|
Mac GeekeryGet your geek on. |
|
blog advertising is good for you
recent popular content
User login
|
What with all the hubbub about the thermal paste application on the MacBook Pro, I figured I’d throw my officially-trained, Apple-Certified bits into the mix. We have two MacBook Pros in the Geekery Compound, and we’re a little tired of roasting chestnuts, if you know what I mean… There’s been conflicting reports as to whether reapplying the paste really makes much difference, but no one’s really done a bang-up, knock-down, scientific experiment on the issue. Well, that’s about to change. Read on for our results. We have included the raw measurements for those interested in which areas were what temperature. For each reading there is a high, low, and last recorded. These were averaged together for the final numbers. The ProcessWe tested our MacBook Pros first with the stock, factory applied thermal paste, exactly as they shipped from Apple, and then we reapplied a mere dot of God’s very own thermal paste, Arctic Silver 5. We took three independent measurements from eleven points on the case in four sets; idle on AC power, idle on battery power, heavy load on AC power, and heavy load on battery. We took a fifth set of measurements in the Arctic Silver test, this time intentionally disconnecting the main sensor connector on the logic board, causing the fans to run full-blast. Three measurements were taken from each of the following areas during each test and the results averaged below:
Internal temperature was measured with the SpeedIt kernel extension and Temperature Monitor. Temperatures were only measured once both load and temperature were stable for five minutes. All measurements were taken with a LaCrosse infrared thermometer by moving it slowly around the specified area for several seconds. The ResultsWe found that applying sane amounts of Arctic Silver 5 made a small difference, but nowhere near the twenty-plus degree Celsius improvement reported by one site. We’re lazy, slovenly Americans, though, and we measured everything in Fahrenheit. We’re not NASA and conversion errors won’t waste a multi-billion dollar probe, so you can do the math on your own if you think in “C”. Apple’s Overzealous ApplicationNo doubt, Apple applies way, way too much thermal paste. The quantity I found in our MacBook Pros was simply insane. Properly applied, the paste on one chip could have been spread across several machines. My initial thought was that Apple is possibly wasting a tremendous amount of money. We took all of our measurements here before opening the machine and seeing this, though.
Arctic Silver 5After wiping off the insanity of Apple’s factory pasting, we applied mere dots of Arctic Silver 5, reassembled the ‘books, and re-ran the same battery of tests.
As promised, we also tried these tests with the sensor disconnected, as first discovered by James Duncan Davidson. Disconnecting this sensor caused the fans to run full blast, full time, as expected. We only bothered to run this test under load, on AC. The “delta” is comparing against Apple’s paste on the AC/Load test.
For the skimmers, here’s a pretty chart:
Only AC temperatures are represented because they are the highest. What’s It All Mean, JC?The short version is that reapplication helps… but I’m not sure it helps enough to bother with. For those of you too lazy to convert F to Cs and do a little arithmetic, the largest delta was -11.7F, or about -6.5C, while the test with the fans crippled to run full blast actually reproduced the mythical “fixed” temperatures. Et toi, codepoet?It’s bunk, the whole lot of it. The pressure forces the paste out as designed, the computer is a freaking twin-core 2GHz beast of lightning drawing 85W from the wall into a metal case and who said that should be a cool machine? It’s a desktop replacement. It’s a portable computer. It’s a notebook. It is not a laptop, and they haven’t been for years. It’s no hotter or cooler than I expect (it’s my daily machine) and, frankly, if I’d known this is all the difference this would have made, I wouldn’t have gone through with it. Instead, what I did that dropped the temperature over 20ºF is to just prop the back of the unit up. There’s a myriad of notebook stands out there to do this for a better typing position, so get the one that leaves a nice air gap in the middle and you’ll see at least a 10ºF drop. I got a little stand with USB-powered fans on it (the Targus Tornado Chill Pad) and it goes down anywhere from 20ºF to 40ºF depending on load. JC uses a Targus Traveler CoolPad and we saw about a 10ºF drop in temperature on the idle test. If you want to drop your notebook’s temperatures, don’t mess around with the guts of the beast, just ventilate the thing. It’s one big honkin’ heat sink, after all, so blow some air over it and quityerbitchin. The Definitive MacBook Pro Thermal Roundup
About JC
 Author Biography JC is a former Mac Genius and Mac-centric IT worker with a background in print advertising. He earned a reputation as a miracle worker when he saved the day at a new business pitch with the arcane knowledge that Apple’s ADB cables were nothing more than poorly shielded S-Video cables. JC runs the Heroic Efforts Data Recovery Service and writes Ungenius, a tawdry tale of the life and times of a former Mac Genius. You can contact JC via IM or via the contact form. |
6.5C is a big difference on it’s own, but over time it will be even more – Arctic Silver says you need about 200 hours of burn-in period and the temperature will drop for additional 2-5C. This means you should load your cores to maximum and let them work like this for some time and then turn the computer off or put it asleep for few hours so cores would cool down to room temperature. Then repeat few times.
Also – I reapplied the paste in my 17”. I did this just because I couldn’t use my laptop on my laps and you may say whatever you want, but I want to use it this way. Right after the procedure I could use it on my laps even when it was at full load and core temperature was about 90C! This looked like a pure magic! After the recent SMC firmware update things got even better – it’s just worm, nice and quiet as it should be, core temperature at idle is between 46C and 51C (depends if it’s plugged or unplugged and type of surface you use it on). It was 60-65C before all those moddings.
The final note – the procedure indeed is very difficult and dangerous! I’m an IT pro with a lot of experience in similar areas and it took me over 3 hours! Now I would do this a lot faster of course, but still would be very careful. Don’t even think about this modding if you never did something like this before and you don’t have the apple’s service manual and all the tools they list for the procedure!
We let the AS set for about three weeks before testing it. Forgot to mention that. That’s one of the reasons this has taken a while to get up. It’s not 200 hours, but we felt it enough.
Ok, but most likely it’s still going to decrease after few more full-load-cool-down cycles.
Also – would you post your measures for the bottom case only? Probably this is the most important area after all and this would be very interesting to know what was the difference between highest bottom temperatures before and after.
Surprisingly not. Most of the heat went out the top case at the upper right. Heat rises after all.
I’ll clean up the data and post the rest tonight.
All the readings have been attached to the article now.
I have to both agree AND disagree with the results here.
I found my system becoming unstable under load – kernel panics once a day at least. The fans were ALWAYS audible even while not heavily loaded. I was monitoring the temps and they looked high but still reasonable. I ran the OpenSSL speed tests and the processor shot up to 100 and system crashed.
I cleaned the Apple goop off and applied AS5. Upon firing the MBP up, I noticed the fans stayed quiet. I did normal work – PhotoShop, InDesign, and FCP. Still NO fan noise. Temps were still about 60C but without the fans spinning up.
I reran the OpenSSL bench and the proc temp shot up to 92C and stayed there – finally the fans came on but still not to full power. As soon as the bench completed the proc temp fell back to 60C within 15 seconds and the fans wound down!
This tells me one thing – the old HS grease wasn’t transferring efficiently to the heatpipes and the fans stayed on. Basically it could only transfer so much heat per second. With the AS5 it was able to transfer all the heat the proc generated with no buildup in the chip itself.
My experience shows me it’s not just about the temps reported, it’s also how fast the cooling system can get the heat out of the cores. AS5 properly applied works wonders to the MBP. I have since done several other machines for coworkers but none were as bad as mine with the crashing.
BTW, no crashes whatsoever in the three weeks since I did the HSG upgrade.
Six weeks or so later after applying AS5 to my MBP.
Still no kernel panics or crashes where before I could count on it if I loaded down my processor.
I reran the OpenSSL benchmarks yesterday in similar environmental temps as immediately after applying the AS5.
Processor temp peaked at 85 and stayed there. It took over a minute to even reach 85. The fans stayed at 2000 rpm for over a minute into the test and they maxed out at 3000 once the temp peaked. Never went to full power (6000 rpm).
Within ten seconds of the speedtest completion the processor was back to 65. Ten seconds later it was back to idle temp of 60.
This tells me the AS5 has thermal cycled enough times to become fully effective.
Now I am a true believer that Apple needs to rethink their heatpipe installation procedures AND I am now a believer in the AS5 product.