Thinkpads Forum

Does anybody know how to stop this? Sometimes when my screen goes blank after idling for a few minutes as defined in the power scheme, it turns itself on again, triggered by the trackpoint.

That is a bit of a surprise, as I haven't seen trackpoint drift light up a screen. Indeed, if I press the power off button while the screen is black, the shutdown progresses a long way before the screen lights up.

So maybe there is some other application (I don't know what) or spyware that is activitating both the screen and the trackpoint.
... JDH

You could try to lower the sensitivity in the trackpoint-properties.

I have already pushed the trackpoint 4 notches down from maximum and it still occurs.

Comage, how do you know that this effect is caused by the trackpoint?

call a priest?

however i did have a similar problem with my T21

sometimes it may move by itself. no worry, this is common across all the model, auto-recovery for the track point to its origin position.

Hi gunston, thanks for the assurance. I've sometimes been having the same problems with my old A20p too.

Is there really any real way to resolve this? This is because I normally leave it to turn off after 3 minutes, and I leave my laptop unattended most of the time (don't really want the kind of screen burn-in that happened to that guy in the other thread).

Not sure if this helps but I know this is an issue with dell laptops. To resolve the problem, the keypad needs to be replaced. once done the problem is fixed.

Sorry for digging this old thread up, but I just recently found out that this may not be the trackpoint after all.

This problem only occurs when the CPU is under constant load, and if idle, the screen never turns back on.

There is a process involved that spikes up to 10-20%+ CPU usage when the screen turns back on by itself. It's called "ntkrnlpa.exe". I tried googling for it but haven't had much luck.

I suspect the new speedstep or temperature throttling is the culprit for this "turning back the screen on" issue.

Any advice will be greatly appreciated.

pretty sure its still the problem of the trackpoint drifting by itself.

I have mine turned up to the highest sensitivity and after zooming all over the place it has a tendency to keep moving the small amount.

What the trackpoint software needs is a "dead zone" that many joysticks have in their calibration software. It may be that different trackpoint heads exhibit different issues - I don't use either the eraser head or convex points - I prefer the indented concave instead.

The TrackPoints have a software dead zone - it detects when the TrackPoint stays in a certain position for 3 seconds, and then makes that the center. Usually, it moves a little bit each way when moving normally, so it doesn't auto-center (such that when you let go it shoots the other direction.) Also, increasing TrackPoint sensitivity decreases the likelyhood of an auto-centering. (I can get it to happen if I REALLY try with my mouse speed at near maximum, and my sensitivity at maximum, but not in normal use.)

Anyway, TrackPoint drift usually happens in two situations: when it's REALLY hot (my X21 would occasionally do it in very hot conditions,) and when the TrackPoint is faulty. Call IBM and tell them that your TrackPoint is drifting, they'll send out a new keyboard.

Two things I have tried:

1. I disabled Trackpoint in BIOS. The screen still turns back on after some time.

2. If I run on battery, the above does not happen ever.

Note that this problem happens when the CPU is under constant 100% load. If I leave it to idle, there is no problem ever either.

Out of curiosity...

Is this a tablet? If so, what position is the screen in when this happens, and where is the digitizer pen? (I'm thinking possibly an over-sensitive digitizer?)

Have you tried killing that process? Worst that happens is you'll have to reboot.

This is a first generation X60 from Mar '06.

The process is hidden under the "System" process, and it's one of the 10 or so threads that's spawned under it. Killing it does not work; some other thread with the same name just "takes over its job".

I have tried a factory restore, no go there.

I certainly don't remember this problem when I first got it. And I'm pretty much up against a dead wall because nobody else runs the laptop at constant full cpu load (so that the problem can be replicated).

All advice would be appreciated. Thanks in advance.

I would HIGHLY suggest running Process Explorer, it can tell you which thread is causing it, and let you kill that thread.

I use it on all of my Windows systems in place of Task Manager - much more info, and much easier to diagnose issues.

Yes, I am already using ProcessExplorer, that is how I managed to find out that it was the process "ntkrnlpa.exe" that was causing it.

*Edit*: I have tried killing the process, but with an error message: "Unable to terminate thread thread: Access denied."

Ah, being a tricky little thing...

Follow these directions (just use the at trick, if you're only doing this once) to get cmd.exe running, then start Process Explorer from the command prompt. Voila, LocalSystem permissions (much higher than Administrator,) just like that.

Same error message occurs; I can't kill the thread.

I did however, try to suspend the threads one by one (there are 24 of them nasty critters). But the spike occurred with the next thread after I suspended the first one.

This came to a point when the system simply locked up because I suspended too many of the threads.

I simply can't believe that I'm the only one around here experiencing this problem...

Any specific thread names that you recall?

This is looking like a tricky hardware problem.

You know... for S&G, run memtest86+ overnight. I'm suspecting RAM...

Thanks a lot for your help bhtooefr. I really appreciate that.

I'll try running Memtest86+ later tonight. How do you think that might be the problem? The thread names are all called "ntkrnlpa.exe". It seems to be an NT kernel process, which explains why it can't be killed. Or it might be that the process is not the cause of the problem, but a symptom. I'm not sure how to go about troubleshooting this issue already.

I've tried flashing back the BIOS and it doesn't help either.

What's S&G btw?

Well, the threads that you're killing are causing system instability, which tells me that it's in the kernel, and/or it's something that works with hardware.

Weirdness with the kernel and weirdness for hardware... first (or second, after temperatures) thing to check is if the RAM is any good.

Second thing to check in this case, if it doesn't feel like it's running hot, is malware, which I do wonder if it is... run your favorite anti-virus (I prefer Avast) and your favorite anti-spyware (honestly, I prefer Windows Defender, but on a system where I know something's seriously wrong, I hit it with Spybot Search & Destroy and AdAware.)

Edit: Oh, and S&G means sh*ts and grins.

8 passes of memtest, with 0 errors. I did try a factory restore but it still exhibited the same symptoms. Spybot reports nothing.

Well, it was worth a shot.

You might also run PC-Doctor.

Thanks for the advice, I'll try that later.

Just a quick update, I tried to remove one of the sticks of ram I had, leaving it to boot up with only one stick.

WinXP went into a 16-bit color mode, and the disk chugged along pretty sluggishly like it was running in safe mode.

I think I managed to resolve this issue: I changed the processor scheduling under the system properties to adjust for programs instead of background services.

*Edit*: Spoke too soon: The problem seems to be less pronounced when I've got 1 stick of RAM running in it, compared to 2 stick before.