Constant BSOD, data corruption

[garyoa1]

Ok guys, riddle me this. New build. Vista business. Now, there is NO OC'ing involved. Just running stock

Gigabyte x48-dq6 - 4g crucial ballistix rated at 5 5 5 15. Which the bios is reading at 5 7 7 24 at default settings. (which memtest is also reading at 5 7 7 24) with a Q6700 quad.

Ran fine for about 3 weeks. Then BSOD started. Got virtually every BSOD error known to mankind over a few days and finally it was dead in the water.
Ok, dumped vista, swapped the HD out, got back to xp pro.
Withinn 3 days... same thing. Corrupted data.

FWIW, coretest shows idle at 48c. A bit hot but not near critical.

Ran memtest86+ .... instantly errors up the ying yang right from the get go with the 4 sticks installed. Pulled ram, started memtest with one stick. 8 passes... no errors. Switched modules... again, no errors. Rather than going thru all of the ram, just reinstalled xp with 2 gig. Few days... data corruption again.
Running memtest for 12 hours with the 2 gig... no errors. Replaced the 2 gig with the other 2 gig... overnight... no errors.

Data corruption continues.

Now, question is, since I got a truckload of errors instantly with the 4g installed and none with the two sticks, is it possible I have some bad memory slots, therefore bad Mobo? I have the two sticks in banks 0 and 2. Didn't try them in banks 1 and 3.
.

[RyanJ]

Bad mobo. If you're getting all sorts of memory errors AND it's not OS related, it's your mobo. It COULD be your memory, but I doubt it.

-r-

[sswilson]

Originally Posted by garyoa1

Ok guys, riddle me this. New build. Vista business. Now, there is NO OC'ing involved. Just running stock

Gigabyte x48-dq6 - 4g crucial ballistix rated at 5 5 5 15. Which the bios is reading at 5 7 7 24 at default settings. (which memtest is also reading at 5 7 7 24) with a Q6700 quad.

Ran fine for about 3 weeks. Then BSOD started. Got virtually every BSOD error known to mankind over a few days and finally it was dead in the water.
Ok, dumped vista, swapped the HD out, got back to xp pro.
Withinn 3 days... same thing. Corrupted data.

FWIW, coretest shows idle at 48c. A bit hot but not near critical.

Ran memtest86+ .... instantly errors up the ying yang right from the get go with the 4 sticks installed. Pulled ram, started memtest with one stick. 8 passes... no errors. Switched modules... again, no errors. Rather than going thru all of the ram, just reinstalled xp with 2 gig. Few days... data corruption again.
Running memtest for 12 hours with the 2 gig... no errors. Replaced the 2 gig with the other 2 gig... overnight... no errors.

Data corruption continues.

Now, question is, since I got a truckload of errors instantly with the 4g installed and none with the two sticks, is it possible I have some bad memory slots, therefore bad Mobo? I have the two sticks in banks 0 and 2. Didn't try them in banks 1 and 3.
.

These are the new single sided ballistix? Sounds like a textbook example of why folks are avoiding this latest batch of DDR2 Ballistix.

I'd bet dollars for doughnuts that changing the ram out will make all of the issues go away, but just to be sure... did you set the mem voltage up properly in bios? (probably needs to be 2.1V).

[DarKStar]

That's one trap some people fall into - When you buy RAM advertized with lower latency or high speeds, you have to always look at the specified voltage it was tested at - When installing RAM on your system, it will default to the JEDEC timings - so to run it at advertized specs you will have to manually raise your voltage, (2.1V for example) or as stated on the packaging.

Windows errors, data corruption, that is a classic case of memory problems - Some motherboards are very picky about RAM - try to swap your RAM with something else, and use the same type, same timings for a pair, if it does that no matter what RAM you put in, then it looks like your motherbaord is failing - have you done any visual inspection on your board to look for anything unusual oh wait I mean anything usual :D like burnt resistors, bulging, leaking caps ? ;-)

[garyoa1]

Thanx, Ryan. That's what I'm leaning towards. Can't be OS related if I used vista and xp. (then again, it's microslop) Machine is completely new. Coolermaster case and coolermaster 680w PSU, pny 9800GT.

Since everything is new it could have been anything really. That's the downside of building your own. Can't reasonably swap out parts to get it right. Since memtest showed errors instantly with the banks all full, and none with one or two banks loaded... doesn't seem to be memory to me. I'm assuming a problem between the slot banks on the mobo.

I could just try it with one stick, but I'm kind of tired of reinstalling windwoes. :) And one gig won't cut it anyway.

Got the Mobo from newegg and it sat on the shelf for a month before I got time to build. Since their warranty is 30 days and it's now about 3 months old... well, emailed gigabyte (their warranty is 3 years) but they have a bizarre way of contacting them. I guess time will tell. Also ordered a new heat sink to keep the temp levels down, even tho it's not really critical. I have no intention of OC'ing, but you never know what the future holds.

FWIW, this is the second gigabyte board with a bad slot (if that's what this is) The current old machine has a bad pci slot. Don't need it, but it's annoying. Makes me wonder about gigabyte.
------------
The ballisitx is ddr2 1066 - PC2 8500 - BL12864AA1065. But since they're sealed... no idea if they're double sided or not. But loading the bios at failsafe. But still running memtest as we speak (this time 10.5 hours so far) with two sticks in and so far 28 passes with no errors.

And yeah, crucial's another one that's been making me wonder lately. I heard patriot is the new up an coming king of ram.

[misterd]

My vote is RAM also especially after going through the same sort of deal with my 2x1GB PC8500 Ballistix. I was leaning on my board but finally realized it was one of the 2 sticks went crazy.

[DarKStar]

Originally Posted by misterd

was one of the 2 sticks went crazy.

....I guess you can say they went "ballistic"

They are named appropriately too ;=)

[sswilson]

If it's memory that's been bought in the last 6 months or so, it's pretty well guaranteed that it's the new single sided stuff. (Unless there was a lot of dust on it when you got it..... :) ).

Running fine for a month or so, and then BSODing is a classic example of the problems folks have reported with the single sided stuff.

[garyoa1]

Lovely. You guys certainly know how to screw with one's head! :) I'm not seeing anything in the bios that gives an actual reading of the ram voltage. Again, I'm no guru at this. I do see in the health status section case open section it says ddr218v set 2.224v. And in the m.i.t. looks like I can only increase it. Case opened status is disabled and it says yes, BTW. I have no idea what that's supposed to mean.

Also FWIW... I don't see this particular model ram on neweggs site anymore... hmmm...

Well... I have 4 gig of mushkin in my old machine... maybe I'll steal two and see what happens. The annoying thing is that it's not an instant crash. Just over time. So productivity for me is in the trash can. Since the crashes I'm pretty much on a first name basis with adobe trying to get @##% photoshop reinstalled since I can't deactivate after the crash. I don't game "much". Maybe some AOE once in a while but nothing really intense. These are business machines, generally. :(

FWIW, I've also heard that sometimes running all four banks full creates a problem on some boards.

[Eldonko]

Prob just mem incompatibility, got anything else to test dual channel with?