BSOD :( Please help!

[WyWyWyWy]

Recently upgraded...

Parhelia
2 sticks of 256MB DDR333 2.5-2.5-2.5
A couple more fans (0.10mA each, running at 12V)
SoundBlaster Live

Since then... quite a lot of BSOD

Parhelia on its own IRQ.
SB Live and LAN card share same IRQ.
DDR running at 266.

The strange thing is... when it is BSOD'd, it doesn't provide info on what caused it. My original suspect was that Parhelia runs too hot. But now I have three 40mm (2 deep 1 thin) blowing at it so it is unlikely. Plus I don't even use a case.

Tried Sandra burn-in for a few hours but no BSOD.
It usually occurs after some gaming (after!), or just suddently when I'm doing important work or browsing.

Notice that before I upgrade, it didn't happen.
I'm running WinXP sp1.

So could this be, Parhelia, Live, DDR or my PSU?

Please advise... many thanks!

[DGhost]

could be any and all of the above.

Is the blue screen giving you *any* error at all? or is it just going by too fast?

If the blue screen is going by too fast, try enabling memory dumping by going into System Properties, Advanced, Startup and Recovery, and either enable a complete memory dump or disable automatic restarting. that should enable you to get the blue screen error, as it *should* have and error associated with it.

[WyWyWyWy]

Yes I know but there is NO error info at all!
It just BSOD and say contact hardware manufacture blah blah blah and no informative text.

Saying that, there was one occassion it said something like IRQL_LESS_THAN_OR_EQUAL but it only happened once.

[The PIT]

Could be anything from faulty memory, memory clocked to hard, other faulty hardware, Crappy drivers. If it's an older Sblive don't install in acpi mode as it doesn'talways play well also try it own irq.
Check that you've got all the latest drivers and Bios. Use the bios defaults as well and see what happens. If still crashes clean install of XP. If it still don't work clean install of XP with the just the video card. Also if your memory is in pairs try one stick at a time and also different sockets.

[Wombat]

What does the Event Log have?

Are you running multi-monitor?

[WyWyWyWy]

When I'm digging through my Event Log... I found this...

Event ID 11: "The driver detected a controller error on \Device\Harddisk0\D."

After some googling, nobody has a definite solution and many NT4, 2K and XP users are having this problem... I don't know if it is related to the BSOD though.

And yes, one 17", one 15" and a TV.

Now I uninstalled the SB Live driver (Creative one) and WindowsUpdated it to a WHQL. I don't know but I just have a feeling that it is the SB Live causing the BSOD, not playing well with my LAN card and my USB ports... which in turn is a mouse and a webcam.

However, if it is not SB Live then it HAS to be the Parhelia running out of spec. No I didn't overclock it, but I got another feeling that 133MHz memory would make PCI/AGP out of spec on this crap motherboard... in that case, I'll run my DDR333@200

[rugger]

Just to be sure though,

You should run IBM DFT on your hard drive. Make sure it isn't failing.

A dieing hard drive will often lock up the bus it is connected to (IDE bus ect) and cause blue screens.

[UtwigMU]

Cheap roundcable EBuyer.com

The driver detected a controller error on \Device\Harddisk0\D

Are they shielded? Try some ribbon cables and see if it makes a difference (You can try other rounded ones, but it will be more expensive just for a test.)

Are you using good PSU? What brand is it? I replaced 128MB PC 133 CAS3 stick with 2x256MB PC133 CAS2 sticks. Due to crappy Codegen PSU I could not pass Memtest86. (Try that to see if ram works). Are the woltages clean?

Is your RAM ok? (Memtest86 passes)?

Are you overclocking?

Try different Chipset drivers.

IRQL NOT LESS OR EQUAL

Originally posted by Hat Monster in Explanation of BSODs
IRQL_NOT_LESS_OR_EQUAL, IRQL_NOT_GREATER_OR_EQUAL, IRQL_NOT_DISPATCH_LEVEL
A kernel mode process or driver attempted to access memory to which it has no permission to access, or should not be accessing at this time, or at this dispatch level. The most common cause is an invalid pointer, either due to bad programming or memory corruption.
What does not cause this are IRQs. This has nothing to do with enabled/disabled ACPI, IRQ sharing or IRQ conflicts. An IRQ is a very distinct entity and is not related to an IRQL.
RAM overclocking is the prime suspect, followed closely by CPU overclocking/overheating. You know what to do. Software can also cause it, usually a bad driver.
For IRQL_NOT_LESS_OR_EQUAL, it may be given in the form "STOP: 0x0000000A (parameter, parameter, parameter, parameter)". This can be decoded, the second parameter shows the IRQL which caused the error. If it is not 2, then the interrupt came from a driver. The fourth number is the address, compare this with the base addresses of the drivers in the driver table on the BSOD screen to find the driver that tripped the system. The third parameter is just the action that was being taken, 0 is read, 1 is write. I cannot find any literature on the first parameter.

For Live/128PCI problems and general info read this article:

http://arstechnica.com/guide/building/bios/bios-1.html

[WyWyWyWy]

Look at what just happened in the last 30mins or so...
1. BSOD while I was shutting down, this time it was 0x0000007F
2. Done IBM DFT, no errors found
3. Removed some fans

And actually I think the IDE cables are alright, coz they've been running a few months now.

My PSU... On my current setup I think I'm using about 300W, from rough calculation. And my PSU is rated 400W... so I think its alright...

Voltage read:
+12.00 - 12.14
- 12.00 - 11.86
+ 5.00 - 5.10
- 5.00 - unreadable
+3.30 - 3.28
They look alright.

System pass all Sandra benchmarks just fine.

NOW MY DISCOVERY.
I found that the first time disk controller error occurs is the day I put my Parhelia in, 30th Oct 2002. And now it looks like it is the Parhelia's problem. But... is there anything I can do?

THANK YOU GUYS.
You guys are of great help! Very much appreciated! Thanks.

EDIT:
The 7F error is a kernel panic --- divide by zero.

And now, I'm going to down-clock my DDR333 to DDR200

EDIT2:
Now installing new SiS IDE & AGP driver... dated 24 Oct 2002

[UtwigMU]

Macase 300W
3,35
12,18
5,08
1,51

I see those voltages all the time and they never fluctuate after bootup is finished.

Quality PSU is very important for system stability.

[Greebe]

it's not the voltages while idling we are concerned with. If you would post up the voltages with like Prime95 runing after an hour. Also note the room ambient temp and cpu core temp.

I've seen this issue before

[Wombat]

I'm still leaning towards a hardware thing as well, but if you'd humor me....

Do you have the latest Parhelia drivers? Could you try running single-head for a little while? There were some issues with &gt<!>1 display on Parhelia when XP-SP1 first came out. This could be some weird manifestation of that.

If you don't have this problem solved by the time you go to bed tonight, please download <A HREF="http://www.memtest86.com/">Memtest86</A> and have it run ALL night. I've seen faulty systems only show an error after a few hours of running it, but a healthy system won't glitch at all.

[WyWyWyWy]

Hi all again

I'm trying to get absolute highest system load before I read the voltage but... there is just too much swapping!

After few hours of attempts... I think I've found out the bastard who is causing it.

And no surprise --- SB Live.

Now I'm running this PC without a sound card, and not a crash. The driver controller error however is still there, it showed up only once since last though. Comparing to every five minutes when I have the SB Live in the PC, it seems to be getting better...

On the other hand, I think it actually could be the PSU problem since SB Live draws too much power? I don't know... because I haven't got a PSU to swap and see, well actually I have but that one is only 250W.

And yes I have Memtest86, will run it overnight before I go to bed!

Back to the drive controller error, it definitely has something to do with Parhelia because of the date of its first appearance. I will take Wombat's suggestion and run single-head for tonight's productive work... which is a pain but anyway. Three (or 3.5) courseworks waiting for me I need absolute stability for these few weeks... see now I don't even have a soundcard

By the way,
room temp is 24C,
mobo temp is 25C,
CPU/min is 38C,
CPU/max is 49C,
HDD1 is 32C,
HDD2 is 34C,
PSU is 28C to 32C,
Parhelia is gotta be over 45C!!! though I cannot check. I can only check with my fingers.

To Mr. Greebe:
So what was the cause? PSU? RAM? heat? IRQ? shielding? WinXP?

I've seen this issue before

Again, many thanks all!!!!

[WyWyWyWy]

Oh and some observations:

1. BSOD usually happen while XP is shutting down, for today at least.
2. Official error code is... 0x0000007F (0x000000D, 0x0000000, 0x0000000, 0x0000000) but no more info available even on Google.
3. If it is a PSU problem, error code can't be the same everytime, can it?
4. I .H.A.T.E. this ECS board and am bidding on a MSI645Ultra now.
5. Is that all SB Live cards are so troublesome? I have some Live Platinums coming soon, already paid

[Wombat]

3. If it is a PSU problem, error code can't be the same everytime, can it?

Yes, it can, but that doesn't mean it's necessarily the cause. I learned a lot about stuff like this, through some interesting experiences. If there is a sequence of events that will cause a high load on the PSU before the sequence finishes, the sequence is very likely to cause the same failure. Also, don't assume that the sequence is only happening once. Tight code loops, or frequent types of memory tranfers, may be executing hundreds of times per second in exactly the same way, but it only takes one bad execution of the loop to get the familiar error message.