K8V - The Next Step?

  • Thread starter Thread starter Jurgen Chiang
  • Start date Start date
J

Jurgen Chiang

This is a long read. I apologise. And thanks for listening, those who do.
I'm rather desperate for new places to look in troubleshooting the new K8V
system I've put together for myself here, and I've exhausted everything that
occurs to me at this point as far as possibilities go.

It's a new Athlon 64, K8V (not K8V Deluxe, but K8V) system, and from
assembly to now, three weeks later, it's been non-stop crazy instability.
Windows installer errors throughout installation, with initial install
attempts all failing outright (eventually managed a fresh install of WinXP
Pro), then once Windows was installed successfully, crashes to desktop,
random reboots, etc. Whether it's Firefox or its a game, it eventually
crashes. Boots result in odd STOP errors of more or less random nature
about 20% of the time.

The details at present are:
Antec TruePower 550W (also tried Enermax 430W)
Athlon 64 2800+ w/ in-box retail heatsink/fan combo
Asus K8V (non-deluxe)
1GB Infineon DDR400 in 2x512MB DIMMs (also tried Infineon DDR266)
Radeon 9700 Pro
6 Hard Drives
1 CD Burner

------------------
Idea #1: Memory
------------------

So I was thinking maybe it's the memory. I'd been using the memory (2x512MB
Infineon DDR266 chip id HYB25D256800AT-7) in my A7V up until this point, and
it had always been fine, so I was disinclined to replace it in this new
system. But I figured an upgrade to DDR400 would be nice anyway, so I
decided to spring for it in the hopes it would resolve my problems.

Nope. Same problems. Mind you, this was more Infineon RAM, just a
different generation. But Asus's K8V RAM compatibility list includes
Infineon DIMMs with a virtually identical chip id (the manual lists
HY25D256800BT-5B and if I recall correctly mine is HY25D256800BT-5B2), so I
thought I'd try it. This time, I'd be thorough, as I wouldn't be paying for
new DIMMs with the same problem:

So I create a Memtest86 boot floppy and run all tests. No errors. None.
Boot into Windows. Tried a single DIMM, tried underclocking, tried tweaking
voltage. No improvement. And still no errors in Memtest86, but the same
problems in Windows XP. Crash to desktop, reboot,

----------------
Idea #2: Power
----------------

Then I thought, well, maybe it's the power supply. The system's on a good
725 Watt UPS which has served me well over the past six months so it's not
unreliable power from the wall jack, but I figured maybe the PSU couldn't
handle the load.

I have six hard drives and a burner in the system after all, and not only
all those, but the case fans, and the Athlon 64 processor ALL draw on the
12V rail (the Radeon 9700 Pro draws on it as well as the 5V rail). My prior
PSU was an Enermax 430W, which was none too shabby, but six HDs, a vid card
with high power needs and a CPU drawing the same voltage as most of my other
components might be a problem.

So I upgraded my 430W Enermax, which was speced for 20A on the 12V rail to a
550W Antec TruePower which was speced for 30A on the 12V rail. A true king
among power supplies.

Same problems. No change. The BIOS was indicating I was only getting
11.71V-11.76V on the 12V rail, but that's within spec. I used a multimeter
to measure manually and my multimeter's showing not 11.71V but almost
exactly 12V being provided, so all seems well. Couldn't really be a power
issue, that I can tell.

I popped my old DDR266 DIMMs back in just to experiment with the RAM again.
Ran Memtest86. Zero errors. All tests passed perfectly. Same as for the
DDR400.

---------------------
Idea #3: Overheating
---------------------

CPU temperature is 36 degrees celsius. Case temperature is 30 degrees
celsius. That's one possibility down. This is a very well ventilated and
cooled case, with front, rear and side panel fans. But, I had six HDs in
this system, and they were stacked. One possibility: my boot drive was
overheating and experiencing read/write errors. So I rearranged my drives.
Mounted two of them in 5'1/4" bays, so there was going to be no heat buildup
due to stacking. Just the 30 degree case temperature. No change. Still
unstable insanity.

---------------------
Idea #4: ??????????
---------------------

Does anyone have ANY other ideas where I go from here? At present I have NO
evidence whatsoever that there is anything wrong with my components, so I
can't very well return any of them the OEM rationally, but I also can't use
them functionally at present either.

Thanks
 
Hm yep that was a long read so I wont repost it. ;)

Your attempts were pretty much what I would have done as well. That being said
its really sounding like a memory issues. I have the K8V (deluxe) and I know
its a damn picky board about memory (ran in to the exact problems you described
and it was the RAM) so you may want to try testing the memory in a known good
system if you can (I usualy reocmend at least 12 hours test). Barring that, the
harddrive heat thing may be something to try. Strip all but the boot drives out
of the machine and see if maybe that has any effect. Also maybe try the primary
harddrive in a known good system. Hell do it with all the bits and pieces if
you can.

If you have the machine in a bare bones setup, all the componats work fine in a
known good system and its still doing this I'd say try and RMA the board. If
the only problem comes from adding the motherboard to the mix it pretty much
narrows the possiblities.

Good luck
~A
 
Jurgen Chiang said:
This is a long read. I apologise. And thanks for listening, those who
do.
I'm rather desperate for new places to look in troubleshooting the new K8V
system I've put together for myself here, and I've exhausted everything
that
occurs to me at this point as far as possibilities go.
Click to expand...

snip


Try a different graphics card. Some ATI 9700's had a variety of
incompatibilities with at least one ASUS mb.


formerprof
 
I'd put my money on the video card. The manual for the K8V (and K8N as
well) states:

"If installing the ATi 9500 or 9700 Pro Series VGA cards, use only the card
version PN xxx-xxxxx-30 or later, for optimum performance and overclocking
stability."

You didn't say if you were overclocking, but even if you're not I'd be
worried. That statement indicates to me that there is definitely a problem
with older revisions of the 9700 Pro and the K8V/K8N. Out of curiosity,
what revision of the card are you using? I myself have a 9700 Pro and
ordered a K8N the other day, but didn't see that warning until after the
order shipped. I checked the sticker on my card and it shows "P/N
1029420630 070835," which obviously doesn't match the format specified in
the ASUS manual. Unless that first part is supposed to be "P/N
102-94206-30," it's got the right number of characters. I'm wondering if
I'm going to have problems too...

If so I guess I'll just have to upgrade to one of the newer cards.
 
"Jurgen Chiang" said:
This is a long read. I apologise. And thanks for listening, those who do.
I'm rather desperate for new places to look in troubleshooting the new K8V
system I've put together for myself here, and I've exhausted everything that
occurs to me at this point as far as possibilities go.

It's a new Athlon 64, K8V (not K8V Deluxe, but K8V) system, and from
assembly to now, three weeks later, it's been non-stop crazy instability.
Windows installer errors throughout installation, with initial install
attempts all failing outright (eventually managed a fresh install of WinXP
Pro), then once Windows was installed successfully, crashes to desktop,
random reboots, etc. Whether it's Firefox or its a game, it eventually
crashes. Boots result in odd STOP errors of more or less random nature
about 20% of the time.

The details at present are:
Antec TruePower 550W (also tried Enermax 430W)
Athlon 64 2800+ w/ in-box retail heatsink/fan combo
Asus K8V (non-deluxe)
1GB Infineon DDR400 in 2x512MB DIMMs (also tried Infineon DDR266)
Radeon 9700 Pro
6 Hard Drives
1 CD Burner
Click to expand...
<<snip>>

I'll relate a short story for you. I currently own a P4C800-E and have
two sticks of Azenram (cheap stuff with LEI chips on it, and LEI
doesn't sound like a major ram manufacturer).

I've had the P4C800-E for a while, and as I advocate the use of memtest86
on new builds, I ran memtest86 on the P4C800-E and the dual channel ram
during the build. No problems found in an overnight test. And I haven't
had a bit of problems with the P4C800-e since I built it (zero crashes).

I bought a A7N8X-E and a mobile 2600+ to play with. Stuck the cheap
PC3200 ram in it. It passes memtest86 but won't pass Prime95. It gives
an error in 30 minutes or less. I run three copies of Prime95, using
a Knoppix boot disk. The failure seems to be in the same area of ram
each time.

After playing with the A7N8X-E for a week, I put the ram back in the
P4C800-E, and as Knoppix will boot both of them no problem, I repeated
the test on the P4C800-E. Sure enough, it gave an error in Prime95.
It just took 2.5 hours, instead of 30 minutes on the AMD processor.

The moral of the story, is there is more to life than memtest86. It
seems either pattern sensitivity, or the random pattern of the Prime95
data, uncovers more problems than memtest. (By using Knoppix on a
CD disk, I don't have to risk corrupting a hard drive if the
computer is not stable. That is one of the reasons I like using a
read-only Knoppix boot disk as a test source. I put the Linux
version of Mprime on a floppy, keeping the Knoppix image in original
condition, because I am not clever enough to figure out how to
add stuff to the .ISO yet...)

So, after owning my cheap ram for a year, now I see it is in fact
defective :-(

If your ram will pass memtest86, but won't pass Prime95 torture
test (mersenne.org), then return it to your vendor, while there
is still time for an exchange. Maybe eventually you will be lucky
enough to find RAM that works. I'm not as confident about that as
I used to be. (I didn't buy this ram outright - it was removed from
another pre-built system. I wouldn't buy this brand of ram normally.)

Which reminds me, I have to try this test on the older computers now,
to see if any of them will actually pass the test.

Paul
 
You appear to be correct. Swapping the 9700 Pro out for a GF4 4200 has so
far seen problems cease. It's been running for 5 hours with no errors,
reboots or crashes, and a Doom 3 timedemo and KoTOR ran successfully.

I had discounted the video card as a possible cause simply because there
seemed to be no sense in blaming what was a perfectly good, popular vid card
for, of all things, Windows XP install read errors, boot errors and web
browser crashes, but...well..that all went away when I switched out the
video card, so yeah, there does appear to be some MAJOR problem there.

I can't begin to express my thanks for this tip.

The PN on my card is 109-94200-30.

That number consists of the first part of the PN on the sticker, as you
point out, but is printed on the other side of the card as well, right on
the silicon, without the rest of those numbers: just "PN 109-94200-30"

Cheers,
JC
 
Well, it wasn't exactly what I wanted to hear for myself, but I'm glad that
it resolved your problem.
 
Back
Top