Cold boot issues - advice before I kick off RMA

[d3fiant]

**reposted here because I stuck it in the 3xs forum by mistake**

Recently (06/09) bought the following from Scan;
GB X58A-UD3R mobo
Corsair XMS3 6GB DDR3 1600 3x2GB
Intel Core i7 950

To join recently purchased SSD and GTX460 (not scan)

Basically I have a cold boot issue where I get repeated BSOD with various messages, mostly memory related, just before or at the logon screen (W7 HP x64). However after 5-7 minutes of the system being active it will boot fine and I wont get another issue until the next cold boot, but it happens at EVERY cold boot

What I've tried
Latest ver of Memtest+ gives all three dimms clean bill of health and using one dimm doesnt change the outcome
SSD and GTX460 worked flawless on old system, same case and PSU used now
Tried OS installed on different SATA controllers, there are 4 of them, no change
I have tried several bios for this mobo, clean cmos and dmi each time
I have applied the XMP profile 1 for the memory as the mobo doesnt supply enough voltage at defaults
No OC involved
Temps are very stable and not high or of concern, case airflow is outstanding on the eclipse
Turned of tons in the bios bar items essential to get the system up and running, things like unused sata controllers, sound, firewire etc etc

I'm not new at this, this is the 4th system I have built on this mobo with very similar components in the last couple of months but the first v2, but Ive been building gaming PC's for over a decade. My gut feeling is possible cold solder joint on the mobo somewhere, now I may have missed something obvious and welcome all and any suggestions but I believe my only next logical step is to RMA the mobo/cpu/memory or at least just the mobo

Scan, what are your thoughts, I'm loathed to do an RMA as when it runs this system is faaaast but I shouldnt have to put up with these issues and would appreciate your help getting this resolved

Thanks in advance

d3fiant

[d3fiant]

I have now logged a returns query for this, I have tried everything I can thing of, and I am not alone with this issue, several people have reported extremely similar issues with this mobo as below on the GB forums

http://forum.giga-byte.co.uk/index.p...ic,2860.0.html

http://forum.giga-byte.co.uk/index.p...ic,2907.0.html

[d3fiant]

Seen ur response

So far, booted fine into a Win7 64Bit I already had on a drive from a previous install, detected drivers, rebooted fine. I would expect that would have had more chance of a BSD than an install to this board, which I will now do. I will leave under load all day, then try it in the morning after it has been on the bench all night (turned off)

I would also expect that but NEVER had the BSOD during the install process, only once the initial W7 build is complete and then left off for a few hours then its BSOD on every cold boot

[d3fiant]

On its way back with no fault found profoundly pissed off as I tested shed loads before I sent it back but to be honest I cant fault their testing methods if they did do what they stated and have given me a few pointers on next steps

[d3fiant]

Its always fun talking to myself but anyways....

Received the goods back from Scan, rebuilt the system using their suggestion of inserting new Intel F6 drivers during the build. All runs fine during build as it did before.

This morning was first cold boot...........REPEATED BSOD

Now Im pretty pissed off with this issue. I couldnt fault Scans testing EXCEPT when I got it back all SATA controllers where is IDE mode, now if Im using an SSD I need to use ACHI to enable enhanced features such as trim, at least that is my understanding. Why would you test it in IDE mode??

Im posting in the OCZ forums as a long shot but I am also now going to investigate my rights to get a refund, the system isnt usable if every cold boot I get several BSOD and every so often one of them corrupts the OS requiring yet another rebuild

[d3fiant]

OK, just arranging RMA #2, testing with a mechanical drive proved to me it is not the SSD and that the fault may have been missed in RMA #1 due to the tests being carried out in IDE mode. Incredibly annoying but in Scans defence its a bloody odd one and their response to the issues has been fast so lets see where RMA #2 gets me, although it wont be 'no fault found'

[kalniel]

Are you testing with the same power supply as SCAN? Cold/warm boot issues are sometimes PSU dependent.

The other thing is bend CPU socket pins, but that shouldn't be a problem if SCAN can boot it fine.

edit: what is the manufacturer code for the RAM?

[MaddAussie]

Have you bread boarded it out of the case? It might be a slight grounding, that cures itself once the components have warmed up a tad

[d3fiant]

@ kalniel - the PSU was an exisitng one, 850W akasa so scan would not have had that to test with, however I did borrow an 800W corsair from a friend and that didnt change anything. As for bent pins I did check, I have one of those electrcians lit magnifying glass contraptions and could see no bent pins

@ MaddAussie - first thing I checked, my case allows for the entire mobotray and backpanel to be removed which allows me to be very carefull when mounting the mobo, it allows me to install practically all components and then slides into the case

[kalniel]

But you might still be grounding to the motherboard tray?

The other thing is to check that the backplate isn't gounding anything.

This has me slightly worried because I've just ordered virtually the identical specification :S

Though I would recommend using IDE over AHCI and making sure you're using intels hdd ports rather than gigabytes, I remember having problems with a build a while back before I sussed that one out.

[d3fiant]

I know what your saying, what I meant was that doing the installation outside of the case allows me to very careful with placement, the mobo isnt shorting and has been on and off the tray alot due to the unsuccesful RMA, even had it running on my desk briefly, same issue. There is a clear gap between the backplate and the mobo. I appreciate the concern about your order, but this is the 4th 'almost' identicle system Ive built in the last two months, the other three had no such issues granted two were on v1 mobo's.

I would prefer not to use IDE due to using a SF-1500 SSD, while I now know TRIM works in IDE I have found clear examples where AHCI allows optimisations in W7 for SSD that results in a fair improvement in performance. I am using the ICH10R ports as the Marvell SATA 6gbps ports sucked and the GB ports are even worse. I have done extensive tests around using MS AHCI drivers as well as Intels latest RST F6 drivers, both stable and beta. I have also tried several different bios for the mobo.

I have spent a ridiculous amount of time trying to bug search this issue, the RMA was last chance saloon

Thanks for your suggestions, its always good to bounce things off of other enthusiasts You never know when u have made a n00b mistake

Oh and by the way, when this system does eventually run it runs like 'poo off of an incredibly hot shovel' I'd still buy the same components again, this is just a 'glitch'

thanks

d3fiant

[d3fiant]

OK, short term fix, well not a full fix Basically I am now running an older beta bios, B16, and this removes the BSOD on cold boot and replaces it with a single error in post stating 'recovering lost memory' which then does one reboot and boots fine so minimal risk to the OS. Will test new bios as they come out but makes me think its a bios or memory error. I opted not to go for a 2nd rma