I took my 15.4 inch late 2008 2.53GHz MBP to the Genius bar today. Actually, I took it in last week as well. Under stress (Prime95), my CPU will shoot up to about 110C. The fans will stay at 2000rpm for a bit (5-10 minutes) before they ramp up. Once they kick in, they'll slowly speed up until eventually maxing out at 6000rpm. My CPU will then stabilize at about 98C-105C, depending on circumstance.
When I took the MBP in last week, the Genius didn't bat an eye when I told him this. I asked if he wanted me to demonstrate, and he didn't. I showed him a screenshot. He said that the computer will shut down when it reaches an unsafe temp, so not to worry about it. He also expressed skepticism about the "third party" temp monitoring software (iStat Pro), and told me that he gets at least "3 people a week" complaining about heat. They check their notebooks in, he said, and never find anything wrong. He was hinting that leaving the computer with him wouldn't me anywhere, and he turned out to be right. They ran diagnostics on it, and couldn't replicate the problem. I told them I could replicate it for them, and he suggested I make another appointment.
Today, I walk into the bar and explain my problem to another Genius. I told him that the thermal spec (tjmax) for the chip (T9400) is 105C. Of course, this is the maximum safe operating temperature per Intel's specs. Once the CPU reaches these temps, the chip will begin throttling itself in order to reduce its temp. I fire up 2 "third party" temp monitoring programs, and 1 CPU frequency monitoring program. I start Prime95 and, after a few minutes, the temp on both of the monitoring programs reaches 110C. I show the Genius that the CPU is throttling down to 1.6 GHz, indicating that the temps are reaching the tjmax. The Genius retreats into the back room to get his buddy, "who knows about this stuff."
Genius number 3 emerges and I get him up to speed. I tell him that the tjmax is 105C and he says "Who told you that?" "Intel," I responded. He asks where I saw it online, and I show him the Intel datasheet I preemptively downloaded the night before. He, again, repeated the "It will shut itself down if the temp is unsafe line." I told him that the shutdown temp is about 125C, which is way above the tjmax. He didn't know what a tjmax was, and thus didn't really understand what it meant. He also expressed skepticism about the monitoring applications I had. I showed him that the CPU was throttling down to 1.6GHZ, which independently shows the tjmax is being reached. He said they could check the MBP in and run diagnostics on it, but that they'll just use the same diagnostics they used last time (which didn't find any problem). I asked him if the 110C temps indicated to him that there might be a problem, and he said they were irrelevant because they weren't the official Apple diagnostics. Again, I pointed to the throttling as evidence and he said Intel CPUs throttle "all the time," and that such behavior is normal. Of course, I said, the CPU throttles when it's not being used in order to save power, and it throttles when it gets too hot. It shouldn't, I insisted, be reaching such a temp with normal use. I told him I'd paid for a 2.53GHz CPU, and didn't want it downclocking itself whenever I try to convert a video or something. He, again, said this was normal operation. I told him it wasn't.
We argued, politely, for awhile. He suggested that, because I have 1.5 years left on my AppleCare, that I ought to try to burn the processor out in that time. If there's a problem, he said, it'll manifest itself as a component malfunction. In effect, he said that he didn't want to proceed with any repairs unless there was a component malfunction detectable by Apple's diagnostic software. Again, I told him that Intel's specs indicate that the current performance is liable to degrade performance and reduce lifespan. He said that Apple doesn't necessarily follow Intel's specs, and effectively dismissed Intel's guidelines.
Eventually, I told him that I wasn't looking for a replacement, I just wanted them to replace the thermal paste and reseat the heatsink. The thermal sensor which controls the fans is on the heatsink, and if the heatsink isn't seated properly (or if the paste is misapplied), then the heatsink temp will lag behind the CPU temp, and the fans won't ramp up fast enough. He agreed, probably just to placate me. On the paperwork I received, it said that he visually saw the temp readouts, but that Apple has no internal support document indicating this is a problem. This, I think, is the larger problem here. Apple trains their techs to believe that all temperatures below the CPU shutdown temp are safe. This, obviously, is not true. If reapplying the paste and re-seating the sink doesn't work, I'll have to do a bit more work in educating the Genuii.
When I took the MBP in last week, the Genius didn't bat an eye when I told him this. I asked if he wanted me to demonstrate, and he didn't. I showed him a screenshot. He said that the computer will shut down when it reaches an unsafe temp, so not to worry about it. He also expressed skepticism about the "third party" temp monitoring software (iStat Pro), and told me that he gets at least "3 people a week" complaining about heat. They check their notebooks in, he said, and never find anything wrong. He was hinting that leaving the computer with him wouldn't me anywhere, and he turned out to be right. They ran diagnostics on it, and couldn't replicate the problem. I told them I could replicate it for them, and he suggested I make another appointment.
Today, I walk into the bar and explain my problem to another Genius. I told him that the thermal spec (tjmax) for the chip (T9400) is 105C. Of course, this is the maximum safe operating temperature per Intel's specs. Once the CPU reaches these temps, the chip will begin throttling itself in order to reduce its temp. I fire up 2 "third party" temp monitoring programs, and 1 CPU frequency monitoring program. I start Prime95 and, after a few minutes, the temp on both of the monitoring programs reaches 110C. I show the Genius that the CPU is throttling down to 1.6 GHz, indicating that the temps are reaching the tjmax. The Genius retreats into the back room to get his buddy, "who knows about this stuff."
Genius number 3 emerges and I get him up to speed. I tell him that the tjmax is 105C and he says "Who told you that?" "Intel," I responded. He asks where I saw it online, and I show him the Intel datasheet I preemptively downloaded the night before. He, again, repeated the "It will shut itself down if the temp is unsafe line." I told him that the shutdown temp is about 125C, which is way above the tjmax. He didn't know what a tjmax was, and thus didn't really understand what it meant. He also expressed skepticism about the monitoring applications I had. I showed him that the CPU was throttling down to 1.6GHZ, which independently shows the tjmax is being reached. He said they could check the MBP in and run diagnostics on it, but that they'll just use the same diagnostics they used last time (which didn't find any problem). I asked him if the 110C temps indicated to him that there might be a problem, and he said they were irrelevant because they weren't the official Apple diagnostics. Again, I pointed to the throttling as evidence and he said Intel CPUs throttle "all the time," and that such behavior is normal. Of course, I said, the CPU throttles when it's not being used in order to save power, and it throttles when it gets too hot. It shouldn't, I insisted, be reaching such a temp with normal use. I told him I'd paid for a 2.53GHz CPU, and didn't want it downclocking itself whenever I try to convert a video or something. He, again, said this was normal operation. I told him it wasn't.
We argued, politely, for awhile. He suggested that, because I have 1.5 years left on my AppleCare, that I ought to try to burn the processor out in that time. If there's a problem, he said, it'll manifest itself as a component malfunction. In effect, he said that he didn't want to proceed with any repairs unless there was a component malfunction detectable by Apple's diagnostic software. Again, I told him that Intel's specs indicate that the current performance is liable to degrade performance and reduce lifespan. He said that Apple doesn't necessarily follow Intel's specs, and effectively dismissed Intel's guidelines.
Eventually, I told him that I wasn't looking for a replacement, I just wanted them to replace the thermal paste and reseat the heatsink. The thermal sensor which controls the fans is on the heatsink, and if the heatsink isn't seated properly (or if the paste is misapplied), then the heatsink temp will lag behind the CPU temp, and the fans won't ramp up fast enough. He agreed, probably just to placate me. On the paperwork I received, it said that he visually saw the temp readouts, but that Apple has no internal support document indicating this is a problem. This, I think, is the larger problem here. Apple trains their techs to believe that all temperatures below the CPU shutdown temp are safe. This, obviously, is not true. If reapplying the paste and re-seating the sink doesn't work, I'll have to do a bit more work in educating the Genuii.
