Computer hell - help please!

[schwinn8]

Ok, so here's the story on this computer that I have been having tons of issues with over the past month:

Athlon XP 1800+ (133FSB)
Spire WhisperRock IV CPU cooler
MSI KT4AV-L (MS-6712) board (VIA KT400A)
Antec Trupower 430 (I think that's the power level... I know it's an Antec)
PNY Verto Geforce 6200 AGP
Santa Cruz Sound card (onboard sound disabled in BIOS)
Cendyne 52x CDRW (2nd IDE channel, master)
Liteon DVD-ROM (2nd IDE channel, slave)
160GB WD drive (1st IDE, master)
20GB WD drive (1st IDE, slave, currently disconnected)

This computer has run Windows 2000 in the past. Recently, we installed WinXP Pro on it and it was operating, for the most part. We would see occasional lockups, but nothing too frequent or severe. The other curious thing is that when you enter the BIOS setup, the screen flickers (LCDs and CRTs) for about 1 second before going into the BIOS setup. Not sure what this is, but I figured I would mention this. Lastly, there would be times that you would turn on the computer, the BIOS would post, then hang at a black screen and do nothing. You would never see any Windows or other displays during this time - not a flicker or anything. A reset (sometimes a few times) would bring it back to life and boot through. The owner dealt with this situation throughout all of this, as it still allowed him to play his games.

As an upgrade we decided to install an Athlon XP 3000 on the board instead (since it's the fastest supported CPU for this board). Long story short, something went grossly wrong during this process, and the board totally stopped posting at all. MSI provides a "D-bracket" With 4 LEDs to tell you what stage of BIOS it's in, but at no point would it ever even pass the first stage. This was supposedly indicative of a bad CPU, but neither the new CPU nor the old previously working one would allow the board to post. The board was sent to MSI for repair, and has been replaced by MSI, with no further information on what passed/failed.

So, now with the replacement board back, I reinstalled the hardware. The board posts, and it still does the BIOS setup flicker. The old CPU works fine as well, thereby implying that there is nothing wrong with it. So I go ahead and try to run the existing windows installation and get it all squared away, and sure enough we see the after-BIOS black-screen hang again. Determined to fix this problem, I dig deeper.

Fast forward a bit - at this stage, I have deleted the partitions on the main drive (160GB) and have prepared to install WinXP (Retail) from the CD. And I am getting a new slew of issues.

The first issue was a SESSION3_INITIALIZATION error during WinXP setup. This error happens right after the setup program finishes loading the initial drivers and says that it's "Starting up Windows" in the white bar at the bottom of the blue-setup screen. After some reading on the internet, this could be due to bad RAM, bad CDROM, bad HD, so I performed the following:

- Ran 5 hours of memtest86 and another 3 hours of memtest86+ with no issues
- Scan the HD for errors using SMARTUDM (to check SMART status), Spinrite, and WD's Data Lifeguard Tools - no issues found
- Remove the second hard drive from the equation - error persists
- Remove one of the 2 sticks of RAM (256MB each) - error persists
- Remove the other RAM stick, and restore the previous - error persists
- Remove the DVDROM cables - error persists
- Remove the CDRW cables and install with the DVDROM - fixed... the SESSION3 error is no more!
- Just as a check, I connected the DVDROM to the IDE cable at the same place as the CDRW was, and again, no more session3 error.

So, I think I am home free, as it seems there is a bad CDRW here. I allow setup to create the partition and format the drive NTFS (slow mode, to ensure there are no drive errors). That passes through. Setup gets to the point of copying files. The DVDROM drive is noisier now, not sure why... seeks are pretty loud... then they quiet down. Copying files gets through about 38% then stops. DVD stops spinning, and nothing is happening. WTF?!

So, I think, maybe the HD or DVD IDE cables are messed up, and I replace them with 2 brand new cables I have leftover. Repeat the setup process, this time is freezes at 37% during the copy.

As you can see, I have been beating myself up on this machine for quite some time, and I can't seem to find the root problem. I have a few more things I plan to run, but I can't understand what the hell is wrong here. Does anyone have any ideas on what I may be missing?

My current plan is to:
- Swap for a different HD to see if Windows setup can proceed
- Scan the WinXP setup CD to see if there are any errors, just in case
- Try a WinXP OEM CD that I have (that I know works) to see if that one will function properly
- Remove the sound card and try again
- Try another video card
- Try an old Epox KT133A board I have that's working in another machine, but use all the other hardware from this one on that board (not a true apples-apples test, but it's something).

Any help is appreciated... I still *feel* that this MSI board is crap... and I wonder if replacing that will help. I don't want to go buying parts that I don't need, though - I'd like to find the problem and replace that, rather than everything inside.

 

[nebula]

Go into the BIOS and check your CPU temp. make sure that's not overheating. Also take out every non-essential part until you get XP installed, i.e. no sound card, no NIC, only one optical and one hard drive.

 

[FlyingPenguin]

I didn't see you mention anywhere that you tried a different PSU. Next thing I'd try. I used to swear by ANtec myself but I've had issues with their True-power 430. Got one here that, although the rails were fine, apparently spiked the HDD when you powered down the computer. It killed 3 drives before I figured this out.

The flicker before entering BIOS may just be the peculiarity of the mobo, or your monitor having a fit when the video resolution changes.

Hope this helps...

 

[nebula]

FlyingPenguin - Really? I have that PSU! I guess I haven't seen any weirdness yet, been using the PS for a few months now. Maybe you just have a bad unit. Ever contact Antec about it?

 

[schwinn8]

nebula:
Good advice on the remove all other hardware. I figured this wasn't the issue, since I am not dealing with anything "fancy" here - just a file-copy process during windows setup. Still, I will add that to my list - probably the first thing to try.

CPU temps in BIOS are reading about 48C, typically. Doesn't sound like overheating to me.

NIC is disabled in BIOS. I will pull the sound card. I am already down to 1 HD and one DVDROM.

flyingpenguin:
This is the second PSU put into this system. The first was a crappy noname CompUSA brand "500W" PSU which wasn't worth crap. The Antec is certainly heavier and seems stronger. Wouldn't this PSU cause other problems too? I have used this PSU on my main home computer, too, and it worked fine in there...

I thought that the BIOS flicker would be a monitor refresh rate change or something, but I have never seen this happen before on these same monitors (again, LCD and CRT).

 

[schwinn8]

A few more questions:
1) What about APIC - should I enable this in BIOS for WinXP? I have heard (in the past) about issues with APIC... should I be worried?

2) Should I setup XP under "Standard PC" HAL on this board?

3) What about the CPU - is it possible to have a CPU that works (POSTs) but still has lockups and other such issues? I have another AXP3000+ CPU on order, and will be trying that again... hopefully the board doesn't croak like it did last time.

 

[birdpup]

It is mentioned that the NIC is disable in BIOS. If the NIC is on its own PCI card (not familiar with the MB), then the PCI card needs to be removed to properly test the system and find out which hardware component is causing the problems. Leaving the card in place and merely disabling it in the BIOS will not help, it needs to be physically removed from the system with any other cards that may be present. Also, the one DVDROM should also be removed. Everything, except the minimum 1 stick RAM (depends on MB), one hard drive, and all external connections except for monitor, keyboard, and mouse. What is happening is you are trying to troubleshoot the system and find which component is causing the problems. Leaving unneccessary components in place will cause ambiguity by allowing multiple components to be possible causes of the problem.

Also, try updating the bios with the latest and greatest version. Although this could possibly cause more problems if the system hangs in the middle of installing a new bios.

Good luck with it.

 

[schwinn8]

birdpup:
The NIC is onboard, and is disabled in BIOS. There is no separate NIC card installed.

The only non-essential board in there now is the sound card, which I plan to remove for my future operation attempts.

You mentioned that the "one DVDROM should also be removed"... I have to assume you mean the CDRW that was causing the SESSION3 error. That has already been completely disconnected (power and IDE cables), and hence is removed.

But, are you suggesting I remove the currently working DVDROM drive as well? If so, how do I go about installing windows xp without an optical drive?

BIOS is the latest version available (V5.4) , and it was shipped by MSI to me in this fashion.

 

[birdpup]

It seems I was getting confused as to the exact place you are in with your troubleshooting. If the system works fine through the bios POST, then yes, the optical drive needs to be installed for installing windows. You are correct.

And your current plan as described above seems to be good.

To your question regarding bios options and APIC, it would be logical to disable unneccessary bios options when attempting to find or bypass the problem.

 

[schwinn8]

birdpup:
Yeah, I am through BIOS now... on to installing windows. However, your answer about APIC... if I leave it disabled, then I am without it on the windows installation. I cannot enable it later, as it's part of the HAL that windows chooses during setup. This is why I am wondering if it's "safe" to run APIC (and/or ACPI) these days... it used to be poorly supported in the past (Pentium 2/3 era) so I am not sure if it's been "fixed" since then, and hence, whether I should use it. That's why I am asking.

To be more direct - anyone running an MSI board (in general) with APIC and ACPI enabled? Or are you running a forced HAL for "Standard PC"?

 

[BadThad]

Originally posted by: schwinn8
birdpup:
Yeah, I am through BIOS now... on to installing windows. However, your answer about APIC... if I leave it disabled, then I am without it on the windows installation. I cannot enable it later, as it's part of the HAL that windows chooses during setup. This is why I am wondering if it's "safe" to run APIC (and/or ACPI) these days... it used to be poorly supported in the past (Pentium 2/3 era) so I am not sure if it's been "fixed" since then, and hence, whether I should use it. That's why I am asking.

To be more direct - anyone running an MSI board (in general) with APIC and ACPI enabled? Or are you running a forced HAL for "Standard PC"?

Definately want to use ACPI!

 

[schwinn8]

Ok, so I have some success to report. Nebula was right on - removing the sound card allowed the windows install to go through without a hitch. I have never seen this happen before, as I have sucessfully installed WinXP on many machines without having to go through a system strip-down to get it to install... so this is new territory for me. Of course, for that reason, I have a few questions.

Now that WinXP is installed without the soundcard, does this mean that the soundcard is flaky, or is this just a "bug" during windows installation that should never rear it's head during operation? Maybe it was the slot that I installed the card into that causes this specific issue?

Same questions about the CDRW drive that gave me the SESSION3 errors. Do these setup errors mean the drive is flaky, or was it simply some weird issues when coupled with the soundcard (remember, the sound card was installed during those attempts).

I plan to test the reinstall with the CDRW drive reconnected, since I am in an investigative mood. I don't know what kind of outcome to expect... but I just want to find out if the drive was the culprit, or if it was the sound card, or some weird interaction between the two.

So, in summary, now that XP installed:
- Is the sound card flaky or bad? Should I not use it anymore?
- Is the CDRW flaky or bad? Should I not use it anymore?

 

[FlyingPenguin]

If you've tried two PSUs then you've likely eliminated that possibility.

 

[FlyingPenguin]

If you've tried two PSUs then you've likely eliminated that possibility.

As for the sound card I have seen XP installs stall because of a card that XP insisted on installing the wrong generic drivers for. I generally tend to pull all cards and disable most onboard devices for the initial install. Usually goes smoother. There are also the occasional IRQ channel sharing issues, although they're rare these days (SBLive cards were particularly prone to that problem).

The sound card may be fine or it may be flaky. You'll just have to test it. You can always enable the onboard sound and try it with that.

 

[schwinn8]

I could easily accept the sound card driver explanation in most cases, but in this case the drivers weren't even being installed or activated in any way. The lockup would happen during the file-copy that setup performs before the first reboot. So, I don't see how this would affect that function, other than due to an IRQ issue or other conflict, as we both suspect.

I will play around a bit more tonight and see what happens - particularly if I can get the CD to install off the CDRW now that the sound card is removed.

As for testing the sound card - I assume that installing it and seeing it operate is sufficient to say that it's fine? Is there any test software or method I can perform to ensure it's fine beyond this simple process?

It just amazes me to have seen so many issues with this install. I mean, I have installed many other machines, including my own, with all sorts of hardware physically in place (mine had a sound card, modem, PCI NIC, firewire, and SCSI all in there at once) and none of these other systems ever caused an install failure of any kind, nor any operational issues.

 

[nebula]

Originally posted by: schwinn8
I could easily accept the sound card driver explanation in most cases, but in this case the drivers weren't even being installed or activated in any way. The lockup would happen during the file-copy that setup performs before the first reboot. So, I don't see how this would affect that function, other than due to an IRQ issue or other conflict, as we both suspect.

I will play around a bit more tonight and see what happens - particularly if I can get the CD to install off the CDRW now that the sound card is removed.

As for testing the sound card - I assume that installing it and seeing it operate is sufficient to say that it's fine? Is there any test software or method I can perform to ensure it's fine beyond this simple process?

It just amazes me to have seen so many issues with this install. I mean, I have installed many other machines, including my own, with all sorts of hardware physically in place (mine had a sound card, modem, PCI NIC, firewire, and SCSI all in there at once) and none of these other systems ever caused an install failure of any kind, nor any operational issues.

Then maybe this means the sound card is toast. I've seen a bad CDRW pull down an entire system before.

I guess if you can install it and you get nothing weird going on, then I guess you can assume it's ok. Also, if you do have problems, move it to a different PCI slot, maybe the slot is bad.

 

[schwinn8]

So, I am testing more tonight. Left the sound card out and did some investigating, changing only one item at a time, and here's what I find:

- Used the same IDE cable from the DVDROM drive on the CDRW and reattempted to install WinXP. Got another SESSION3 error, so I think this CDRW is toast. It's been behaving kinda weird, so it doesn't surprise me.
- Used the original IDE cable that was in the case on the DVDROM only and tried to reinstall WinXP. Locked up during file copy again, at 38%... what's with this point of the copy?! Anyway, this leads me to believe that cable is flaky, so I leave it out of the system.
- Scanned the WinXP CD to ensure it's not a bad disc (used CDCheck on my other computer). No errors found. Ok, that's good information.
- Reattached the known good IDE cable to the DVDROM, and attached the second drive to the system (primary IDE channel, slave to the XP drive). Retry WinXP install - got a message during install saying that one of the drivers is corrupt on the drive, and if I want to retry... interesting. Told it to go ahead and retry, and then it locks up at about 91%.
- Noticed the video card (GF6200) has it's second power cable disconnected. Don't know if this has anything to do with it, but I reconnect this, and repeat the XP install. Now it locked up at 51%. WTF?!
- Tried installing a known good Acer 52x CDROM I have leftover to replace the DVDROM drive. Repeat install... initial setup "booting" seems faster than with either of the other optical drives. 37%... 38%... at 50% got another error saying "setup cannot copy paqsp.dll" and enter to retry. Retried, and it finally completed.
- Being a glutton for punishment, I wanted to make sure this wasn't a fluke. I replaced the CDROM IDE cable with another brand new, never-used IDE cable and retried the install. Install goes through... ok, maybe it was another bad cable?
- So, once more with the DVDROM drive (I can't believe I have 2 bad IDE cables AND 2 bad drives?! And, I must be bored to want to go through all this, I know!) Reinstall... 80% and another random file gets the "cannot copy" error. Retry doesn't even work, as it' retries then drops back to the same error message. This can't be the drive or the cable's fault... at least not on the optical drives.
- Let's try something different - the power supply. Replaced it with a known good Enermax 300W supply - setup locked up at 51% again.

Throughout all this, I would check the CPU temp through BIOS - highest was 48C, usually saw 46C. Nothing too bad there that I can see. Voltages are all in spec too.

That's enough testing for today. I think the next test sequence will be to try a different video card, and to try my old KT133A motherboard with all this same hardware. I'll be pretty pissed to find out the MSI board is bad... since it was just replaced a week ago under warranty because the previous simply died.

Again, any suggestions and insight are welcome...

 

[nebula]

OK wait, the install went fine with the Acer and the new cable. The way I'm reading it, it sounds like two bad cables and two bad drives. That seems highly unlikely so if you don't want to RMA the board yet, I'd try the combination that worked, the Acer and new cable again, maybe even twice. If XP install goes flawless twice then two bad cables and two bad drives otherwise
I hate to tell you but I'd be RMAing the board.

Do you happen to have an ATA controller card? You could use that instead of the onboard IDE ports and see what happens. I'm thinking one of the mobo IDE ports is bad.

 

[schwinn8]

nebula:

The Acer drive did work once. I then reconnected the slave HD to the mix, and then the Acer did not work. Throughout the testing yesterday, I got some installs that went through (after a copy-retry) and some that didn't. I don't consider a retry event as a success, as it implies there is something wrong with the communications here. Simply assuming it's all fine would be a bad assumption on my part, I would think, as it would lead to other weirdness during normal operation later on.

Also, do note, this is an RMAed board already. The first board died and wouldn't even start to POST... so I was assuming this board was fine. But, I agree, it may still be the board.

I actually do have an IDE card that I could try... I'll add that to the list of trials. Right now I am also going to try a different video card, and possibly my old KT133A MB. It's just a lot of work to take that board out of the old computer that it's currently running on. I may also need to try my known-good working memory from my other computer... maybe this is still a memory issue.

RMAing the board is certainly on my list of possibilities, but I believe if I have to do this, I will be ripping into MSI for sending me a crap board. I may simply buy another replacement board instead... we'll see.

I don't think I will be working on the computer tonight, but I welcome any other suggestions anyone may have.

 

[nebula]

OK sure. From your post I didn't know how far you went into things, thanks for elaborating. And I too wouldn't accept a retry as a success.

You've already done Memtest. I'd try the controller card and if no dice, RMA would be my guess. I've read posts before with people getting a bad part after RMA. Did you RMA to MSI directly or the vendor?

I've never installed XP with a HD on a controller card, do you have to F6?

 

[schwinn8]

I, too have never installed XP with an addon card. Not sure how that works, but I don't think an F6 is needed.

I don't think I will try that install method anyway, as I think my plan to use the old MB will be just as thorough of a comparison of board-to-board. Heck, if I see that the CDRW works again, then that's pretty damning evidence to say the MSI is bad. We'll see what happens... hopefully I will have a good answer after this weekend!

Thanks for your help, nebula and everyone else who has contributed so far!

 

[schwinn8]

Well, it turns out that it was bad memory after all. DAMNIT! I replaced the RAM with the known-good set from my computer (it's the only DDR stick I had around). Sure enough, I went through 5 installations of XP without a hitch. In fact, I was able to reconnect the sound card to the board, and it still installed with no problems (the CDRW is still toast, though).

All this, even though I tested the RAM with memtest86, memtest86+, and I pulled one or the other and swapped slots. Granted, at those times, I also had the CDRW installed, which was its own problem, as we have seen. Still, it's amazing that the failure was so repeatable that it always failed at about the same point - 37-38% of the copy (though, more recently, it was failing at other points, too.

The moral of the story? Don't put too much faith in memtest86. I have used memtest to find bad memory in the past, but I guess there are some situations where it simply doesn't find problems, yet there certainly are issues.

Now, the machine is back to operating state - gotta order some new Crucial memory (the ONLY memory I have NEVER had a failure with). But, at least the machine is finally going to be fine.

Thanks again, everyone, for your help...