I've been told to use a pea-sized amount. I'm tempted to open up my Macbook Pro again and see how much paste has been slobbered on.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
(HOT HOT HOT) When thermal paste attacks
- Thread starter byke
- Start date
- Sort by reaction score
https://forums.macrumors.com/posts/6748970/
There are also some pic's of Rev A & B in the sticky take apart thread.
Something else I like to do is give a bit of a bend to the "L" shaped tension piece mounted to the top of the heat sink to make sure the paste makes good contact with the heatsink.
For fun, I went back to my 2.33GHz manual - it shows 0.1 to 0.12CC, about 1/3 of a syringe. One-tenth cc is really more like it. The MBP, early 2008 - 0.1 to 0.12 per device, same 1/3 syringe. The 17" UB MBP shows "1/3 of the syringe" for each device, which is probably inline with the other syringes. The 15" UB MBP - same. The MacBook UB - same.
Although the MBA manual did show a good sized blob, I tend to agree it's not really "1cc". They show marking the same syringe in fourths, so that has to be less than 0.1cc, going by the other model's instructions.
I think the picture itself shows the correct amount, but there is an error in the service manual as to the "exact" amount (they don't say it's "exactly" any amount, just that each device gets 1/4 of the syringe.
Sorry if I created a storm. The text was copied and pasted right from the manual (I looked again - no decimal point!)
I wasn't picturing the real volume of ONE cc. Again, the blob shown [in the MBA pic] is nowhere near 1cc - that would cover the entire device
, but I think it's close to the real, 0.1cc amount.
Your guide was brilliant and it helped me tons to get my mba rb repasted.
However there were a few open questions I had to research first.
It would be great if you could tell everyone on your OG post :
**** They need to remove all 10 screws from the base to open it up and to record which screw goes where as they are different sizes.
**** After removing the bottom,you need to remove the battery connection.
To remove this you need to slide the connector to the right.
**** to remove the heatsink plate you need to remove just the screws in that area and not those around the fan.
You are correct sir! Actually, I believe when I posted that I was replying to a member who already opened 'er up but the addition certainly wouldn't hurt. Thanks.
OF COURSE it's working. They've applied the CORRECT AMOUNT. I re-mounted a Q6600 quad core chip into a windows machine today because I was unhappy with the temperature, and used the same amount of paste as they did, and saw 5 degree C drop in idle temperature over how how the machine was before I did that, and something like 10 degrees improvement over the temperature the machine rose to when playing GTA4.
The pictures from their opened up MBA are pretty bad. I can see how that could happen in a factory process, but it certainly shows a lack of pride in your work for a good "craftsman" computer guy to be doing that. Apple are crazy if they want you to slop on paste in those amounts.
For fun, I went back to my 2.33GHz manual - it shows 0.1 to 0.12CC, about 1/3 of a syringe. One-tenth cc is really more like it. The MBP, early 2008 - 0.1 to 0.12 per device, same 1/3 syringe. The 17" UB MBP shows "1/3 of the syringe" for each device, which is probably inline with the other syringes. The 15" UB MBP - same. The MacBook UB - same.
Although the MBA manual did show a good sized blob, I tend to agree it's not really "1cc". They show marking the same syringe in fourths, so that has to be less than 0.1cc, going by the other model's instructions.
I think the picture itself shows the correct amount, but there is an error in the service manual as to the "exact" amount (they don't say it's "exactly" any amount, just that each device gets 1/4 of the syringe.
Sorry if I created a storm. The text was copied and pasted right from the manual (I looked again - no decimal point!)
I wasn't picturing the real volume of ONE cc. Again, the blob shown [in the MBA pic] is nowhere near 1cc - that would cover the entire device
Did you read my latest post? (above) The manual was off by a factor of TEN in the text, but the CORRECT amount would have been applied using 1/4 syringe (and, I'm sure by any tech doing the repair). The picture cannot be ONE cc, just as others have said.
I was simply stating there was a change made and that seemed to fix the problem. It's not PROOF of what the original problem was. If it's working now, then something changed. I looked again at the original OPs picture and looking at the square device (I forget which is which) - it looks like there could have been too little applied and/or misplaced (off center). Take a look at both the device and the pad - there's a large area with NO paste on either part. So, no need to be snippy - there is more to an engineering problem than jumping to conclusions.
They are still using too much thermal paste?
I remember when people were doing this to first generation MBP's and it cooled it off a lot.
I remember when people were doing this to first generation MBP's and it cooled it off a lot.
UGH.
With the weather heating up, I've really started noticing a problem with my CPU temperatures -- to the point where even with just Safari and Transmission (maybe 4-5 active torrents) running, the MBA starts getting very unresponsive (jittery cursor, lagging trackpad clicks, etc.). In the winter, I'd push the computer two or three times as hard, and it still wouldn't max out the fans.
As such, I'm considering doing the whole thermal paste thing. Since I'm placing an order with Monoprice anyway, I thought I'd try their Silver Cool thermal compound. Anyone have any experience with this?
N.B. Thermal compound is obviously a controversial thing, because some people seem to think that toothpaste can have the same effect as Arctic Silver, while others would point to stark differences between Zalman ZM-TG2 and Shin-Etsu MicroSi G751.
With the weather heating up, I've really started noticing a problem with my CPU temperatures -- to the point where even with just Safari and Transmission (maybe 4-5 active torrents) running, the MBA starts getting very unresponsive (jittery cursor, lagging trackpad clicks, etc.). In the winter, I'd push the computer two or three times as hard, and it still wouldn't max out the fans.
As such, I'm considering doing the whole thermal paste thing. Since I'm placing an order with Monoprice anyway, I thought I'd try their Silver Cool thermal compound. Anyone have any experience with this?
N.B. Thermal compound is obviously a controversial thing, because some people seem to think that toothpaste can have the same effect as Arctic Silver, while others would point to stark differences between Zalman ZM-TG2 and Shin-Etsu MicroSi G751.
You should always let the heatsink spread the paste around. The idea is to avoid cavities or bubbles that can cause hot spots. The whole idea of a heatsink is to dissipate the heat and therefore it needs to be in contact with the entire surface.
I've done new installs and also service repair work and have never been told to pre-spread the paste. Also never take the heatsink off and then put it back on after you reapply the paste. That will break the seal and likely cause bubbles and cavities to form. Wipe it off and reapply.
I've done new installs and also service repair work and have never been told to pre-spread the paste. Also never take the heatsink off and then put it back on after you reapply the paste. That will break the seal and likely cause bubbles and cavities to form. Wipe it off and reapply.
A grain of rice is the e-standard measure for desktop cpus, not mobile cpus and gpus like the one in the MBA. Grain sized rice is still too much. Try 1/3 of a grain of rice. It only makes sense since the desktop cpus are about 3x larger than the mobile chips.
And to PinkyMacGodess:
If you know what you're doing, you can use a CC or a flat smoother (I use guitar picks) to spread it evenly. Relying on just squishing the heatsink to spread it evenly is much more unreliable to provide even contact than thoroughly smoothing across the grease by hand. It takes time and experience and skill and its therefore very apparent why production lines and manufacturers would discard this process. Takes lots of time and hands-on expertise.
And true, once the heatsink makes contact with the thermal paste, you do not want to separate them, otherwise you'll have pockets and inconsistencies.
couple tips:
Basically, use a 70%-or-up ethanol solution to wipe clean your contact points on your cpu and heatsink. Use an Aircan to dry it through (Ethanol should evaporate in room temperature quickly so yea, wait for the liquid to disappear). Make sure there's no dust or impurities. It would be best to do this in a dust-free environment.
I gave this a shot on my Rev B the other day, I have some interesting observations. First try I put on about a grain of rice worth and smoothed it out with a CC, seemed a bit thick but I proceeded anyway. Temps were worse than before, I left it a couple of days to burn in. Second attempt I smoothed it with a CC VERY thin as advised. This made my temps even worse than the first time, which were worse than stock.
Finally my last attempt I put about the size of a rice grain in a dollop in the middle of the GPU and CPU, then I carefully squeezed the heatsinks down and locked them down. Lo and behold my temps were below the stock settings.
I used Arctic Silver 5 and on their website they have instructions to use this method, where you put a small amount or line on the chip and squeeze down the heatsink. At least for me it seems to have worked nicely.
Finally my last attempt I put about the size of a rice grain in a dollop in the middle of the GPU and CPU, then I carefully squeezed the heatsinks down and locked them down. Lo and behold my temps were below the stock settings.
I used Arctic Silver 5 and on their website they have instructions to use this method, where you put a small amount or line on the chip and squeeze down the heatsink. At least for me it seems to have worked nicely.
Like I said from above, a grain of rice is still too much. And, spreading it out is a very delicate and difficult process. You have to be in a sterile environment and spread with a sterile and debris-free spreader, evenly yet quickly.
Between pastes, even if you are using the same, you would want to completely cleanse the contact points. Since the paste attaches literally at a molecular level, you'll need a removal solution. Don't bother with the overpriced kits from newegg or whatnot--just use ethanol or alcohol rub, nice and careful. Glad this helped. Wait for another 1-2 weeks for the paste to begin its full curing process and see another 1-2 degrees C drop in temp =D. Now enjoy the 10x better heat conduciveness and transferrence to your aluminum underbody and to your lap.
Between pastes, even if you are using the same, you would want to completely cleanse the contact points. Since the paste attaches literally at a molecular level, you'll need a removal solution. Don't bother with the overpriced kits from newegg or whatnot--just use ethanol or alcohol rub, nice and careful. Glad this helped. Wait for another 1-2 weeks for the paste to begin its full curing process and see another 1-2 degrees C drop in temp =D. Now enjoy the 10x better heat conduciveness and transferrence to your aluminum underbody and to your lap
.Camino 2.0 beta 3. Best with flash, least intensive on resources.
a source: http://www.macosxhints.com/article.php?story=20090302135816608&query=flash%2Bcamino
It isn't a very perfect test, but Camino also uses the least processes when I tried it.
The amount used in the manual is LUDICROUS. This isnt a desktop chip with a IHS. This is a bare die. The die itself is already very flat and smooth. It does not take much thermal paste to fill crevices. For die the best way to apply it is to use a credit card to smear a thin layer. I guess Apple chose the easiest way and just told the employees to drop a glob...
Heat does not magically disappear. It still has to be dissipated. The outer casing has to dissipate the same amount of heat. The only difference is the CPU will run cooler.
Heat does not magically disappear. It still has to be dissipated. The outer casing has to dissipate the same amount of heat. The only difference is the CPU will run cooler.
What I'm curious about is why are the heatsinks so thin? There looks to be plenty of room to put in some thicker beefier heatsinks. I'm surprised no one has put in a slab of heat conductive material in there to channel the heat more effectively.
There's a fine balance between removing the heat and retaining it. Yes, a beefier heatsink would remove more heat from the processors but then where would it go? It would heat up and stay there longer which then would defeat the purpose. The heatsink is made so the heat transfers to the sink, then the fan turns on and intakes cool air from outside the laptop and passes it over the heatsink exhausting it out another set of holes in the rear. A beefier heatsink would require a larger fan/more energy to remove the heat. If you take a careful look at the redesigned (after Gen. 1) heatsink, you'll notice little nubbins. Those little nubbins are important. What Apple has done is effectively created a larger heatsink with the same footprint. I could go on and on about this but then my brain would explode and I'm out of paper towels.
Pics. showing gen.1 vs. 2:
https://forums.macrumors.com/posts/6540582/
Physics
OK - lets think about this together - if there is less thermal paste between the CPU & the heatsink, the flow of heat from the CPU to the heatsink (and therefore the bottomcase / rest of Mac) will increase - this is the design. Casing has great surface area to move heat away from system into surroundings.
Your Mac should feel warmer, because the heatsink & casing can do a better job of distributing the heat through a reduced bottleneck caused by excess thermal compound. This should result in a cooler CPU - this is the whole idea. A cooler CPU would mean throttling of the CPU would take more Apps / threads / torrents - making the Mac in this case more effective.
F
OK - lets think about this together - if there is less thermal paste between the CPU & the heatsink, the flow of heat from the CPU to the heatsink (and therefore the bottomcase / rest of Mac) will increase - this is the design. Casing has great surface area to move heat away from system into surroundings.
Your Mac should feel warmer, because the heatsink & casing can do a better job of distributing the heat through a reduced bottleneck caused by excess thermal compound. This should result in a cooler CPU - this is the whole idea. A cooler CPU would mean throttling of the CPU would take more Apps / threads / torrents - making the Mac in this case more effective.
F
I don't think the case is designed for heat removal. There is no thermal contact from heatsink to case. Case heating is a consequence of the heat being transferred by air, a relatively low conductor to the case. The heat if perfectly designed would go from proc's. to heatsink and be vented 100% out rear via the fan. The case heating is directly related to the small tolerances between heat generating components and the case. I forget which gen. of MBP's went from wee tiny holes to much larger ones for the same purpose. You have to get cool air from outside the laptop over the heatsnik and vent it. Air comes in right hand side of Air as seen from front and exhausts out left.
Nice explanation, TY. I'm wondering though how much energy it takes for the smaller current fan to be at 6200rpm pretty much all of the time versus a larger fan/larger heatsink that runs at a lower rpm and is on less of the time. The issue here seems to be battery life.
It just surprised me when I took of the casing to see that tiny little fan in there and those tiny heatsinks.
It just surprised me when I took of the casing to see that tiny little fan in there and those tiny heatsinks.
Ideally, if just surfing the web with non-flash intensive sites and a couple of app.'s running, the fan should only be spinning at 2500+/-rpm. I usually find something causing my fans to crank in the background and surprisingly enough Safari and iTunes can get it going quickly. Empty cache in Safari works wonders.... Office (2004) products can also get it going and for some reason if I leave Word, Excel, PPt open unattended they'll hog my processors and send my Air into combustion mode. I always need to force quit them at that point. Hulu and YouTube can get it hot too but I'm alway sure to elevate the rear so there's airflow for viewing.
One of the design "flaws" IMO of the Air are the air holes residing underneath. I totally get why they're there by design but it tends to restrict throughput. I take my bottom off and clean all the spider webs/eggs, ghost turds and other pieces of foreign substances that accumulate in the fan and air path every two months or so. It does make a difference especially since I use my Air a bunch in bed and in places that are less than a clean room.
I tend to ramble about these things too much. I've spent the good part of 26 years designing high current circuits and thermal compensation is always the buggaboo of their implementation thus my interest.
I just started to follow this thread and I am shocked that Apple still cannot properly apply thermal paste. I mean, I know that the MacBook Pro and the MacBook was bad, but this is just adding to the list. Check out these Santa Rosa MBP pics (not mine)...the Air is nothing compared to this. It's really sad that Apple can't do this right.
