Author Topic: Tracking a lockup problem, opinions/ideas please  (Read 657 times)

Offline CavemanJ

  • Silver Member
  • ****
  • Posts: 1008
Tracking a lockup problem, opinions/ideas please
« on: July 23, 2002, 09:52:22 PM »
Ok, I'm thinking I've got this one figured out, but thought I'd ask for opinions from some of you more knowledgable folk.

Back in Feb I built a new system.
AMD Athlon XP 1600+ w/ Thermaltake (or Thermalrite?) SK6 HS w/ 7k delta fan
MSI K7N420 Pro board - nForce chipset
256mb Crucial PC2100
Maxtor 7200rpm 40gb hdd
Teac fdd
52x Sony CD-Rom
Antec case/PSU with 3 intake fans / 2 exhaust fans running (+ PSU fan)

Box ran great for a month, then the temps started rising.  To make a long story short, the HSF wasn't mounted properly, shifted a little when I used the can of air to blow any dust out it, and the chip fried.  Apparently it took the CPU fan header on the mobo with it, as we tried a brand new chip that my friend had and 5-6 fans and none worked.  RMA'ed the board, bought a new chip (1g duron, couldna afford another XP at the time).

Get the board back, install the chip and HSF (properly this time), put it all together, and nothing.  So take it all down cept the chip/HSF, remount everything, double check all connects (had'em right to start with), hit the power, nothing.  Was going to RMA the board again, but missed the window cause I took my time getting the box back up (read had to wait a couple weeks to buy a new chip).
On a side note, I got the impression the board I got back from the RMA was one someone else had returned and they sent it back out.
Ok, so I'm thinking my RAM mighta gone bad.  Got a wild hair, moved the RAM from slot1 to slot3 and hit the power, just to see what would happen.  To my surprise it booted up just as pretty as you please.  Everything seemed to be running along nicely.

Started up EQ, played awhile, system froze up hard.  Screen went black, monitor led still green though, no input devices working.  It was kinda like I had hit the reset button but it stopped as soon as I released the button.  Hit the reset button.  Thought it mighta been just a quirk, it wasn't.  On very rare occasions the screen goes scrambled or just freezes whatever is on it.  Two or three times since the first lock up the machine has rebooted itself whent this happens, but only 2 or 3 times.

Thought it mighta been a software problem, as I had only noticed it in EQ.  Talked with the EQ techs, sent him the tech type files he asked for (dxinfo, an eq log, system info in text format), and only thing he could find was the pop3trap and webtrap of pc cillin might be freezing EQ.  Shut down PC Cillin completely before starting EQ, still get the random lockups.  That was the day I got the first one in AH, then another one in AH about 30min later.  Eliminated it being an EQ software problem.

Updated all the drivers for the hardware/board.  No cure there.

Heat, HS cool to the touch when running under a load.  Checked it playing EQ, AH, and running Prime95 stress test.  Don't think that's the problem.  PC Alert's highest reported temp under load was 42c.  Also ran with the case open and a floor fan pointed into the system, even though my case has massive airflow, just to be sure.  Still locked up.

Tried moving the RAM to different slots.  Still had the lock ups, but now the system will boot with the RAM in slot1.  Weirdness.

Ran Prime95 in stress test mode and left it running when I went to bed.  Woke up in the morning and checked, Prime95 had stopped after 4hr 37min.  Pretty much eliminates any video issues, unless slow scrolling text in a notepad looking window will trigger a video issue.

So now I'm left with the board as my primary suspect from when I first brought the box back up.  But I haven't been able to eliminate the RAM as the cause.  Tried memtest86, but every disk I've tried to boot from gives me the bios initialization error on the fdd (according to memtest86's readme anyway).

So, any thoughts suggestions on anything else to check?  I did add a GF2 GTS to the system instead of running with the onboard video.  My performance improved, but it didna solve the lockups.

Offline bloom25

  • Silver Member
  • ****
  • Posts: 1675
Tracking a lockup problem, opinions/ideas please
« Reply #1 on: July 24, 2002, 12:39:11 AM »
Cave, my guesses would be the board or ram.  It looks like you've ran memtest86, which typically pin-points bad ram, so that's unlikely now.  I'd RMA the board if it were me.  Of the boards I've used, that particular MSI board has proven to be one of the most troublefree I've used, so I'm sure the board is not just inherently unstable.

Offline CavemanJ

  • Silver Member
  • ****
  • Posts: 1008
Tracking a lockup problem, opinions/ideas please
« Reply #2 on: July 24, 2002, 12:52:51 AM »
Bloom I've not gotten memtest86 to run.

Quote
from the memtest86 readme.txt
Sometimes when booting from a floppy disk the following messages scroll up
on the screen:
        X:8000
        AX:0212
        BX:8600
        CX:0201
        DX:0000
This the BIOS reporting floppy disk read errors.  Either re-write or toss
the floppy disk.


That's what I get, and I've tried 5-6 disks around here.  Thought maybe it was the disks cause all the ones I've got laying around are old, and i keep forgetting to a grab a few when I'm out and about.

I love the board, odds are my system would be running sound if we hadna screwed up mounting the HSF the first time.

Like I said above, when I rma'ed the first board the one I got back gave me the impression it was one returned by someone else, and mwave sent it back out (to me unfortunately).
« Last Edit: July 24, 2002, 12:59:02 AM by CavemanJ »

Offline Gunthr

  • Gold Member
  • *****
  • Posts: 3043
      • http://www.dot.squat
Tracking a lockup problem, opinions/ideas please
« Reply #3 on: July 24, 2002, 10:22:05 AM »
Can you throw in a different p/s (400 watts) and see what happens?
"When I speak I put on a mask. When I act, I am forced to take it off."  - Helvetius 18th Century

Offline CavemanJ

  • Silver Member
  • ****
  • Posts: 1008
Tracking a lockup problem, opinions/ideas please
« Reply #4 on: July 24, 2002, 10:27:22 AM »
Quote
Originally posted by Gunthr
Can you throw in a different p/s (400 watts) and see what happens?


Not right now.  Don't have a spare ATX PSU laying around.

Offline Gunthr

  • Gold Member
  • *****
  • Posts: 3043
      • http://www.dot.squat
Tracking a lockup problem, opinions/ideas please
« Reply #5 on: July 24, 2002, 12:03:59 PM »
Its worth it to get one to rule out the p/s as the source of your problem. Its an easy thing to do.

A bad/malfunctioning p/s can cause:

Windows crashes
Memory failure or parity errors
Data drop-out
Data corruption on hard or floppy drives
Premature hard drive failure
Intermittent glitches
Serial communications won't work
Drives won't spin up
Automatic system reset
System won't boot
Fan noise
Chirping sound coming from inside the computer
Mysterious problems you just can't figure out


Sounds to me that a p/s problem is possible based on your computer's behavior.  

Be cautious about how many fans you run off the motherboard... I use a fan buss to get power straight from the P/S rather than overtax the mobo...

G'luck
« Last Edit: July 24, 2002, 03:33:19 PM by Gunthr »
"When I speak I put on a mask. When I act, I am forced to take it off."  - Helvetius 18th Century

Offline bloom25

  • Silver Member
  • ****
  • Posts: 1675
Tracking a lockup problem, opinions/ideas please
« Reply #6 on: July 25, 2002, 01:47:56 AM »
Cave, you didn't mention what size powersupply you had, but honestly I don't find that likely.  A 300W Antec supply is more than sufficient (if that's what you have), but you could load up MSI's probe software and watch the voltage levels to be sure.

I'm using a 340W Enlight personally and it's more than adequate.

The system I built that is nearly likely yours had a 300W Sparkle power supply and has been perfect.  (I've never been called about a problem with that system, and it's been running about 8 months now.)

Offline bloom25

  • Silver Member
  • ****
  • Posts: 1675
Tracking a lockup problem, opinions/ideas please
« Reply #7 on: July 25, 2002, 01:51:37 AM »
Get a new floppy disk, I too had to try 3 to get memtest to run.  (I just don't use floppys that often. ;) )

Offline CavemanJ

  • Silver Member
  • ****
  • Posts: 1008
Tracking a lockup problem, opinions/ideas please
« Reply #8 on: July 25, 2002, 04:07:32 AM »
Bloom, I guess I did forget that.  I bought an Antec SX830 case for the system, with a 300w PSU in it.

The only 2 things not listed in the original post that are in the box are a GF2 GTS and a Motorola SM56 modem.  I do use the onboard lan port, and at present my machine is net connection for my LAN (no broadband yet :mad: )

I keep forgetting to grab a thing of disks when we go out.  Gonna have to cover the monitor in post-its I guess :D

I watched the voltages in PC Alert for a bit and they all seem pretty stable.  Vartions of +/- .04 on the +5v, 3.3v, 5v standby.  Don't remember seeing the battery move.  +12v and -12v were had the largest changes, at up to .1 in the varience.  Vcore mainly stayed at 1.76, with an occasional dip to 1.74 or rise to 1.78.  Dinnae how much of a difference those variations might make, but they're right in line with reported readings when the box was running the XP 1600+ save the Vcore, which was a hair lower with the XP chip.  Also, there were NO problems of any kind with the system until I disturbed the heatsink's precarious seating and led to the frying of the chip.  The problems started after getting the board back from RMA.
« Last Edit: July 25, 2002, 04:16:48 AM by CavemanJ »

Offline bloom25

  • Silver Member
  • ****
  • Posts: 1675
Tracking a lockup problem, opinions/ideas please
« Reply #9 on: July 25, 2002, 04:53:03 AM »
Those voltages and variances are excellent.  (BTW:  -12V isn't used for anything anymore, so don't worry about it.)

The lockups really bug me.  That board and my new board (Asus A7N266-C) are without a doubt some of the most stable boards I've ever used.  My system has yet to lockup in 3 weeks and I run it very hard occasionally.

When I did the system with an MSI K7N420 board I ran a burn in tool for 5 hours and played some games on it and it was solid too.

The newest version of memtest 86 is 3.0, and I just messed with it on my computer and it seems to work fine.  It took 20 minutes to complete with 512 MB of DDR PC2100 ram.

Offline weazel

  • Parolee
  • Silver Member
  • ****
  • Posts: 1471
Tracking a lockup problem, opinions/ideas please
« Reply #10 on: July 27, 2002, 05:47:10 PM »
I'll bet you didn't disable the 10/100 NIC in the BIOS before installing Windoze.

Offline CavemanJ

  • Silver Member
  • ****
  • Posts: 1008
Tracking a lockup problem, opinions/ideas please
« Reply #11 on: July 27, 2002, 06:18:18 PM »
Quote
Originally posted by weazel
I'll bet you didn't disable the 10/100 NIC in the BIOS before installing Windoze.
 why would I want to do that since I intend to use it?

sides, it's a hardware problem, not a software one.

The question is, RAM, PSU, or board

Offline weazel

  • Parolee
  • Silver Member
  • ****
  • Posts: 1471
OK.....if you say so
« Reply #12 on: July 28, 2002, 01:49:25 AM »
"why would I want to do that since I intend to use it?

sides, it's a hardware problem, not a software one."


Want to bet?   ;)

The reason is because Client for Microsoft Networks detects your NIC and *will* attempt to connect to the network while setup is ongoing.

The SiS 900 NIC on my mobo gave me hell until I figured out what was happening,.... corrupt vxd registry entries all related to MSN.

If you have an extra hard drive do an install with the NIC disabled until it's complete, then enable it in the BIOs and install the drivers.

Offline DAVENRINO

  • Silver Member
  • ****
  • Posts: 1084
Tracking a lockup problem, opinions/ideas please
« Reply #13 on: July 28, 2002, 03:21:37 AM »
weazel,
I have the same chipset as Cave and did not disable NIC in BIOS before Win XP Pro install.  No problems with my box or the NIC which I use for DSL connect.  I have had some lockups and discos last few days in AH but so have a LOT of other folks.  Also, as far as I know, MSI is the only nforce board that actually uses the Nforce LAN instead of a third party LAN.  My Guam PC has an Asus Nforce board but no OB LAN.
DAVE aka DJ229-AIR MAFIA
CH USB HOTAS/ONKYO 705 7.2 SURROUND SOUND/ 60" SONY A3000 SXRD  TV

Offline CavemanJ

  • Silver Member
  • ****
  • Posts: 1008
Re: OK.....if you say so
« Reply #14 on: July 28, 2002, 09:44:47 AM »
Quote
Originally posted by weazel
"why would I want to do that since I intend to use it?

sides, it's a hardware problem, not a software one."


Want to bet?   ;)


Yes, I do.  The only real changes to the system from when it was running fine to when the lockups started were going from an XP 1600+ to a 1g duron and the motherboard, which was rma'ed.  Did you actually read all of the original post?  I have had zero problems with my net connections in general, either LAN or to the i-net.  Unless ya wanna count the wife squeaking about losing her connect when my box locks up :D

Davenrino this isn't an AH lockup, it's a total system lockup that mainly happens while playing EverQuest and occasionally happens playing AH (don't really play anything else :D )