox browser with 6 tabs opened, CPU is around 40-50 C! It was used to be around 50-60 C or over. It doesn't feel 'toasty' any more while browsing the internet.

So it seems that 'blob' or 'dot' method works much better with MBPs.
 
Last edited:

lewdvig

macrumors 65816
Jan 1, 2002
1,416
75
South Pole
As I am reading this thread, on battery, 3 tabs open in chrome with nothing else running my temp is 68 Celsius and climbing.

I might try one more Apple notebook after this 2012 MBP 15" due to my love of OSX. But I am getting tired of this. From what I have read, the 13" rMBP runs very cool unless playing games.
 

MH01

Suspended
Feb 11, 2008
12,107
9,297
Well, considering that millions upon millions of Mac notebooks are running without failure for many, many years, Apple is right. I'm posting this from a 5 year old MBP that I've hammered extensively for 14+ hours a day, almost every day, with no reapplying thermal paste, no fan control apps, no cooling pads, nothing beyond just using it. It still runs like it did day one. The vast majority of Mac users can report similar experiences. If temps were a significant problem, there would be mass recalls, floods of news articles, etc.

There is a small minority who freak out about Mac temps, but it's largely unfounded, as most Macs just keep on running. The OP has made these arguments in several threads before starting another redundant thread on this beaten-to-death topic. The fact is that for most Mac users, temps are perfectly fine.

When you say hammering 14 hours a day, what are you actually doing on it?

Reason I ask, I used to game on MacBook pros, giving them a hammering and I killed a few over time. It's not a sudden death, but heat kills electrical devices.

I agree that for most users mac cooling is fine. Though people who are going to hammer their machines, like extensive gaming, should be informed that heat does kill electrical devices, and macs can get very hot.
 

jayrammac

macrumors newbie
May 9, 2014
26
0
Very interesting thread. As a brand new convert to Mac, this is the first time I have really engaged in the various forums. For a long time I thought that the reviews you see all over the internet were totally biased towards Apple. Heat, battery life, screen quality are all called out on other laptops as issues, but I can not once remember seeing anything of the sort in relation to Macbooks, yet the more I read the more I realise that Apple falls behind the competition on all at least two of these areas.

Reading these forums its pretty obvious Apple has its own sub standard manufacturing processes and poor build quality to deal with.

Seems that those reviewers really are partial to Apple and not being objective.

Even so, really looking forward to the delivery of my 15" rMBP tomorrow!
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.