T30 lock up after a few minutes, with few symptoms

[rjes]

Hi,
My T30 (Type 2366 - 876) started playing up yesterday. Initial symptoms were the cursor / keyboard locking up, with no other symptoms. This has seemed to get more frequent, any the only response I can get is to hold the off button down until it powers off.
After much reading on here I have found a few things to try, like memtest, but nothing definitive, yet. I thought I'd post anyway in case anyone else can help. The symptoms are as follows:

After crashing it will sometimes boot up again straight away, and sometimes not unless I leave it shut down for maybe 1/4 of an hour.
On powering up again after the failure, I get no safe mode or 'windows failed to shut down properly' messages.
Occasionally it also powers itself off (typically when I've come back to it after leaving it working normally).
At best I seem to be able to get it to run for around 10 minutes now before the cursor ans keyboard lock up, or powers itself off.
The fans runs normally when the system boots up, then cycles on and off occasionally, as normal (not cycling like the loose inductor problem I have read about). This seems to rule out my first thought, that maybe the CPU is overheating.
There are usually no errors on boot up, although once I did get beeps in the sequence 4 - 4 - 4 - 4. This does not sound too hopeful from what I read on here, as it may mean a motherboard failure? I have rebooted it many times since getting the beeps, and they have not come back, although it's continued to lock up.
The symptoms seem to be similar to what is referred to as 'blink of death', but that mostly seems to be in relation to T2x's, and to be fairly terminal for the mothernboard. Can it affect T30's too?
The same seems to happen when running on AC, battery, or both (it's a genuine IBM battery).

I have just downloaded memtest 86+, and will be running it soon, if I can get the machine to run for long enough.

I am familliar with the T30 ram socket problem, as I have had, and successfully fixed it before. The symptoms this time seem to suggest it's not what's happening this time though (no 1 - 3 - 3 - 1 beeps on boot up), although I'm about to try an old 256 ram card just in case the 512 which is in it now has failed.

I have also replaced the CPU fan about a year ago, and the LCD backlight about 2 months ago, both of which have worked fine since. Maybe it's time to give up fixing the T30, and buy something newer? - I hope not though!

If I make any progress, I'll post whatever I've found. Thanks,

[RealBlackStuff]

After reading your story, I would still tip on an overheating CPU.
Open her up, replace the thermal paste with e.g. Arctic Silver.
While you are at it, take out the CPU and reseat it, just in case.
Check that the keyboard connector is in properly when you reassemble.

Also, use an eraser to clean the memory-module contacts on both sides.
They are probably dull/corroded. Clean off debris before you put the RAM back in.

[Reo51St]

When you replaced the CPU fan did you use a thermal compound.

Just a thought.

Good luck.

[madkat]

apply GOOD and FRESH thermal paste (e.g. AS5) - and install mobmeter or NHC - to monitor the temperatures of the cpu/hdd

[rjes]

Thanks for all the replies!

I fitter a thermal transfer pad (ShinEtsu PCS-TC-11, www.asaka.com.tw) after cleaning the fan and CPU with alcohol when I replaced the fan, which seems to have been OK for the year or so since fitting. Does anyone know whether these termal transfer pads are any good? The guy in the shop told me they were when I bought them!

I've now ruled out faulty ram, as the fault still happens with a different ram board. However, since swapping the ram around, the time it takes to lock up has increased a lot, from a few minutes to maybe an hour. This has got me thinking along the lines of broken ram socket solder joints (again). When I resoldered them before, I didn't use any flux, so it probably wasn't the best job ever. I think I'll get a flux pen and try again. Has anyone else had ram socket problem without the 1 - 3 - 3 - 1 beep sequence on start up? I guess the exact symptoms could vary depending which pins are loose?

The ram contacts on both the boards I have used look very clean, although cleaning then can do no harm.

Also, I noticed that when holding the power button down after the fault has occurred (the only thing it will respond to), this only makes the machine hibernate (the backlight goes out, but the screen is still on). I am not sure whether this is normal T30 behaviour (possibly determined by bios?), as I always shut it down via windows. I then removed the battery and AC, to make sure it shut down. This triggered the usual 'safe mode' option on restarting, as the previous windows session had not closed correctly.

Next I think I'll take the keyboard out and check the connector, and install mobmeter. What's the max CPU temp for a 2 Ghz T30? If that fails, I'll remove the fan / CPU. Thanks again for your help.

[madkat]

thermal pad is not good for cpu's... and t30's are hot machines - use only thermal paste... even paste gets old every 1.5-2 years or so ...
i've seen on my former T30 (2.0 cpu) a maximum of 86 degrees celsius (with bad paste) and 80 with AS5 - and this with the weak fan

try removing the heatsink and changing the thermal paste

off: a friend of mine brought me last week a Dell 1501 - to look at it - with the same simptoms (freezing etc.) - it is an AMD... installed the NHC - and under load just before crashing it showed 75-77 deg C - after cleaning and applying fresh paste (ordinary - not AS) - temperatures lowered by 10-15 degrees - it never reached more than 63-64 degrees under similar load...
ah - the machine was only 1 year 8 months old...

[rjes]

Thanks. it looks like the thermal pad was a bad choice then. I'll get some good paste instead.

I removed the keyboard, and its connector was firmly attached, as was the one for the trackpad.

I've not managed to check the temperature yet, but have sussed out why it would only reatart after 15 minutes or more after locking up. As it had been hibernating rather than shutting down when I held the power button down, and I'd pulled the AC lead out, 15 minutes is about how long my (pretty useless) battery lasts. I'll continue investigating with the battery removed, and a different power supply, just trying to rule more components out.

[RealBlackStuff]

Indeed, thermal pads are crap. Clean up the CPU and cooler with isopropanol or isopropyl alcohol (from a pharmacy).
Buy some Arctic Silver 5 at the likes of Maplin. Not the cheapest stuff, but very, very good.
Put it on sparingly, about the size of a rice corn, and spread that out evenly with e.g. a credit card.

If your RAM slot fails, it's normally the one closest to the front of the laptop.
And nearly always it's only the 4-6 pins on each outer side.

RAM error 1-3-3-1 normally only comes up if there is no RAM at all, or (with one stick only) the RAM slot is faulty.

[rjes]

Funnily enough, the thermal pads were from Maplin. I'll call in again when I'm passing to get some proper thermal paste, although from the results of my testing (below), the thermal pad seems to be doing an OK job.

I installed MobileMeter. When keeping the CPU busy, the fan speed increases noticeably when the temp reaches 64 degrees, and decreases again once it drops to 54 degrees, and this cycle either continues until the CPU load drops, or the temp levels out at 60 degrees. 66 degrees is the highest I've seen, and 44 degrees is the highest temp I've seen it lock up at so far. I now suspect overheating isn't the problem. The HDD temp has not gone over 40 degrees. Hopefully, if I find out what the problem is, this info will help someone else diagnose a problem.

[RealBlackStuff]

If you seem to be able to cancel the CPU out of the equation for the moment, leave it as is for now.
Concentrate on the RAM, run MemTest, beg/steal/borrow another stick.

Also, run a full antivirus/malware check.

[Mike Blake]

Clean up the CPU and cooler with isopropanol or isopropyl alcohol (from a pharmacy).

And I recommend getting the 91% alcohol. The 50% type is half water and the 70% not much better. I don't know why they always stock it on the lowest shelf--oh, my aching back!

[Maccess]

I have a T30 thqat is experiencing the same problem, but the cause is the driver for the Huawei 3G E220 dongle not loading properly. I haven't fixed it yet, but hope to do so soon. I don't think it's a USB 1.1 issue, as the same dongle works flawlessly attached to an X22 which also has USB 1.1 ports. FWIW, the same dongle also works atached to a T43's USB 2.0 ports.

[rjes]

Thanks for the tips. I have ruled out the ram by tring an old 256 MB module which the machine had in it when I bought it. I don't have time this week to continue investigating the problem (and have transferred all files to my desktop so I can continue working). I have tried removing all unnecessary hardware, but not the drivers. The last things I installed were a USB 2 card (no driver needed), and a bluetooth dongle.
To determine whether it is hardware or a software issue, I'm going to install Windows and the basic IBM drivers on an old hard drive, with no extra drivers, and see whether it continues to lock up. It'll probably be next week before I get a chance now.

[poshgeordie]

Just come across your post so a few days late.

My own T30's developed a GPU fault the same as the T4x's - I just can't be bothered to strip and reflow it!

It uses the same ATI mobility 7500 chip as the T4's and therefore has the same design 'hangups'.

I'm not saying you necessarily have this fault, but it does exist on T30.

Other suggestions - physically check how hot the CPU heatsink's getting (remove the kbrd bezel).

I guess you cleaned all the dust and hairs etc from inside the fan when you renewed the heatsink compound?

Also make sure the fan assembly's correctly seated on the CPU.
You can check this by removing the assembly and checking you have even squares of heatsink compound on the assembly and CPU.
Also incidently, use the thinnest amount of thermal compound - less is more here so to speak.

[rjes]

Thanks for your replies. I've got back to fiddling with my broken T30 today. The plan was to format an old drive and reinstall the operating system and IBM drivers, to rule out anything software related. However, the problem has got much worse today. It's symptoms are now near identical to the T20 'blink of death', which is not a good sign, as it seems pretty terminal. I've never read of anyone having 'blink of death' with a T30 though.

Most times, on power up, it now fails to boot at all. The fan starts, and the power LED comes on, the DVD drive spins if it's fitted and has a disc in, and that's it. No response on the screen, backlight off, no beep codes, etc. When it does boot, up, it appears to do so normally. However, most (but not all) times the keyboard and trackpad get no responses. Using a USB keyboard does not help either. Once it started to boot up normally, then locked up half way through, after the backlight had come on and the big 'IBM' logo, but before there was anything else on the screen. Overall it seems to be the exact same state as it used to end up in after working normally for a decreasing amount of time. When it does respond to keybaord / trackpad input, it seems to work normally, but continues to stop responding after a while.

I also checked whether the CPU temps were accurate with a thermocouple meter on the heatsink. The fan cuts in at 55.9 degrees, which is very close (within a degree) of what mobilemeter says. This seems to suggest that the thermal transfer pad I fitted when I replaced the fan us not as bad a people told me they are. It took (a very, very tiny bit of) cellulose thinners to remove the remains of the pad.

I decided to strip it down again and look for anything stupid like loose connectors. Typically, there is nothing obvious. I then stripped it down as much as possible to, eliminate faulty removable components. With the keyboard, trackpad, battery, HDD, ultrabay device all removed, the same still happens (I powered it up and switched it on from the docking station). Sometimes, it boots up, but usually not. It is more likely to boot up if it has had all power except the BIOS battery removed for a while than it is if I just turn off and on again. I've also had the CPU out and refitted it (complete with another thermal transfer pad).

'Poshgeordie' has PM'd me to say that he has the gear to infrared reflow solder components. Thanks. I am interested, but unsure whether it is worthwhile unless it is possible to get the whole board reflowed? He's suggested it could be the GPU, as he's seen these have broken solder joints before. However, it does not get remotely warm, and pressing down on different parts of it does not seem to make any difference. Poshgeordie - I've replied to your PM. Thanks.

I'm convinced it is a hardware problem, and reasonably convinced time for a new laptop after reading about the amount of effort which has gone in to the T20 blink of death problem, which remains unsolved. The only problems with getting a new laptop (apart from the cost!) are that I have a lot of accessories for my T30. Ideally, an new T30 would be nice, but I very much doubt that they still exist! If all else fails, I'll be breaking the T30 for spares, and listing all the parts and accessories on eBay. I've not quite given up yet. I have pretty much run out of ideas though, unless anyone has any suggestions?

After reading about the 'blink of death' problem with T20's, I am fairly convinced that the fault with my T30 is similar - something to do with the power management system thinking something is wrong, and either preventing power up, or shutting off power.

[poshgeordie]

OK, remove the HDD, CD, and any modem wireless ethernet cards, battery (but leave the CMOS), in fact everything except the CPU, CPU fan, keyboard, and only one stick of memory (try swapping them around the slots too if the problem persists).

By removing everything that's not actually needed to
Stick in the mains adapter.

Press hard on the GFX chip (not hard hard, but hard - I had a T4x with some of the ball grid array (BGA) balls smashed by someone pressing too hard!- use two fingers on top of the chip and press down).

Power up the TP.

Does the display work properly, or in any way?

If it works press F1 to get into the bios and then press F9 to reset to default settings followed by F10 to save the settings. That will take care of any funny bios settings which could be causing other issues.

Take your fingers off the gfx chip and try wiggling the mobo around a little. Does the screen go off? Then it's probably the gfx chip.

If it doesn't, power down (use the on / off switch until it shuts down) and remove the lcd connector from the mobo and connect the TP to an external monitor.
Same again with the gfx chip - any display? what happens when you take your fingers off?

I don't think there's a blink of death problem with the T30's (least I HOPE there's not!), but it's certainly a strange one you have here!

Replied to your PM too thanks.

[rjes]

Hello Nick,
Thanks for your continued suggestions. I've just got back on to messing with it after having visitors. I've just tried all your suggestions. One thing I'd forgotten about removing was the WiFi card, and I'd not previously tried disconnecting the screen as I was unusre whether this could do any harm. With the T30 screen connected, it worked for long enough to get in to the bios and reset the defaults, then locked up again. Pressing on the graphics processor made no difference. With it still running OK, pressing on or tapping the graphics processor or flexing the base frame slightly did not induce locking up.

I then repeated the above with an external monitor and the T30 screen cable disconnected. Nothing induced it to lock up again, but also, it didn't lock up on it's own in maybe 1/4 of an hour. This is the longest it has run for a while.

I've just put the HDD back in, and have been actually using it with the T30 screen disconnected , giving it a fairly hard time (playing video while searching for a file and working in Excel), etc, and no lock ups for about an hour. I really thought i was on to something, but it locked up again as soon as I moved the cursor to select another video clip. I held the power button down, and the graphics have not locked up, as it came up with 'entering sleep mode'. I left it in sleep mode for maybe a minute (with the power still connected), and hit the power button again to see if it would wake up. I got the not very hopeful sounding 4-4-4-4 beep pattern, the fan on, and nothing from the screen. Since then it back to the normal pattern of locking up, and occasionally booting up then locking up.

I am pretty convinced it is not graphics releated, as the screen still seems to be working even when it's locked up (stopped responding to all inputs except the power switch).

[rjes]

I have bought another faulty T30 on eBay, which apparently just needs a new backlight. I wouldn't be too surprused if there is more wrong with it than that, as I only paid £39, but you never know. First I'll see if I can make one good one from the two, by swapping the screens or replacing the backlight. If that works, I'll then continue seeing if I can get the broken one to work, as I wouldn't mind having two working laptops.

[richk]

On a T30, that symptom is most likely a bad memory slot. Try it with the outside slot empty and only one stick.