hard lock-ups caused by hardware (nvidia card)? [solved]

maartenstorm · n00b Joined: 09 Jan 2003 Posts: 46

Hello

since a few weeks I have serious problems with my computer. It suddenly locks up without giving a reason. Only thing I can do is a hard restart by pushing the power button for a few secs. it started a few weeks ago and got rapidly worse, in the end the computer wouldn't even start at all. I disconnected a new installed harddrive, then it got up again, the indication lights at the back of my Dell computer indicated a video card problem. I disconnected my nvidia video card and reconnected my harddisk (I needed that one for booting linux). Since then the video card from my intel chipset was working quite well after changing my Xconf. Anyway, after a few weeks, same problems with the intel card. Hard lockups, in the end even after a few minutes. Didn't know what to do, as a sort of last try reinstalled the nvidia driver and guees what, it worked again with no problems at all. And now after a few days the locking-up started again I guess.
Is it possoible this is a problem vaused by power shortage in my computer? How can i check this?
Or has it to do with something else? The new hard drive is working well, no problems with that.
I got no messages in my logs after the sudden locks, so nothing to post. If you need any info to help me sresolving this problem I'll post that of course.
Any help is very welcome!

oKtosiTe · Posted: Tue Sep 13, 2005 10:28 am Post subject:

Have you taken a look at the temperature of your CPU and video card GPU? Perhaps your computer is getting dusty. This is how I see most computers die (except my own, of course).
_________________
Ask Ubuntu | Super User

reub2000 · Guru Joined: 31 Jan 2004 Posts: 364

To check the powersupply, check the voltages. It should be in a hardware monitor section. I think they should be within 5%, but I could be wrong on that. However, voltages that are off will usually cause restarts. I'd check the RAM with memtest86+ if I where you.

However. could you provide the following information:
Motherboard and chipset
CPU (Overclocked?)
RAM amount and speed
PSU Power and Brand
etc

oKtosiTe · Posted: Tue Sep 13, 2005 12:10 pm Post subject:

Yeah, I would also try what reub2000 said.
Memtest86 or memtest86+ will install to your /boot/ directory, and you will need to add them to your boot loader's menu.
_________________
Ask Ubuntu | Super User

maartenstorm · n00b Joined: 09 Jan 2003 Posts: 46

Some specs from my computer. It's quite hard to find all information because it's a Dell...

I seem to have a Dell motherboard.
Intel Pentium 4, 3.0 gHz, not overclocked, hyper-threading turned off, bus speed 800
Intel Chipset 82801
250 Watt "dell" power supply
512mb RAM
NVIDIA GeForce FX5200 128MB
2 hard disks (1 sata 160GB, 1 ide 250GB), 1 dvd drive 1dvd rw drive

Was trying to install lm_sensors, doesn't seem to work. Can't find anywhere which hardware sensor chips are used in this dell.

I checked all the hardware with a Dell Diagnosis Utility. Everything seemed to be fine, also memory.

I think it has to do with the graphics adapters or power, but 250w should be enough for my hardware, shouldn't it?

Where can I check the voltages for the power supply? Couldn't find anything in my BIOS.

groovin · Posted: Tue Sep 13, 2005 4:51 pm Post subject:

some mobo bios have a temp reading u can look at. not sure about dells though. you can also turn off your computer (unplug it from the wall) and open it up and check how hot is is inside (dont burn yourself). sometimes a low-tech touch test works well enough. i had a problem like that too... so i opened it up, felt that everything inside was smoldering hot, and then pointed a regular home circulation fan into the case and ran my computer... it stopped crashing so i knew it was the heat.

reub2000 · Guru Joined: 31 Jan 2004 Posts: 364

Sorry, I meant the hardware monitor section of your bios.

A little google search revealed that the 82801, aka ICH2 is a southbridge chip, and is part of the 815 chipset. However, the 815E chipset is a Pentium 3 chipset. Seems strange.

maartenstorm · n00b Joined: 09 Jan 2003 Posts: 46

Still can't get lm_sensors to work, give it up, probably not possible with Dell.
Had about three crashes last days. No information about these crashes in Xorg or /var/log/messages (kernel log?). Where else can I look to find a reason for the crashes? Two of the three crashes were using firefox and the other while running the xscreensaver. Turned screensaver off and using mozilla instead of firefox. See if that helps.
Think videocard might have something to do with it. Using the nv driver with Xorg.
Checked my Nvidia card with the dell utility, but the dell utility doesn't find any problem....

maartenstorm · n00b Joined: 09 Jan 2003 Posts: 46

Think I'm getting to know the cause of the problems.
Had a crash about one hour ago. Checked with the dell utility, again nothing wrong with the videocard.
During the Dell utility it crashed again and now it didn't want to start up again, as a few weeks before.
The diagnsing lights at the back of my computer indicated a video card failure and with another try to startup it indicated a "possible expansion card failure".
Took out the nvidai geforce fx5200 card, using the intel onboard. That seemed to be broken, only shaking video output and after starting the computer it shut itself down within a minute.
If the computer shuts down itself this could be a possible overheating cause?
However, when I took out the nvidia card it was quite hot. So I let it colling down for a while. Plugged it in again and for now its working again, at least long enought to type this post.

It seems for me that the video card is getting too hot. I found out in Dell forums the sensor chips are automatically linked with the fan to cool it down, that was why I couldn't get lm_sensors to work with my dell computer.
That leaves me with a few questions:

1) Is it possible the temperature of my videocard is not noticed by my Dell, so it causes a hard lock up?
2) Is there a sensorchip on a nvidia geforce fx5200 card I can read with lm_sensors?
3) Is it better to use the nvidia driver instead of the nv with xorg? Is it possible it makes a difference for the possible temperature on my card?
4) What can I do to keep it cool? (I have no home circulation fan...)

At the moment I'm back to XP for a while to see if the problem occurs there as well.

Hope you can give me answers!

maartenstorm · n00b Joined: 09 Jan 2003 Posts: 46

Problem is solved. A heavier power supply did the trick. Now the CPU didn't get overheated anymore because of the too-hard-working power supply. The power supply Dell put in there Dimension 4600 is 250 watt what was way too less after adding a extra hardrive and dvd player...

Corona688 · Veteran Joined: 10 Jan 2004 Posts: 1204