my T60p hates me

[cheeebs]

Hi All,

I have a T60P (2613-CTO). It has had recurrent problems since day one. I have always had thoughts on the problem, but now I believe it must be hardware. I'll start with an example of the problem, then more detail following.

My laptop will:
1. suddenly & unexpectedly turn off -- black screen. not freeze, but die, as in instantaneous power off. upon restart it will ask me to launch R&R, which will hang on file loading about half the time, or hang on the system restore the other half.
2. if i bypass R&R for "Start Windows Normally" (any windows version, 7-64, VPro-32, XPPro-32, dozens of installs from various disks) it has a 50% chance of hanging -- screen freeze while still in VGA and no HDD activity.

Last week Lenovo System Toolkit scanned my system and said HDD failure was imminent due to SMART test results. So I replaced my HDD with a new non-refurb Hitachi, and then I did a clean fresh install, not a restore nor an image. Guess what, my system hung on boot during the 3rd boot after completing the install. This is the third fresh HDD I have installed thinking I had duds for drives, but apparently I was wrong.

My laptop won't boot 50-70% of the time if it's undocked. It will often boot if I redock it, but not always.

I have tried reinstalling the OS with Lenovo Vista 32 disks many times and I get the same problem. So I ordered a set of Lenovo XP 32 disks and reinstalled the OS many times with the same problem. Finally I bought Win7-64 directly from MS (non-Lenovo disks) and did a full wipe & reinstall, same problem. I have, of course, attempted to run repair utilities, restores, crapware/malware/rootkit scans and all that... same issues.

Out of pure frustration I finally gave up on Windows & installed Linux. In the last 2 weeks of Linux it has not black-screened on me (yet?), however it HAS failed to boot (hung on boot) when undocked in the same fashion that Windows hangs on boot.

This problem is intermittent (OH JOY!), and although I can sometimes have 2 or even 3 full weeks without a shutdown symptom, once they start they come in waves. Yesterday I had 3 sudden deaths in Win7 combined with multiple failed reboots in Windows & Linux both.

Here's a few points to note:

1. This problem is not related to having Linux onboard... this symptom predates any dual boot installation.
2. I do not get any memory dump nor error log from this shutdown symptom... it happens immediately and leaves no obvious trace.
3. I do not have any other hardware issues of which I'm aware.
4. I have tried many varied BIOS settings over the life of this laptop including disabling multi-core etc etc. Maybe there's still something in there not optimally set, but that wouldn't seem right since the problem is so intermittent and seemingly not directly related.
5. My battery is in fine condition, if that's of question.

I'm about to either trash this **bleep** thing which is hard since I can't afford a new one, or pray for another option, but running a machine with this kind of symptom is far more frustrating than not having a laptop at all. If you have any thoughts, PLEASE contribute! Thanks!

[Neil]

I'm guessing you don't have warranty left, right? Sure seems like you have a faulty system board that Lenovo would replace under warranty. Hopefully someone else will have some better suggestions for you...

[richk]

Is your battery Lenovo or aftermarket? I have seen 2 with bad batteries that report OK, but cause those symptoms. I have a 9-cell battery here on my disk made by Agptek that has a low cycle count, reports a full charge, will drive the laptop, etc, but will act as you describe. Try running with a different battery or, if you don't have a spare, no battery. Other than that, I guess motherboard.

[Brad]

I would have to concur with Neil and Rick. A motherboard replacement would likely end the problem.

Brad

[crazyfrog]

You may also consider memory issue. When your t60p runs, turn it into hibernation and then reboot. If you immediately see blue screen, you got bad memory stick.

[SHoTTa35]

I had that on my T60 and i think it was a overheating issue. To combat it i would have to go in the BIOS and turn the CPU settings from balanced to performance or something like that. That way it would boot instead of hang on the Windows loading screen (after a random shut off like that).

[cheeebs]

Thank all you guys for your help and suggestions. I am in the process of trying a few more things to whittle down possibilities, but in the end I will likely have Brad swap the motherboard. I gotta keep this baby running... where else am I gonna find such a sweet UXGA Flexview 4:3 Hydis panel as the one in my T60P? Nowhere, apparently!

Just set all my BIOS power mgmt settings to Max Performance hoping that will help with boot & docking issues. If I can boot into Win7 then I'll run some thermal diagnostics.

[cheeebs]

i did the full boot time mem test and all is clear with 0 errors.

i installed tpfancontrol as a service. no shutdowns today, still here are my current numbers:
1 cpu 63°C (0x78)
2 aps 41°C (0x79)
3 crd 43°C (0x7a)
4 gpu 76°C (0x7b)
5 no5 50°C (0x7c)
7 bat 36°C (0x7e)
9 bus 47°C (0xc0)
10 pci 54°C (0xc1)
11 pwr 63°C (0xc2)

anyone know what ranges i should or should not be seeing on a T60p? also, why would the temp not be properly controlled by OEM software?

[richk]

Yours look too high.

[cheeebs]

Yours look too high.

ok, and what's interesting now is that this was after booting when cold, browsing to find tpfancontrol, download & install, and rebooting, then there's the reading... maybe 15 minutes of usage.

looks like the ghost in the machine is beginning to show itself.

[cheeebs]

just after waking up my machine this morning... it's been awake for maybe 20 minutes:
1 cpu 88°C (0x78)
2 aps 39°C (0x79)
3 crd 42°C (0x7a)
4 gpu 96°C (0x7b)
5 no5 50°C (0x7c)
7 bat 20°C (0x7e)
9 bus 48°C (0xc0)
10 pci 61°C (0xc1)
11 pwr 71°C (0xc2)

fan's running full blast, and i'm just reading email & checking forums, nothing more. does this look unusual to anyone, or just normal for a T60p? looks like i'll have a DIY for xmas day... after everyone homes to the TV...

[SHoTTa35]

Wow those numbers are totally nuts! Check the vents on your machine, gets some compressed air and clean them out. They do get clogged with dust sometimes.

There's a thread that was here (unless it was in the T4x section which wouldn't be valid) that showed temps for lots of systems. Try to see if you can find it as i can't seem to find it right now

Actually, just found one of my posts about it. MY thinkpad was a T60 (not P) with the ATI X1400 and a 2Ghz chip.

http://forum.thinkpads.com/viewtopic.ph ... 82#p380082

Also read this guys post from the same thread:

http://forum.thinkpads.com/viewtopic.ph ... 45#p522645

[GrandMasterKhan]

96 for your GPU is too high. Nvidia?

[GrandMasterKhan]

The GPU usually draws a lot of power on boot up. It's my guess that your GPU gets so hot on boot up it shuts everything down. Maybe open it up and redo thermal paste?

[cheeebs]

96 for your GPU is too high. Nvidia?

ATI

will do a full cleaning, inspection, and thermal paste re-do on xmas. it's the perfect distraction from crazy family members.

[GACrabill]

Just set all my BIOS power mgmt settings to Max Performance hoping that will help with boot & docking issues.

just after waking up my machine this morning... it's been awake for maybe 20 minutes:
1 cpu 88°C (0x78)
4 gpu 96°C (0x7b)
fan's running full blast, and i'm just reading email & checking forums, nothing more.

Setting the BIOS to Max Performance is the same as turning off Intel Speed Step .... Max Performance equals maximum energy and maximum heat at all times. "just reading email" will still result in maximum heat since you effectively turned off Speed Step.

There are also settings for the graphics adapter (not in the BIOS) to control max performance, max battery, and balanced (or adaptive). If you choose max performance for that also, I would expect max heat there too.

But that kind of heat for both cpu and gpu does indeed lead to the belief that some serious cleaning is in order.

[cheeebs]

thanks... i'll change those settings and look for more configs. cleaning on xmas day along with a regrease and possibly some other modifications... whatever i can do that's easy enough and makes sense.

[cheeebs]

Setting the BIOS to Max Performance is the same as turning off Intel Speed Step .... Max Performance equals maximum energy and maximum heat at all times. "just reading email" will still result in maximum heat since you effectively turned off Speed Step.

There are also settings for the graphics adapter (not in the BIOS) to control max performance, max battery, and balanced (or adaptive). If you choose max performance for that also, I would expect max heat there too.

But that kind of heat for both cpu and gpu does indeed lead to the belief that some serious cleaning is in order.

ok, just took that advice and changed my BIOS settings. i changed thermal power management to 'Balanced' from 'Max Performance'. after the change, it hangs on boot (when attempting to exit VGA mode)... just a black screen and no activity. i changed it back to 'Balanced' and it booted into linux like normal. my GPU temp monitor show 92F, but i'm not sure that's correct... shouldn't that read 92C ? anyway, what does this say to you who have experience? does it reveal anything wrong with the m'board or BIOS? seems that the black screen hang during boot is not what the box is designed to do, even if it's too hot to boot. wouldn't it just shut down?

[cheeebs]

well, here's the story:

yesterday i did the following:
- disassembled and cleaned a ton of clogging dust out of the heatsink fins and fan assembly
- removed both heatpipes completely, sanded/cleaned, and re-epoxy'd with Arctic Alumina thermal epoxy
- did the 'penny mod' but used a slightly thicker piece of copper, which seemed to put a tad more pressure on the gpu & northbridge than the coins would have
- cleaned and prepped the chips and the contact plates of the heatsink, and replaced the oem thermal paste with Arctic MX-3 paste

all in all it took 6 hours, but went smooth. unfortunately, the results are not so attractive. i did not do a real before/after so there's no real benchmark, but i'm not so interested in that, i'm just interested in cooling this thing down. it may be cooling better than before, however. a few days ago i booted her up and was doing some email. after about 12 minutes of uptime i saw this:
1 cpu 86°C (0x78)
2 aps 38°C (0x79)
3 crd 41°C (0x7a)
4 gpu 94°C (0x7b)
5 no5 50°C (0x7c)
7 bat 20°C (0x7e)
9 bus 46°C (0xc0)
10 pci 59°C (0xc1)
11 pwr 69°C (0xc2)

right now, after about 30 minutes of uptime AND after a Toolbox ATI stress test (finished about 10 minutes ago -- now i'm just back on firefox, nothing more) i am seeing this, and it appears to be continually cooling every couple minutes from after the stress test:
1 cpu 52°C (0x78)
2 aps 27°C (0x79)
3 crd 30°C (0x7a)
4 gpu 71°C (0x7b)
5 no5 50°C (0x7c)
7 bat 16°C (0x7e)
9 bus 33°C (0xc0)
10 pci 42°C (0xc1)
11 pwr 49°C (0xc2)

wow! looks promising, you say? hold on a few minutes while i run the stress test again...
1 cpu 100°C (0x78)
2 aps 30°C (0x79)
3 crd 30°C (0x7a)
4 gpu 105°C (0x7b)
5 no5 50°C (0x7c)
7 bat 16°C (0x7e)
9 bus 38°C (0xc0)
10 pci 56°C (0xc1)
11 pwr 62°C (0xc2)

well, i stopped the stress test before it climbed any further. this was about 4 minutes into the test, which was only the ATI test, not the cpu test.

here's the cpu stress test:
1 cpu 100°C (0x78)
2 aps 31°C (0x79)
3 crd 32°C (0x7a)
4 gpu 91°C (0x7b)
5 no5 50°C (0x7c)
7 bat 16°C (0x7e)
9 bus 45°C (0xc0)
10 pci 52°C (0xc1)
11 pwr 63°C (0xc2)

i stopped this one about 10 minutes into it since the temps were holding... 100C was the peak.

EDIT:
i ran the Toolbox System Stress Test from beginning to end, and here were the results, and this was also the peak temp of the gpu:
1 cpu 95°C (0x78)
2 aps 34°C (0x79)
3 crd 33°C (0x7a)
4 gpu 110°C (0x7b)
5 no5 50°C (0x7c)
7 bat 16°C (0x7e)
9 bus 47°C (0xc0)
10 pci 63°C (0xc1)
11 pwr 66°C (0xc2)

at this point i'm not sure if there's anything more i can do. i guess i can run a lighter load, but seriously, the whole reason i'm hoping to salvage her is because of my 15" UXGA Flexview display. so, without the eye candy, what am i trying to preserve? maybe i'll just keep running tpfancontrol, avoid excessively stressful video loads, and hope for the best. as far as i'm concerned killing kwin effects, compiz, or win aero are not an option... i'd rather run it into the ground. oh well. any other thoughts are certainly welcome, please.

MERRY XMAS!

[crazyfrog]

After the cleaning and heat sink mod/reset work, has the original die on black screen problem gone?

[Neil]

Normal temps would be about 60~68c for 2D work and 85~95c for 3D. Don't worry about anything less than 104c. Your temps look OK for normal use, just don't run the stress test, it seems to overheat the GPU.

[GACrabill]

... maybe i'll just keep running tpfancontrol, avoid excessively stressful video loads, and hope for the best. as far as i'm concerned killing kwin effects, compiz, or win aero are not an option... i'd rather run it into the ground.

What settings are you using with tpfancontrol?

When tpfc starts on my T60, it's default is Mode=Smart ... which immediately turns off my fan. I then immediately change it to Mode=BIOS and the fan comes back on.

A lot of folks use tpfc to make their tp quieter by running the fan less often or at lower speeds .... less fan noise and more heat is not my goal so I don't use it (and am therefore not an expert on tpfc).

What happens to your temps when you use tpfc with Mode=BIOS ?

[cheeebs]

this situation has been resolved.

i ordered a replacement fan/cooling unit from an ebay seller. the fan unit was harvested from a used tp (had a little dust in it). i cleaned it and installed that one in place of my old unit, which had some damage to the heat pipe from my previous "mod" work. i also noticed that, although the new fan unit was nearly identical as the last one, there were a couple minor differences. it looked like a subtly improved design (for example, the exhaust fins on the old fan were open, but the fins on the newer unit were closed by another piece of copper at the base of the fin assembly).

when i installed the new unit i also used much less of the Arctic MX-3 thermal paste... just barely enough to cover the chips. i still did include the "coin mod", which in my case was not a coin but a small piece of scrap copper about 1.3 mm thick.

when i ran the full-lenth system stress test (which i actually ran immediately after the video chip stress test) the results were amazing to me. the absolute highest reading i could get on the video chip was 75C. before i hit 114C. under the video chip stress test i hit 71C, but it usually would stay below 70C. HUGE DIFFERENCE!

i can't say the former fan was faulty -- it worked and ran above 4k rpm. but i believe it was a combination of the fan unit's ineffective design, and the ineffective or incorrect application of thermal pastes first by the manufacturer and lastly by my last attempt to repair.

the new fan and install has great results. right now i'm running compiz fusion with compositing etc etc and am looking at 66C on the video chip. before repairing her i'd be looking at about 95C now, and if i had any real video going it would immediate head beyond 110C. these results are consistent between Win7-64 & Linux both.

thank you guys for your help & info. looks like this was a success, and i'm very happy to keep my wonderful 15" 4:3 WVA BoeHydis panel going for another few years!