PC Freezing up - any suggestions?
 

I recently built a desktop computer from components rather than buying a ready-made computer (dell, gateway, etc...).

I have always done this in the past, but am now beginning to regret my decision.

The reason for my regret? The computer is completely freezing up, and I don't know which component is causing the problem.

The problem:
Computer freezes up randomly. By "freezes up" I mean just that - completely stops responding to any input device, with no sign of activity from hard drive. Works fine after hard shutdown and restart. This freezing problem can occur ANY time, and it doesn't matter what is running at the time. It has happened during encoding, while surfing the web, and even when computer is sitting idle with nothing (ie: ran "enditall" to close everything) running in the background. Sometimes the computer will stay on withiout any problem for up to a week, other times, it freezes a few times per day.

My goal right now is to track down the culprit, but I was hoping someone could give me some advice, let me know if I am heading in the right direction, or share their experience if they have experieneced the same.

Possible causes of the problem?

Software - doubtful, as I said this happens any time, doesn't seem to make a difference what is running at the time

Memory - this is my first thought of what is causing the problem - I have 2 512mb DDR400 memory cards. Rather than replace these right away, I wanted to see if I could test the memory first, so I ran "memtest386" overnight, and no errors showed up - could the memory still be faulty?

Temperature/Cooling? - hardware sensors monitor shows cooling fans running, fine, and I am logging CPU and mainboard temps, neither are exceeding 44C, and don't see any correlation between the temp at time of freezing.

Power Supply: Could the power supply be insufficient somehow? (just guessing here)

Hard Drives: my "D" drive has everything on it, including the windows paging file and Adobe scratch disk, and tons of video files - could it be too fragmented? I was thinking of segmenting it into smaller virtual disks - is there any point?

Motherboard: Some googling on general freezing problems with pentium 4s led me to some information about motherboard problems particularly with capacitors. But my soldering skills are not high enough for me to risk trying anything on my own computer. I'm not even sure what I would be looking for to see if there was a problem.

Some more information on the hardware I am running:
CPU: Intel Pentium 4 3G socket 775 1M cache
Motherboard: Asus P5P800
Powersupply: 350W
Hardrives: Operating system and applications on on 40GB WD 7200 rpm, IDE all else on 250GB WD 7200 IDE
Video: ATI All in Wonder 128 pro AGP (this and the 40GB HD are old components from my previous computer)

So, does anyone have any advice on what I should be doing next? My first thought was to just go buy some new memory and try that out.

I know this isn't a general PC help forum, but with all the knowledgeable people reading this forum, I thought it was worth a shot to throw my question out to the masses :)

Thanks,
Icarus

Hi Icarus,

Download this: http://www.damnsmalllinux.org/ (it's around 50MB)
Burn it to a CD, and boot from it.
Open a "console" as "root" user.
That is, after the CD boots and you're in graphical desktop, press "CTRL+ALT+F2" so you get a text screen and there you log in as user: root password: dsl
See if the computer freezes after a while, and see if you get logged errors on the console.
You'll basically be using Linux as a debugging tool to see if there are hardware problem on your machine.
Let us know the results, or if you have any other questions :)
And yes, this is not a general PC help forum, but there are a lot of freeks and hackers here (me included :rotf:), so you've come to the right place :D

-kwag

Hi,

I had this problem with a friend's PC, because of hard drive fragmentation/IO errors.
The freezings occurs randomly, and only a HD change solved it (the pc freeze most of the time in the startup).
In his case, also, the motherboard was somewhat inestable, but the almost systematically startup freeze don't occur anymore with a new HD.

Fabrice

Kwag:

Thanks for the tip about DSL.

It seems that DSL might be more sensitive to my hardware problems than WinXP, though, because it won't boot up at all :(

After 5 tries, the furthest it has gone is to the text "Welcome to DSL" screen before it freezes (usually I just get the 2 penguins and a stuck cursor).

Any idea what I can try next?

Thanks,
Icarus

edit: Don't know if this means anything or not, but on the first splash screen, the graphics are all weird even before it freezes up - could it be that it isn't recognizing my graphics card properly?

You have the same effects when you try to overclock the video card too far. Perhaps there is a problem on your PC concerning the clock rate of the video slot. I wonder if this can be set into the BIOS. Check there.

I checked the BIOS, and don't see anything related to the clock rate of the video card.

The only things I see that might be related are:

AI Overclock Tuner (set to Standard)
AiBooster Support (set to Enabled) --> I disabled this now (right/wrong?)

Graphics Aperture Size (set to 64MB) --> Should I set to 128MB since my card is "ATI ALL IN WONDER 128 AGP PRO" ???

Thanks,
Icarus

edit: I made the above changes and DSL still freezes on startup, so I suppose those weren't the right changes to make

Check that the capacitors on the motherboard - especially around the CPU socket - are not leaking.

I would disable all automatic overclocking in BIOS.

One other thing to try would by raising the vcore by one notch.

Boulder/Dialhot:

Ok, I found the clock settings now (weren't shown before because in "Auto" mode).

I now have AI Overclock Tuner set to Manual
CPU External Freq set to 200 mhz (i can't change this)
DRAM Frequency (I set it to 400 Mhz)
AGP/PCI Frequency - I think this is what Dialhot was referring to before - I set it to the lowest setting 66.6/33.3
CPU VCore Voltage - this was on auto before, so I'm not sure what "one notch" higher means, but I set to 1.4000V
DDR Reference Voltage - No idea, so I am leaving Auto
AGP VDDQ Voltage - No idea... it's on 1.50V, so I'll leave it

Edit: Whatever the changes above did, I still can't load DSL.

:arrow: Try another video card :!:
It seems you could have a bad card, and that's why it's locking up.

-kwag

It could also have bad capacitors. My last card had four leaking capacitors which all had to be replaced manually. I had to underclock the card seriously to even boot to Windows and any gaming caused instant freezing.

Could be the video card...

Now I don't know if I should by a new vid card or new memory!

:?

Boulder: I'll check the capacitors - will it be obvious if there is a problem? Does it make sense that the problem is so intermittent?

I don't know for your location but in France it's alot easier to have a refund when you return a video Card than a memory chip. So i would start with the video card.

Video card is 5 years old.... no chance of returning it :)

but you probably mean returning the new one if I decide I don't need it. :oops:

You misunderstood me : if you don't know what to buy, video card or memory, then start with th VC. If that was not the source of the problem, than you will return it and take a memory chip.

If the top of a capacitor is kind of round or some stuff has been choked out, the capacitor is bad. In general, if they are bulged (is that the correct word?), they are bad or will cause problems sooner or later.

Thanks for the tips everyone - I'll take a look at the capacitors, and if they look ok I'll replace the video card.

What should I do if they don't look ok? Replace the motherboard?

My computer is in an awkward place under my desk, so it will take me at least an hour to take it apart and put back together - I probably won't be able to get to it until next week. But I will keep you updated on my progress when I do get a chance to look at it.

Cheers!
Icarus

I finally had the time to take apart my PC, and I can't seem to find any bulging or leaking capacitors.

http://www.digitalfaq.com/archives/error.gif

Boulder: Can you tell from this picture whether the capacitors look ok, or if that's not possible from the picture, which capacitors I should be paying closest attention to?

Thanks!
Icarus

I don't see any bad ones in that picture, and I wouldn't expect Asus boards to have cheap capacitors anyway but just had to make sure :)

As you have two memory chips, try using just one of them and see if you get any hangs. Then try the other one and repeat. I suggest that before that, try running Prime95 in torture test mode, it will output errors very fast if you don't have a stable system. You could also run it with only one chip installed.

Boulder:

While I had the PC open I switched the slots that the mem chips were in. The computer has been running fine all night long (which is a good sign, but it has happened before). I'm running Prime95 in Torture Test mode right now, and it seems to be running fine, but I'll keep it running and keep my fingers crossed.

Thanks,
Icarus

Edit: The help file for Prime95 says

Before starting prime95, the voltages in my PC are:
+5: 5.46
Core: 2.86
I/O: 3.26
Aux: 3.62

Immediately after starting the torture test, the only one that changes is the "Core" which drops to 2.75 and has been hovering around there since I started the test (about 15 minutes so far). My question: Is a 0.11 drop in voltage considered normal or "significant"?

Thanks again!

@anyone,

I'm looking for good diagnostic software to trouble shoot PC hardware.
I currently use Tufftest, but it's very basic (i.e. harddrive, memory,
diskette drive, video card, etc.) Any recommendations appreciated.

-BP

BlackPrince:

Tufftest looks like just the program I am looking for to run diagnostics on my desktop - unfortunately I don't have a floppy drive. Do you know if there is any way to use tufftest to create a bootable CD without having a floppy?

You could create a bootable CD instead. There used to be bootable CD images at www.bootdisk.com and Nero can be used to burn a bootable utility CD.

I have a good brand PSU with enough output and still the core voltage goes down a bit under heavy load, it's just that the motherboard's regulators can't keep the voltage line steady all the time. I don't have any stability problems so that doesn't matter. Your system has crashed even without any load so I don't think that it is the cause here.

Looking at the tufftest website, there doesn't appear to be any way of creating a CD Bootdisk without first creating a floppy bootdisk. I also can't find a tufftest bootdisk on bootdisk.com.

You are right though, that the load on the system is likely not the problem.

Could it have possibly been a bad memory chip connection or dust in the case causing the problem somehow? When I had the case open all I did was clear out some dust and move the memory chips around. Now my PC has been running non stop for over 12 hours, and prime85 has been running for the last few hours. This is the best my PC has done in months, so I'll still keep my fingers crossed!

Even if my problem has somehow "gone away" a general diagnostics program seems like something I should keep in my collection just in case.

Cheers!
Icarus

http://www.ultimatebootcd.com/ ;););)

-kwag

Thanks Kwag, that looks like a nifty utility!

One question though: Were you suggesting that I can use this somehow to create a tufftest bootdisk without a floppy, or are you suggesting I use the diagnostics already included with "Ulltimatebootcd"?

Thanks,
Icarus

Hi Icarus,

You should create a boot disk, if you cannot boot from CD.
But I suggest downloading the ISO image if you can boot from CD.

-kwag

@Icraus,

Some problems occur due to over heating sensor of the system fan. This
is built into the BIOS of many Dell computers. I had similiar problems
as yours and discovered the computer was over heating. The Geeks website
has diagnostic programs to test such problems. I too would like to create
a bootable CD for Tufftest. Some of my friends don't have floppy disc
drives and I have to use an external one. I like the concept of the "Ulitmate
BootableCD" but, I need DOS based testing programs independent of
Windows to narrow the problems. :)

Update: Try this for Tufftest BootableCD:
http://www.tufftest.com/how-to-use-nero.htm

-BP

Black Prince:

1. My problem appears to have gone away... The only explanation I can think of was that one of my memory slots is bad, becaused I moved the memory cards around and now all is well. However, I would still like to run any diagnostics I can - can you give me a link to the "Geeks website" so that I can run the test to see if I have a problem with an overheating sensor on the fan? (FYI - the computer is not a Dell)

2. I also saw the information on creating a Bootable CD using Tufftest, however, look at the first step in the instructions:

Kind of hard to do if I don't have a "Drive A" :)

Cheers!
Icarus

Heya Icarus

Could I just ask a quick question, your memory dimms, are they in they in the same coloured slots on the motherboard or do you now have one in each?

Reason I ask is thats a dual channel motherboard and for dual channel to work reliably the dimms have to be a matched pair. Any timing differences and you will get lockups when attempting to run them dual channel. If it was dust in the dimm slot causing it then memtest86 would have thrown errors galore up at you.

The other thing to note is that a P4 can draw a lot of power, apart from needing a seperate 12v connection it can draw in excess of 60 watts. As a bare minimum your psu has to be able to sustain 26 amps at 3.3v, 30 amps at 5v and 15 amps at 12v (this should be in a table on sticker on the side of the psu). Intel suggest a 400w psu with dual 12v lines and dedicated 12v auxillary connector, with a 500w psu being recommended for the p4 630 on LGA775. Allthough it doesn't sound like much thats a lot of amps being drawn just for the cpu, without adding in the motherboard, ram, hdd, optical drives, usb devices, fans etc.

Sorry for the late addition to the thread btw :)

Thanks for the input Mark!

For what it is worth, the memory dimms are now in the 2 black slots (before they were in the 2 blue slots).

I have no idea how that solved my problem, but somehow it did. :?

Computer has been on for the past 48 hours now, with prime95 torture test running for 32 of those hours with no errors.

It could be that the chips were not tightly in place or the connectors were a little oxidized etc.

That's what I thought of in the first place, when he said he moved them and it all worked :)
This happens a LOT, specially in places where humidity factor is high, like here in Puerto Rico.
You should see how I jerk memory and PC cards in place when things don't work, and all of a sudden they start working :lol:

-kwag