T43 Type 2668 Overheating w/ other strange issues

[Jason986]

Hello,

Recently the T43 in my sig has begun to go haywire. When I put the CPU and/or GPU under load, the system temp now rises like mad.

Up until a couple of weeks ago, the CPU temp would reach 70C max under sustained load. Now, my CPU hits ~86C under load and rose to 92C the other day during a 20-minute gaming session. This seems far too high for comfort.

It is notable that this occurs only when the system is docked or on AC power. The system temp never breaks 65C on battery, even under heavy, sustained load. Also of interest, the fan never ramps up under load on AC power. It goes through the whole spin-up and spin-down sequence when the BIOS loads, but that's it. It seems to adjust just fine on battery power, though.

Also, I was sitting in a lecture last week when my beloved machine just shut down as I was typing mid-sentence. The machine was not under any significant load when this happened. I knew the battery had 70-80% charge left, so I tried powering up again. Nada. After removing the battery and reinstalling it, I was able to get the machine to boot. I got back to Windows and moments later, the battery icon in the system tray changed from green to gray and a warning message appeared about a battery not being installed. Of course that was impossible, as the machine was running on battery power at the time (and the battery was firmly installed in place). Several minutes later, the battery icon changed back to green and the system continued to work normally. Now once or twice a day, the system will lose track that a battery is installed, regardless of whether it is actually running on battery power (though I've yet to have it randomly shut down again).

I ran System Update and brought all my drivers, ThinkVantage software, the BIOS, and the embedded control program up to the most current versions, yet the problems continue. I tried uninstalling and reinstalling Power Manager and that has not seemed to help any either. Also, the system manages to pass all the PC Doctor tests fine.

So with all that in mind, what could cause these sorts of issues? I called IBM/Lenovo Thursday night and I've got a tech coming out tomorrow with a replacement system board, CPU, fan assembly, and battery. I'm just curios what (if anything) I did to this machine to cause these problems or if some component just randomly failed.

Any thoughts are appreciated.

Thanks!

--Jason

[christopher_wolf]

The fan should really ramp up under load and you really shouldn't be reaching 90°C+ temperatures after a half hour gaming session.

It seems like there might have been some loss in the efficiency of the cooling system. If the airflow isn't seriously impeded by an object, then that might imply that the fan is having trouble. Did you check the RPMs before and after? I can post some of mine under the same loading conditions if that helps since the models are similar.

[Jason986]

Thanks for the quick reply Chris.

As far as I can tell, there are no major airflow restrictions in my machine. I took a can of compressed air to the system a few days ago and got out a bit of dust here and there, but nothing extraordinary.

As for the RPMs, post away. Though where are you getting the fan data from? I've got Notebook Hardware Control on here, but not the ThinkPad Fan Control Program another forum member made. Should I try installing it and see if I can force the fan out of "slow" mode (or at least to see how fast it is spinning)?

--Jason

[christopher_wolf]

At idle CPU usage, I am getting a maximum of 4000 RPM, low of 3174 RPM, and temps in the 45°-52° range. During renders or heavy multitasking, I get temperatures around 72°C and, very rarely, touching 80°C. Nothing near 90°+ temps.

Get TP FCU and see what responses you can change with th efan and cooling.

[agarza]

It sounds like there's something obstructing the fan. It may be a combination of either the fan assembly and motherboard, or the overheating has damaged something in your battery? or in the contacts of the motherboard. Make Lenovo fix that for you.

christopher_wolf: does your CPU has some Arctic silver on it, or you're using Stock thermal compound. Also what's your GPU temp?

[christopher_wolf]

Stock system, thermal compound has not been modified and my GPU temps, at idle, hover in the 48°C-54°C range from idle to moderate graphics loading respectively. No AS5.

Also, it isn't an airflow obstruction as Jason already pointed out, nor a problem with the battery as I can't see what that would have to do with it.

[agarza]

Ahh OK. My CPU has Zalman Super Thermal Grease applied, the GPU as well. But browsing net, listening some MP3's the CPU temp is 47-48C, I think the same as the stock thermal compound. At load, max CPU temp is about 63C.

[Jason986]

Okay, played a 20-minute game of Defcon with TP FCU running and got some interesting results. About five minutes into the game, I started hearing beeps from within the system. Something like BEEP <four second pause> BEEP <four second pause> BEEP ... . It did that at several times during the game. When I exited the game, TP FCU had exited itself, but I reopened it and got some interesting data. When the CPU was 71C, the fan was reading 3960 RPM. 15 minutes later, when the CPU was down to 51C, the fan dropped modestly to 3850 RPM.

So does anyone want to venture a guess what this means? Also, does TP FCU dump a log file someplace where I can check and see what the temps were when the beeping began?

[christopher_wolf]

Beeping from the planar? Not necessarily a good or bad sign but, I have never heard of that in such a situation.

TP FCU dumps a log file out named, fancontrol.log, in its work directory wherever it is that you have placed it on your system. That covers all of the instances of TP FCU up until the last run. If there are any errors, they also show up in the log.

Sure the beeps weren't coming from errors thrown by TP FCU itself? Because it also beeps when it encounters an error state.