System crashing - where to start troubleshooting? 1

[BaudKarma]

I don't even know which forum to post this in, so I'll start here. I upgraded my system last spring - Asus K8N mobo, 2 SATA drives, 1 IDE drive, 1.5 gig of RAM, ATI 9800 video card, ask about any of the other stuff if you think it's important.

The systems been giving me problems since I put it together. I traced the first issues to an inadequate power supply (got 550W now) and then to the video card overheating. Now it's having problems again.

About once a day (more or less) the system will lock up hard. The mouse freezes, there's no keyboard response, and the hard drive light comes on and stays lit. I have to power it off and restart. Sometimes it restarts okay, sometimes it'll lock up the same way almost immediately on reboot. Eventually it'll come back up, and then it's okay for a few more hours.

One of the crashes (or a series of them) managed to screw up the OS, so I reloaded XP and hoped it was some strange software problem. Nope, still locking up. Then I suspected overheating, so I pulled the system out into the middle of the room and pulled off the cover. Still locking up. By the way, device manager looks happy and there's nothing unusual showing up in the event log.

I've done all the quick/easy fixes I can think of, reseating all the cables and the RAM, monitoring the CPU temperature, and nothing has worked. It's time to get my hands dirty, start pullng out memory and swapping out the power supply and disconnecting hard drives and all the rest of it. Given the intermittent nature of this lockup, it's likely to be a fairly drawn-out process. I'm looking for some ideas about which component to check out first. Thanks!


I try not to let my ignorance prevent me from offering a strong opinion.

 

[KevinADC]

first I would make sure the CPU and RAM are set correct in the BIOS: no over-clocking or under-clocking issues.

Make sure the CPU fan is clean and spinning properly (no cables rubbing on it for example).

Do you have hardware monitoring turned on in the BIOS and maybe in the OS too so you can see what the system specs are while the computer is running?

Make sure any on-board features are disabled if using an add-on card in place of the on board chips and that they are in the correct slots and being initiated in proper sequence (if any).

Now in the OS, set the video acceleration all the way down to as slow as possible if you can.

Check the website of the video card manufacturer for any known issues related to the card you are using.

BIOS upgrade might be worth looking into.

( SATA drives use very little power (250mV?) so the use of three hard drives should not be a power problem. )

Run the computer and see what happens.

If the crashing continues, first I would remove a stick of RAM and run again (continue process if necessary until all RAM has been checked). The type of crash you describe sounds like hardware related to me: total system freeze up.

Like you said, this could be a difficult one to pin-point.

 

[electronicsfreak]

Also a couple things to ad on to what kevin has already mentioned. First off on ram check compatibility with the motherboard. I dont mean if the board supports pc3200 making sure its pc3200, what i mean is check on the motherboard or ram site and see if the other is listed as compatible. Otherwords look up your motherboard and look at all makes and model of ram that is compatible. Incompatible ram can cause many problems. Also after like kevin said you check bios and make sure your ram is set correctly, download this memory tester and test one chip of ram at a time. Otherwords have only one chip of ram in the computer while you test it.

http://www.memtest86.com/

Also check your processor temp and make sure its not overheating. And another thing set your bios to safest settings and then go back and manual adjust things to how there supposed to be and dont overclock anything. I know manual adjusting after that sounds as if its defeating the purpose but its not.

 

[electronicsfreak]

As another thing I think it might be either your ram or processor. Ram might be partially bad. If so when going to load sometimes it can forget what its doing think the registry is bad and try repairing the registry then forget again when its doing that and corrupt it badly and crash the os. Had a windows 98 system crash 5 times in a row because of a bad stick of ram. Also if your processor is overheating the windows will get slow or lock up or even burn up itself. Anyways hopefully this information helps some.

 

[garebo]

One suggestion for you. You mentioned taking the cover off, i believe. That doesnt always help with cooling as it can interfere with fan setup and actually be worse.
But what you can do is take the cover off and have a regular house fan blowing hard right on the cpu and vid card area. If you do that and your pc runs fine for a long time then you at least know its a cooling issue.
If so then take the hs\fan\cpu all apart, clean well, re-install well with new thermal grease or even get a better heatsink\fan combo.

However, i also think it sounds like a ram issue and that you should check the ram. You can google and download "memtest" and run one stick at a time with memtest and you will find once and for all if its a ram problem.


Good advice + great people = tek-tips

 

[BaudKarma]

Memory was the way I was leaning as well. I've got 3x512, from two different manufacturers, so there could well be a bad stick or some compatability issues.

The other odd issue (and this one just occured to me this morning) is that the machine only seems to lock up when I'm using it. I leave it on 24/7, and I don't recall ever waking up in the morning or coming home from work and finding it locked up. It's not exactly sitting idle at those times, either. I've got a p2p app running, virus scanner going, and Ghost backing things up the the IDE drive every few hours.

I try not to let my ignorance prevent me from offering a strong opinion.

 

[garebo]

Still leans toward ram.
I do hope you are running with a UPS though. Especially when operating 24/7.


Good advice + great people = tek-tips

 

[KevinADC]

have you tried any of the suggestions posted so far?

 

[BaudKarma]

I've got it running with one stick of RAM right now. No crashes yet.

I try not to let my ignorance prevent me from offering a strong opinion.

 

[BaudKarma]

Still no crashes after a week. I actually at this point have a couple of candidates for what was originally causing the problem. Gonna plug the suspect RAM back in and see what happens. Stay tuned.



I try not to let my ignorance prevent me from offering a strong opinion.

 

[ceh4702]

Check the Processor Cooler Fan and see if it is actually turning all the time.

You can do this by running it with the Side off and just starting it up and running it.

Then look to see if there is lots of dust on the Cooling Fins. Dust kills CPU's all the time. If it builds up the CPU will just slow down or stop inter mittently till it gets cool enough to run some more.

You might also have a bad Mouse. Just try a different Mouse. If you are using a USB mouse might be the cable or the connector or the system might run better with a mouse that can plug into the PS2 Port. USB is sometimes not quite good enough for the mouse. Also try a different USB Port. Keyboard could cause problems also.

Heat often causes problems you would not expect. The Processor can overheat as well as the Memory and the Video card and the video card memory. Some video cards have fans and their fans sometimes quit.

If you do not like my post feel free to point out your opinion or my errors.

 

[ceh4702]

Sometimes a CD has problems reading also. A little dirt on the old eye and it has all kinds of problems.

If you do not like my post feel free to point out your opinion or my errors.

 

[csab]

Hello Guys,

I've got maybe the same issues as Baudkarma!

My config is the following:
mobo = Asus P4C800 Deluxe
cpu = P4 2,8c
ram = 1 X CorsairValue 512Mb
psu = 400W PFC QTech. (Papst Series)
gpu = Asus Ati 9600 XT

The description that I would make of the problem is the following :

The PC switches off without posting any message on the screen, but I have no freeze. There is no influence if I'm working with it or not...
The temperatures are OK, about 35°C for CPU and 30°C for MoBo so this is not a question of cooling for me.
The psu can't be the problem, because I'm monitoring it also and the voltages are constant (+/- 0.010 V).
Unfortunatly I can not test my config with an other RAM, because I have only one for now...

There is also something I do not understand: What the hell is SMB or SM Bus?
How can I desactivate it? (I've already tried in the control panel -> sytem -> hardware... But each time I restart the PC (after a crash) Windows wants to (re)install this stuff. It's surely a caps/pin problem, but I have looked 3times and it seems to me that the pins are correctly configured...
BaudKarma have you got disfunctionment with this setting? (look in the control panel -> system -> hardware under Windows) ...

Do you think we could have the same problem? And is there a solution? Please BaudKarma if you could post the mark of your RAMs it would be maybe helpfull...
And if you can fix the problem tell me the solution also

Thanks for all

Csab

 

[garebo]

csab, most problems are not the same and few people will read this. YOu should likely post on your own, a new, separate post.
It also seems that baudkarma has found his problem, it being bad ram.

You could use google to find and download a program called memtest or memtest86 and test your memory. You have to have only one piece of ram in at a time to test it.

Also, you can find a setting for win xp to not just turn off the pc if there is a problem. You can change that setting so the pc will stay on and it will give you the error message instead. That would help a lot to find the problem.



Good advice + great people = tek-tips

 

[garebo]

csab, cut and paste this link below for your sm bus info:

http://www.google.ca/search?hl=en&q=sm+bus&meta=

I believe the sm bus is part of the motherboard system files. Do you have a cdrom for your motherboard? If not you can go to the motherboard mfgrs website and download the files and drivers you need.


Good advice + great people = tek-tips

 

[garebo]

Here is some info on the sm bus:

http://www.driverforum.com/controller4/2278.html

http://www.computing.net/drivers/wwwboard/forum/4072.html

http://www.iamnotageek.com/a/114-p1.php

You have an intel motherboard and you need the correct sm bus drivers. You have to go to intel and download the chipset finder and install it. Then it will tell you what chipset you have.
Unless, of course, you have the cdrom for your motherboard. If you do then you have to re-install the chipset drivers from your cdrom.



Good advice + great people = tek-tips

 

[csab]

Thanks Garebo, I am going to post another thread ...

Csab

 

[garebo]

Thanks for the star!
You may not need to after installing the sm bus drivers.

On the other hand, you could possibly have a virus\trojan\worm. Most antivirus progs dont catch everything and they are very bad at trojans\worms. Trojans can and will do what is happening to you, pc on and off all the time.
So here is the cure. Go to trend-micro and allow them to do an online scan of your system. They are very professional and do no harm and a lot of good. They will make sure you have to virus\trojan\worms, etc.
But only do the first scan, not the other 2 as they arent needed, only the first scan is needed. The scan is in 3 parts, only do the first part.




Good advice + great people = tek-tips

 

[BaudKarma]

Update: I plugged all my RAM back in, and have been running and using the system since Monday. No lockups, no problems.

So... either the RAM wasn't quite plugged in properly, or the actual difficulty is the "other issue" I mentioned above. It's a software thing, so I'll do some more checking and testing before I render my verdict.



I try not to let my ignorance prevent me from offering a strong opinion.

 

[ceh4702]

Maybe it was a little bit of corrosion or tarnish on the contacts on the RAM or the slot. Sometimes this has caused problems for people. Often just popping the RAM out and pushing it back in has just enough surface tension to make better contact between the two surfaces. IBM use to always try to use Gold Plated Contacts as a preference. I have seen some in copper and also aluminum in the past. Metal does corrode over time and sometimes fails to make good contact. I have seen suggestions like take a pencil eraser and rub the contacts till they are cleaner, but never tried it.



If you do not like my post feel free to point out your opinion or my errors.