Problems with newly built Gigabyte GA-MA770-DS3

I recently built a new machine with a Gigabyte GA-M770-DS3 board. It
is running XP and was stable for a week or so but now has major
problems and I am not sure where else to look for the problem. I need
to make sure that I find it soon before the Newegg return policy runs
out. Any suggestions would be appreciated.

The hardware is listed below. I put the machine together with no
problems and XP installed easily. Once the drivers, etc. were
installed from the Gigabyte CD, I then connected to windowsupdate to
get all the latest patches. The computer is running pretty basic
stuff: Win XP Pro SP2, Office 2003, etc. However, based on the latest
behavior I's say I can definitely rule out a software problem.

I first noticed the problem a few days ago. I was in the middle of
watching a video that I had downloaded from the pay service at
Amazon.com. About 20 minutes into the video I got a BSOD.
Unfortunately, the blue screen only stayed for a second or two and
then the machine immediately started rebooting. A little later the
same thing happened. After the next reboot it was stable for a few
hours. I went to bed and when I got up in the morning it had tried to
reboot again but it said that it couldn't boot because one of th
esystem files was missing or corrupt.

At this point I assumed that I had gotten a bad hard drive. I shut the
machine off in order to wait until I had more time to deal with it on
the weekend. When I powered it up on the weekend it booted just fine.
Since I had already downloaded the Western Digital drive tester I
decided to run a full test. I watched it for about 1/2 hour and it ran
fine. I left for a few minutes and when I came back the machine had
tried to reboot again. However, this time the BIOS screen was all
messed up. The first several lines (video card, BIOS version, etc.)
were normal but then there were some "garbage " characters and a
couple of lines of inverse video and then more garbage.

I shut the machine down for a few hours and when I powered it back up
it booted fine again.

Normally I would immediately suspect a heat problem with this kind of
behavior. I downloaded the Gigabyte EasyTune Pro in order to monitor
temps and fan speeds. And BTW, I don't do any overclocking. The fans
were normal and the CPU temp never got above 110F. Normally it stayed
around 105F. I opened up the box and reseated the memory and all the
connectors but the problem is still the same. The last time it
rebooted, I didn't get the garbage characters but it hung just before
the device detection.

Here's the config:

- Gigabyte GA-MA770-DS3 motherboard
- XFX GeForce 8600GT 256MB PCI X16 video card
- Antec Earthwatts EA500 power supply
- AMD Opteron 1218 Santa Ana 2.6GHz dual core processor
- 2 x 1GB Crucial Ballistix Tracer PC2 8500 memory
- 3 Western Digital Caviar SE WD800JD 80GB 7200 RPM SATA 3.0Gb/s Hard
Drive
- 1 Seagate 500MB SATA 3.0Gb/s Hard Drive
- SAMSUNG 20X DVD±R DVD Burner
- SABRENT CRW-UINB 52-in-1 USB 2.0 Internal Card Reader

I am not running RAID. The BIOS settings are mostly default other than
enabling USB keyboard support.

Any suggestions would be appreciated. Thanks.

 

Set Windows startup and recovery to not automatically reboot after a
crash.

I would start by measuring the power supply voltages to make sure
they're within specs.
Next I would run memtest86+ to check system memory.
After that, I may try a different graphics card.

 

Thanks for the suggestions. I'll run the memtest program tonight and
also change the reboot option. What do you recommend for measuring the
power supply? I did download Easy Tune in order to check the temp,
fans and voltages. The voltages seemed OK. Should I use a voltmeter?

 

In message

I would go along with Andy and suspect a marginal PSU.

Have you got the OS set up to stop the hard drives when they are idle
for long periods? The crashes seem to happen when the system is
unattended, or you were watching a movie (no KB or Mouse activity).

It may be that the PSU is fine when cold and can cope with the start up
load of the disks; hot it can cope with the steady state load. However
if a drive spins down once things are hot, the re-start load may be
enough to tip the voltages over the edge.

If the OS spins down the disks when idle, it may decide after a period
of time to spin them up (to do something for its own internal reasons).
Under those circumstances the system may simply crash when apparently
sitting idle. In fact its more likely to crash when apparently idle than
under load.

You will probably never see that without some form or recording
voltmeter, however one check you might do is to temporarily disconnect
one (or two) of the hard drives (power and SATA) or the optical. If the
machine is then stable, its probably the PSU that is marginal.

Regards

 

I powered up the machine last night. According to the Easy Tune
software, all the voltages, temps and fan speeds were good. I do not
have the machine set up for any kind of power saving scheme like
powering down the hard drives, etc. I even have hibernation disabled.

The first time I ran the memtest86 program, it ran for around 1/2 hour
without a problem so I then decided to boot it up and try some other
stuff. It ran fine for several hours doing videos, web browsing, etc..
I then shut it down and ran memtest again. After a short time (I don't
know exactly how long), the test halted.

I took a picture of the screen:

http://i302.photobucket.com/albums/nn109/mcbill20/IMG_1212.jpg


This is very frustrating. I don't know whether or not to keep
troubleshooting or just RMA the power supply, motherboard, video card
and memory and start over.

 

In message

As I suggested, try your tests with one or more hard drives
disconnected.

If it is a marginal PSU, then it may work on the limit, and fail when
something (anything) drives up the load slightly, or the mains power
drops slightly, or even a decent noise spike.

If it works with one or two disks disconnected, RMA the PSU - it does
have enough power available on ALL the lines doesn't it? Don't forget
to add in the case fans when computing the 12v load.

Regards

 

Problem solved!

It was the memory apparently. I had gotten an RMA for all the major
parts-- video card, motherboard, memory and power supply but I decided
to try one more thing before I went to all that trouble. I stopped at
a local store and picked up some memory (Corsair this time, not
Crucial) and installed it. Memtest86 ran fine so I then booted it.
Later I did some stress testing with Fresh Devices benchmark programs
and then I used Nero to convert an hour long avi file to DVD.
Everything went fine.

Just an FYI--

I had earlier tried adjusting the memory voltage as some had
suggested. The BIOS did not allow it to go as high as 2.2v as was
suggested. The max it would allow was +3.75 volts. At first, I tried
just one extra volt to see if that would make a difference. The board
would not post with the extra volt but luckily it detected it and had
an option to restore to last know good without having to reset the
entire BIOS.

I've also been monitoring the pwoer supply with te Easytune utility.
All the voltages still look fine. As for temps, the highest the
processor has gotten was 113 and that was during the video rendering.

Anyway, thanks for all the suggestions. I appreciate the help.

 

wrote:

[snip]

Get the current version of Memtest86+: it's v2.01 at:
http://www.memtest.org/

Most high performance memory requires more than the default DDR2 voltage
of 1.8V. Go into the BIOS, memory settings and increase the voltage to
2.2V, as that is what is called for for that memory. If already set to
2.2V, increase to 2.3V and run Memtest86+ again.

Link for product:
http://www.newegg.com/Product/Product.aspx?Item=N82E16820148070

* Click on Specifications Tab and look at voltage required. *

 

wrote:

[snip]

This makes no sense. "The BIOS did not allow the voltage to go to
2.2V...The max allowed was +3.75V." First, 3.75V is > 2.2V, so I don't
understand the problem.

Second, you purchased the wrong memory for your motherboard. The
Crucial memory specs state it requires 2.2V to operate correctly, yet
you installed it into a MB that can't supply that voltage.

Replacing the memory with another type that requires a lower voltage is,
in your case, the only way to fix the problem (other than getting a MB
that can provide spec'ed voltage).

In any event, it's running stable, which is all that is really
important. However, when building a system, be sure to check that all
components are compatible with all others.

 

Oops...sorry...I just can't type today. It was a long night with
server failures here. I meant +.375 volts was the max. And the memory
used to be listed as compatible with that motherboard. According to
what I read in another forum, it was just taken off the list. Since I
don't overclock I didn't look at any of the specific details; I just
figured I'd get name brand 1066 DDR2 memory since that's what the
motherboard specifications showed.

 

Yeah, that decimal point makes a bit of a difference. ;-)
Sounds like someone at Gigabyte made a big mistake by putting that
memory on the QVL list. Too bad it's not something that only happens
occasionally (with all manufacturers).

Thanks for the explanation and, again, the important thing is you've
fixed it! Have fun.