Bad board, bad memory, or.... (P965-DS3)

[rubblerubble]

Just slapped together my new system...

Gigabyte P965-DS3, Rev 2, F9
C2D e6300
Freezer Pro 7
Corsair TWIN2X2048-6400
xFX 7600GS
Creative X-Fi
2 SATA hard drives
Antec Trio 650W

...and after initial 24 hours of perfection (basic OS loading, Prime95 testing), I'm starting to get random reboots and failures. At first, I was OC'ing CPU to 400FSB to work nicely with stock settings of RAM. I verified, on numerous occasions, timings set in BIOS to satisfy Corsair specs; 5-5-5-12, normal voltage (1.8) + .1 = 1.9v. Plenty of cooling is in play here (highest CPU temp was recorded at 30C, system temp 39C).

After installing the second SATA drive, which turned out to be defective, the system wouldn't post. I removed the drive, cleared CMOS, reset timings, and reloaded OS for good measure. I also refrained from OC'ing and kept FSB to stock 266MHz just to get stability back. Still, I get random reboots and failures to post. On a non-post, I swapped the 2 RAM sticks in slot 1 and 3 and it still failed to post. When using either stick on its own in slot 1, system posted fine without clearing of CMOS. When moving both sticks to slots 2 and 4, or just 1 and 2, system posted without clearing of CMOS. Initially, I was thinking that perhaps slot 3 could be defective, but after a short time, when the sticks were used in the other combinations listed above, the random restarts came back.

I tried running MEMTEST. On all 4 occasions when combinations 1 and 3, and 2 and 4 were used, the system rebooted during random phases of the test and never completed. I'm going to try more MEMTEST trials tomorrow with only one stick at a time, but can anyone give me some insight in advance? Could it be a bad batch, or something else?

 

[wallis]

i think you should up the vdim voltage to 2.1v

the corsair twin X ram is rated between 1.8-2.2v


my geil ram is the same and there is a sticker on my ram stating that for best results set vDimm voltage to 2.1v


here is a handy thread that may help

http://forums.overclockers.co.uk/showthread.php?t=17636047

 

[rubblerubble]

Wallis,

Thanks for the quick reply. However, I'm just trying to get this rig running stable at stock values (no OC'ing) at manufacturer specs just so I can isolate the problematic component, if there really is one. So for now, I'm keeping the FSB at 266MHz, Auto SPD (2:3), and 1.9 volts running at 5-5-5-12.If you think it would help though, I'll bump up the vdim.

 

[Jizumonkey]

If you're having troubles in 1&3 or 2&4 it's likely dual channel compatibility issues, the only sure way is to exhaust every possibility.

Run a single stick in each bank, any errors always swap the stick out to see whether it's the mobo or the ram. If you get no errors, then your memory is not at fault.

If you get errors when using dual channel banks then you may not get away with RMA'ing it as it may be that the memory type is simply not supported. Either that or simply dont run it in dual channel in slots 1&2, 2&3 or 1&4 (i.e two slots of differing colour).

Run these tests at manufacturer stock, overvolting is a bad fix imo.

 

[rubblerubble]

How's this for interesting...

I never put too much value into the randomness of the reboots, but after testing both sticks individually and watching them pass a few MEMTESTS in row just now, I remembered something odd from last night. Since I'm working with both my old PC and new build, I use a Zonet KVM3004 to switch between the two machines. Sure enough, on 3 of 4 occasions today, and after 3 or 4 consecutive passes with ZERO errors, when I move the mouse on the old PC as I switch the KVM to the new build, WHAM, the new build reboots. Is it possible that the KVM is sending such a drastic charge to the DS3 when switched that it would provoke a reboot? I never heard of this and find it quite alarming!

I will now test without the KVM and see how it goes.

 

[Jizumonkey]

Well from what you've said I'm sure it will help...I didnt know what a KVM was so I looked and see that it's a multiple PC controller.

At a guess I'd also say that it's more likely the problem is being caused not by the signal being sent by the device, but more the fact it draws power from the ps/2 port.

Do some testing and see how it goes, if you cant do without the controller maybe try one of those small adaptors that converts ps/2 to USB.