Gigabyte K8VM800M - What's causing this video problem?

[Sparkey55]

Re: Gigabyte K8VM800M - What's causing this video problem?

I once had a socket A board made by Megatrend Axper K7V400 that was quite thinner then most boards. It would warp just by looking at it. I "fixed it" by putting a hard rubber spacer about 1/8" taller than the chassis spacers under the center of the board between the CPU and the Northbridge. The thing worked fine after this fix. One of those cheap Tiger Direct Specials.

[socketa]

Re: Gigabyte K8VM800M - What's causing this video problem?

Looks like the good video result was just a coincidence.
Tried it couple more times with pressure, with faulty video result
About a dozen more times without touching it, and two times it went OK.
Tried wobbling the board on the edge of the box, as it was running that checksum app, and couldn't get it to fault.
Visually checked the components around the NB, and between it and the agp port, and they all looked good.
The rubber spacer is a good idea - i'll keep that in mind if i ever install 'my' bent socket LGA 775 ASUS board.

[momaka]

Re: Gigabyte K8VM800M - What's causing this video problem?

Try running the board flat on a table, with the CPU heatsink still installed on the CPU but with the retention mechanism disengaged. AMD's stock retention mechanism for socket 754, 939, AM2/3+ have a tendency to warp the board under the CPU quite a bit in some cases. Same goes for the stock Intel socket 478 retention mechanism as well as the plastic push-twist pins mechanism used on socket 775 and later.

That said, problems with IDE channel 1 could also suggest SB BGA issues. In general, VIA chipsets run very cool, so if there is any BGA issues, they probably weren't caused by heat.

And artifacts on video output suggests that the link between NB IGP and RAM is somehow getting interuped - either physically (i.e. bad connection in RAM slots or bad BGA under NB) or electrically (bad caps or something else causing noise).

If you have any 3D games (that can actually run on this VIA IGP ), give them a try as well to see if you get any problems.

[socketa]

Re: Gigabyte K8VM800M - What's causing this video problem?

Released the retainer, but problem still remains.
These don't seem like artifacts, as i've seen them - the entire picture is always instantly replaced with vertical stripes that all have horizontal bars through them.
Checked the couple of mosfets in between the NB and RAM - no change in voltages from good to faulty state.
And it would very likely have to be a RAM controller (if that exists outside of the RAM stick) as opposed to RAM slots, because the problem occurs with one stick of RAM in either slot.
Unlikely to be bad 'lytics because the smaller ones are the same series as the output caps which were sitting next-door to the CPU heatsink, of which all tested good.
It's noteworthy that the issue usually surfaces quickly when running a checksum program on a big file, where the CPU usage is less than 10%; but while playing three videos simultaneously, with the CPU usage at 100%, the video usually doesn't fault in a similar time period.
It's almost like that particular application is accessing a particular component (that has become prone to faulting) more often than the media player application.

[Sparkey55]

Re: Gigabyte K8VM800M - What's causing this video problem?

Released the retainer, but problem still remains.
These don't seem like artifacts, as i've seen them - the entire picture is always instantly replaced with vertical stripes that all have horizontal bars through them.
Checked the couple of mosfets in between the NB and RAM - no change in voltages from good to faulty state.
And it would very likely have to be a RAM controller (if that exists outside of the RAM stick) as opposed to RAM slots, because the problem occurs with one stick of RAM in either slot.
Unlikely to be bad 'lytics because the smaller ones are the same series as the output caps which were sitting next-door to the CPU heatsink, of which all tested good.
It's noteworthy that the issue usually surfaces quickly when running a checksum program on a big file, where the CPU usage is less than 10%; but while playing three videos simultaneously, with the CPU usage at 100%, the video usually doesn't fault in a similar time period.
It's almost like that particular application is accessing a particular component (that has become prone to faulting) more often than the media player application.

If the motherboard has any KZG or KZJ caps still on it then all bets are off
as to WHAT could be wrong with it. Get rid of them and put some good HM or HN back on but check the dates, no 2004 or older. Personally I replace caps that are around heat sources with ones that are a step up in microfarad ratings just to be on the safe side. Also squirt some good quick evaporating contact cleaner into the mem slots then let dry. Hell just clean ALL the slots with cleaner. Change the grease under the NB and no overclocking, try other dimms that are known to work well at lower frequencies, higher capacity power supply if available.
Also I do have some Gigabyte boards that have factory installed Nichicon HM with date codes of 5304, yes that is correct 53 week in 2004. Seems that the plant caught them in time to run a new batch of HM caps for the OE market.

[momaka]

Re: Gigabyte K8VM800M - What's causing this video problem?

Released the retainer, but problem still remains.
These don't seem like artifacts, as i've seen them - the entire picture is always instantly replaced with vertical stripes that all have horizontal bars through them.

Hmm, I re-read your first post, and you said the problem was present both with a discrete and onboard video card. I guess this rules out the onboard video being bad.

The only thing I can think of is perhaps the memory is somehow getting corrupted at a specific block, and maybe that block happens to be where Windows stores the video memory. Seems to agree with your theory as well, where you said, "And it would very likely have to be a RAM controller (if that exists outside of the RAM stick) as opposed to RAM slots, because the problem occurs with one stick of RAM in either slot."

I don't remember if the memory controller for socket 754 CPUs is in the CPU or the NB. For socket 462, it's the NB (you can see the board traces from RAM going to the NB), and for 939 it's the CPU (you can see the board traces for RAM going directly to CPU). But you can check by seeing where the board traces from RAM go to (either NB or CPU).

Anyways, you can also try booting a Linux live CD and running the same (or similar) benchmark to see if the problem persists. If not, maybe try putting back the discrete video card and play some 3D games just to see what happens. If you don't own any games, you can try free demos. I think Need For Speed Underground 2 demo (or other games from that vintage) will run okay on this system. An alternative to that is 3DMark 2001/2002/2003. Either way, play around with the resolutions and settings to see if anything triggers this.

It's noteworthy that the issue usually surfaces quickly when running a checksum program on a big file, where the CPU usage is less than 10%; but while playing three videos simultaneously, with the CPU usage at 100%, the video usually doesn't fault in a similar time period.
It's almost like that particular application is accessing a particular component (that has become prone to faulting) more often than the media player application.

That might suggest problems with the Southbridge <-> IDE drive communication. You said you already have problems booting from IDE1. I wonder if this is actually "the problem" itself, or just an additional problem.

Can you check the hard drive's SMART status and post a log here? In particular, looking for anything under "Ultra DMA CRC error count". I use HDTune usually.

[ChaosLegionnaire]

Re: Gigabyte K8VM800M - What's causing this video problem?

socket 754 is simply the single channel memory version of the athlon 64. socket 939 is the dual channel version. all athlon 64s regardless of the socket have the mch in the cpu.

anyway, i was wondering if this issue could be caused by a bad smd mlcc decoupling cap not filtering high frequency noise properly and thus causing bits to flip wrongly?

[momaka]

Re: Gigabyte K8VM800M - What's causing this video problem?

socket 754 is simply the single channel memory version of the athlon 64. socket 939 is the dual channel version. all athlon 64s regardless of the socket have the mch in the cpu.

Cool, thanks for clarifying.

anyway, i was wondering if this issue could be caused by a bad smd mlcc decoupling cap not filtering high frequency noise properly and thus causing bits to flip wrongly?

Nah, when those SMD ceramic caps go bad, they usually go short-circuit. Would be hard not to notice it - either the PSU will shut down due to overload or the cap will get smoke
Haven't seen them fail any other way.

[socketa]

Re: Gigabyte K8VM800M - What's causing this video problem?

I think that, at least part of, the IDE problem is related to the cables (and also whenever the cable is swapped the LBA selection was wiped from the BIOS and had to be reset - which i figured out after some time)
Have tried 3 different Cable Select cables. (and yes, the jumper is on CS position)
Can't see any blown SMD caps (the only one that i've seen, an another board, made a crater)
Yeah, the traces from the memory slots are going to the CPU.
Booted from BartPE DVD, and ran the same checksum app on those two video files and the video went faulty (tried this several times)
Booted from a LUbuntu DVD and ran 2 full passes of Memtest - no errors were detected.
Rebooted into LUbuntu (Live), and the video faulted while changing directories in the terminal window, but after rebooting it played video for half an hour until i stopped it.
Have attached SMART info below, and note that there is errors in the Ultra DMA CRC error count that you mentioned
Consequently, just had a good idea! - i'm going to replace the PATA drive with a SATA drive and perform the same checksum procedure - If video doesn't fault then that would pinpoint the fault to the South bridge and it's IDE output. And if it does fault, then that might also mean that there is a problem with the SB or before it?

[socketa]

Re: Gigabyte K8VM800M - What's causing this video problem?

SMART info:

[momaka]

Re: Gigabyte K8VM800M - What's causing this video problem?

Have tried 3 different Cable Select cables. (and yes, the jumper is on CS position)

Why not also try forcing Master and Slave.

Can't see any blown SMD caps (the only one that i've seen, an another board, made a crater)

Yup, that's how they go.

Have attached SMART info below, and note that there is errors in the Ultra DMA CRC error count that you mentioned

So that means likely you do really have a bad cable and/or connection from SB to IDE.

I am also going to guess with some confidence that your HDD is a Seagate. Normally, Hardware ECC recovered could suggest drive problems too. But on Seagate drives, Raw Rear Error Rate, Seek Error Rate, and Hardware ECC recovered are Seagate-specific parameters that always show high values.

Consequently, just had a good idea! - i'm going to replace the PATA drive with a SATA drive and perform the same checksum procedure

Good idea indeed. Keep us posted.

[Sparkey55]

Re: Gigabyte K8VM800M - What's causing this video problem?

Would this be the same motherboard that you had the bad CPU socket in awhile back?

[socketa]

Re: Gigabyte K8VM800M - What's causing this video problem?

No it's not that board (i've got 3 of these)
(i got a replacement socket for that one but it was BGA with solder balls attached, but the one attached to the board looks like it has pins)

Spent over a week on this.
Couldn't get XP x64 to load on IDE - BSOD etc.
Couldn't get a backup (from same model board) to be successfully installed on IDE or SATA.
Finally installed XP x64 on SATA drive and couldn't get it to fault while running my checksum app on those two video files that are located on the IDE drive - So it appeard that the SATA ports were OK
But then i removed the IDE drive and just left the SATA drive attached - now the SATA drive refuses to run windows (it gets to the XP logo screen then automatically reboots, over and over)
Moving the SATA cable to the other motherboard SATA connector made no difference.
Removed the SATA drive and connected it to the other (same model) good board, and get the same issue.
It's almost as if removing the IDE drive caused the SATA drive to become corrupted.

[Sparkey55]

Re: Gigabyte K8VM800M - What's causing this video problem?

No it's not that board (i've got 3 of these)
(i got a replacement socket for that one but it was BGA with solder balls attached, but the one attached to the board looks like it has pins)

Spent over a week on this.
Couldn't get XP x64 to load on IDE - BSOD etc.
Couldn't get a backup (from same model board) to be successfully installed on IDE or SATA.
Finally installed XP x64 on SATA drive and couldn't get it to fault while running my checksum app on those two video files that are located on the IDE drive - So it appeard that the SATA ports were OK
But then i removed the IDE drive and just left the SATA drive attached - now the SATA drive refuses to run windows (it gets to the XP logo screen then automatically reboots, over and over)
Moving the SATA cable to the other motherboard SATA connector made no difference.
Removed the SATA drive and connected it to the other (same model) good board, and get the same issue.
It's almost as if removing the IDE drive caused the SATA drive to become corrupted.

I had a PcChips motherboard that had similar problems like that. After reading the PCCHIPS Lottery forum what worked was reflashing the bios using Uniflash and forcing a complete update even the boot block area, tha fixed it.

[momaka]

Re: Gigabyte K8VM800M - What's causing this video problem?

Finally installed XP x64 on SATA drive and couldn't get it to fault while running my checksum app on those two video files that are located on the IDE drive - So it appeard that the SATA ports were OK

Well, that's certainly good news. Now at least we know the RAM<->CPU<->NB path is okay.

But then i removed the IDE drive and just left the SATA drive attached - now the SATA drive refuses to run windows (it gets to the XP logo screen then automatically reboots, over and over)

Disable automatic reboot to see what kind of BSOD message you get. Also try booting into Safe Mode (with or without Networking, doesn't matter). To get to the full Safe Mode boot menu, simply keep hitting the "F8" key after the POST screen disappears.

Moving the SATA cable to the other motherboard SATA connector made no difference.
Removed the SATA drive and connected it to the other (same model) good board, and get the same issue.
It's almost as if removing the IDE drive caused the SATA drive to become corrupted.

Could be just something Windows is not happy about when the IDE drive is removed. Try installing Windows on the SATA drive again without the IDE HDD attached. Better yet, if you have a SATA CD/DVD drive, try installing Windows from that and completely disable the motherboard's IDE controller.

[socketa]

Re: Gigabyte K8VM800M - What's causing this video problem?

Removed the IDE drives and used fail safe BIOS settings
Then plugged SATA DVD drive into SATA 0 motherboard connection, and SATA hard drive into SATA 1 motherboard connection.
Here is what happened upon each time i pressed the reset button.
Y="Boot from CD drive" then, "Press any key to boot from CD" (then i pressed reset button to continue testing)
N="Boot from CD drive", (and booting halts there, then i pressed reset button)
NYYNNNNNNY
YNNYYNYNNN
NNNNNNNNNN
YNNNNNNYNN
NNNNNNNYNN
NNNNYNNNNNY

i.e.
N = Failed to move on from "Boot from CD drive" message to "Press any key to boot from CD" mesage

Then i swapped the drives around - Still the same semi-random pattern.

Plugging back in the IDE drives made no difference, nor did disabling IDE1, IDE2, and IDE DMA transfer access.

Another thing that i've noticed is that over the past week when cold powering it down by holding in the power button, the computer would shut down and then beep to start up again while the button was still being held down, but then completely shut down after a couple of seconds.
(My other machine doesn't do that)
Now it seems to not be doing that as much, but rather shutting down without trying to start up again.

There is still one KZG on the VRM input (two were replaced with new HM's, as well as the three output caps), so what are the odds that this is causing the problem? (since the other 2 VRM input KZG's tested good ESR and capacitance)

[Sparkey55]

Re: Gigabyte K8VM800M - What's causing this video problem?

Removed the IDE drives and used fail safe BIOS settings
Then plugged SATA DVD drive into SATA 0 motherboard connection, and SATA hard drive into SATA 1 motherboard connection.
Here is what happened upon each time i pressed the reset button.
Y="Boot from CD drive" then, "Press any key to boot from CD" (then i pressed reset button to continue testing)
N="Boot from CD drive", (and booting halts there, then i pressed reset button)
NYYNNNNNNY
YNNYYNYNNN
NNNNNNNNNN
YNNNNNNYNN
NNNNNNNYNN
NNNNYNNNNNY

i.e.
N = Failed to move on from "Boot from CD drive" message to "Press any key to boot from CD" mesage

Then i swapped the drives around - Still the same semi-random pattern.

Plugging back in the IDE drives made no difference, nor did disabling IDE1, IDE2, and IDE DMA transfer access.

Another thing that i've noticed is that over the past week when cold powering it down by holding in the power button, the computer would shut down and then beep to start up again while the button was still being held down, but then completely shut down after a couple of seconds.
(My other machine doesn't do that)
Now it seems to not be doing that as much, but rather shutting down without trying to start up again.

There is still one KZG on the VRM input (two were replaced with new HM's, as well as the three output caps), so what are the odds that this is causing the problem? (since the other 2 VRM input KZG's tested good ESR and capacitance)

I do not understand what you are saying with the whole N Y thing but the first boot device should be on sata0, sata hdd and the sata odd on sata1. If you want to change the boot order use one of the hot keys at startup. Pressing th RESET switch is not what you should do to boot from the cd drive, use the SPACE BAR key instead. Of course this is if you have a bootable disc in the drive to begin with. Like I posted earlier all KZG/KZJ need to be replaced otherwise the inputs and outputs to the VRMs are out of balance and an electrical noise situation can develope, not good for filtering. Also try changing out the POWER BUTTON that goes to the board. I have had problems with those tiny switchs before, either shorting or going open or just double switching.

[momaka]

Re: Gigabyte K8VM800M - What's causing this video problem?

There is still one KZG on the VRM input (two were replaced with new HM's, as well as the three output caps), so what are the odds that this is causing the problem? (since the other 2 VRM input KZG's tested good ESR and capacitance)

Possible, but unlikely IMO. With two good caps (Nichicon HMs) out of three caps on the VRM high side, that should be more than enough to "hold the front", so to speak.

Just follow Sparkey55's advice above, and connect the HDD to SATA 0 and ODD to SATA 1, then set boot device order accordingly so that you can boot from CD (and/or diskette too, if you like ) and install Windows again.

[socketa]

Re: Gigabyte K8VM800M - What's causing this video problem?

Ok
The drives are plugged in as requested.
(As previously mentioned) After several attempts at booting and 'freezing' at "Boot from CD message", i finally get it to move on to "Press any key to boot from CD...."
So i push any key, and the screen doesn't change from the "Press any key to boot from CD.." Message
The computer recognizes the keyboard input because as soon as i push any key, the "...." stop.
The SATA DVD drive is good, because i checked it with my other machine, and the DVD is good too.

Same problem occurs if i swap the drives, or disable IDE channels.

[Sparkey55]

Re: Gigabyte K8VM800M - What's causing this video problem?

This is now starting to look like a memory controller issue. What I mean is a problem physically between the mem slots and the cpu socket or pins, could also be electrical in nature too. Based on what you described it is booting properly but loses the sequence just at the point where the ODD boot info would be loaded into RAM, hence the freeze.