D0 I920 and Gigabyte EX58 UD5 issues.

[DavyBoy]

Hi,

I was wondering if anyone could help me out?

I recently built myself an i7 920 rig, with a Gigabyte EX58 UD5 mobo and 6GB of Corsair Dominator 1600MHz RAM, but I am having a lot of problems with it...

When I have everything set to auto in the BIOS, my system is completely unstable, whenever I try to run any benchmarks it either BSOD's, locks up, or resets itself all the time...

But I can manage to get it Prime stable for 11 hours + if I alter some of the advanced RAM timings... But if I attempt to do any overclocking whatsoever, the system becomes unstable again...

I have spent just over a week trying to get it to be stable, but the only stable settings I can get is when the CPU is @ its stock frequency,,,

And like I said, if I leave all settings on Auto for RAM, and CPU it crashes all the time to...

Does anyone have any idea which piece of hardware would be causing this to happen?

Surely it isn't normal for a PC to be unstable when the BIOS is set to auto?

I contacted Gigabyte themselves about this, and they actually had a system configured that was similar to mine (D0 i920, 6GB Corsair Dominator 1600MHz, EX58 UD5, GTX285, 600wPSU) and they ran Prime95 overnight with BIOS settings on auto and it ran perfectly...

I was actually given RMA numbers to send these items back and have them tested, but this was before I found stable settings... but I bought this gear so that I could overclock it, not run it at stock speeds!!!

I am probably spending more time in the mobo's BIOS than I do in Windows with this rig, it is driving me crazy!!!

I hope some of you folks on here could help me figure out what would be causing these problems, because I really dont know what it could be...

I know the Graphics card is working as I have tested that in my other PC and it was OK... Also, I was able to sit and play Left 4 Dead on this i7 rig for a couple of hours and it ran perfectly... I dont understand how I can convert AVI files to DVD, run games, but when I try running Prime, or OCCT I get problems....

Thanks in advance..

System specs:
Core i7 i920 (D0)
Gigabyte EX58 UD5 (F7 BIOS)
6GB Corsair Dominator 1600MHz
Asus GTX280
Zalman CNPS9700NT with 1366 mounting bracket
Hiper Type-M 630w
3 x WD SATAII HDD's
1 x LG DVD Writer
5 x 120mm Case Fans
Silverstone TJ10WMV Case
Windows Vista Ultimate x64

 

[Plec]

I'm guessing you've tried this - but have you looked to see if there is a new BIOS update for your Motherboard?

And what voltages have you experimented with?

You mention Prime95 have you ran the latest memtest?

Have you tried running the system with just one stick of memory for stabilty?

Additional: Did you run a skeleton system outside of the case and test that it was stable'ish before inserting it into case?

Edit: The other thing i would try is a differnet PSU if you have one to hand - as it's a quick and easy test and the psu is the only weak'ish link in your rig.

Edit x 2: Sorry this is a fragmented post - i was just off to bed. (i would definitley try a different psu if you can hold of one as it's a simple thing to rule out. And test your memory if you haven't done so already. )

Good luck.

 

[DavyBoy]

Hi,

I have ran memtest overnight with all 6GB installed, and it never found any errors.. Also, I have tested each module one at a time using Prime95, and I still get the same issues...

Even at stock speeds it still crashes when Priming from time to time...

I did suspect my PSU could be the problem, but I dont have another one at hand to test... I was planning on upgrading this pretty soon anyway, but the guys at Gigabyte only use a 600w in their rig that is similar spec to mine... So I would have thought my 630w would cope with my hardware??

I have never had problems like this with any other PC I have put together, if I leave my RAM and CPU settings on auto in the BIOS on my other PC it is perfectly stable...

I personally cant figure out which piece of hardware is giving me these problems, and it is really starting to drive me crazy... I have tried for a whole week to figure this out but I am not getting anywhere!!!

Maybe I would be better off RMAing it all back to get someone else to find what is causing the problems.... unless anyone else has any suggestions???

Would high temperatures cause BSOD, and system reboots/freezes??

Because I am only using my old Zalman 9700NT from my last rig to keep the CPU cool, and my idle temps at stock speeds are around 42c, and load temps are 66c, but when I start overclocking, @ 3GHz it hits 91c under full load when Priming... Could it be these insane temps??

 

[Plec]

I did suspect my PSU could be the problem, but I dont have another one at hand to test... I was planning on upgrading this pretty soon anyway, but the guys at Gigabyte only use a 600w in their rig that is similar spec to mine... So I would have thought my 630w would cope with my hardware??

I'm not worried about the wattage (although you have a lot of components hooked up) I'm more concerned that the psu could be failing or that it's just not delivering the power that it advertises (Hiper hasn't had the greatest of press recently). Test a different psu in the system as it's a quick and easy test (if a psu is to hand) and it's always worth trying with the type of problems you're having. (beg, borrow or steal one if you have to.)

Last edit If you haven't got a spare 700w psu to hand - unplug 2 of your HDD's, all of your case fans and your dvd-r and anything else that's drawing juice that you can do without - now test your system. The trouble with this test is - if it's still unstable i would still suspect the psu, as the gfx card and cpu could still be drawing enough juice to stress a weak/failing psu even without the other components connected. However, if it was stable during this test - start giving your psu some long hard stares...

Would high temperatures cause BSOD, and system reboots/freezes??

Because I am only using my old Zalman 9700NT from my last rig to keep the CPU cool, and my idle temps at stock speeds are around 42c, and load temps are 66c, but when I start overclocking, @ 3GHz it hits 91c under full load when Priming... Could it be these insane temps??

Yes, but your default clock temps are fine (warm, but fine) and yet you're still getting problems - so i don't think it's temp related. (but i wouldn't rule it out either as your clocked loads temps are insane - and, obviously, forget about clocking for now.)

Edit: have you tried slackening the memory timings and manually setting the voltages for the memory and Vcore? Plus, did you check for an updated BIOS?

Did i mention testing a different psu?

I must go to bed...

Edit: I am going to bed...

 

[DavyBoy]

Hi again,

I have tested out my rig only having 1 HDD connected, and I removed my GTX280 and installed an old 8500GT I had lying around, but I still get the same problems...

I am experiencing instability with stock settings all the time now, I am unable to run OCCT for more than 5 minutes before my system resets itself... This is really starting to drive me crazy!!

In the past I have always used cheap and cheerful AMD systems that have cost about half the amount of money I spent on this rig, and they ran perfectly... And I didn't have to be constantly entering the BIOS searching hopelessly for stable settings....

I have tried loosening the RAM timings but still experiencing the same problem..

I don't think that this has anything to do with BIOS settings anymore, something is failing in my system, but I just dont know what...

I guess maybe it is time to RMA this hardware to get it sorted because I can't find the problem myself....

::edit::

About the BIOS, I have tried F4, F6, and I am currently using the latest F7, and each one is exactly the same....

I was suspecting the PSU may be at fault, because when I run OCCT, the voltage readings on there only show the 12v as 8.5v... infact every single piece of software I have that give voltage readings all give me wrong readings.... I was told that this is nothing to worry about, and it is just the software reporting false readings....

I really don't want to spend out another £100 for a decent PSU only to find this has nothing to do with my problem... I don't want to waste anymore money on this rig to be honest!!!

And I don't know anyone else in my area with a PC that has a good PSU installed... just cheapo generic things, that would struggle to power my graphics card nevermind my entire system!!!

If no one else has any suggestions then I guess I have no other options but to return this mobo, RAM and CPU to have it tested, because I am unable to source the fault myself.....

I thought when I put this rig together that I would have my dream computer that would perform better than any other rig I have built to date, but I have ended up with a nightmare instead, I wish that I could just return all this goods and get a PII 955 rig instead.... I have never had any trouble with AMD hardware,,, this is the first Intel rig I have owned since Pentium III and it has caused me nothing but trouble since I built it....

 

[original.wood]

stick with it, get your stuff RMA'd and hopefully relax before you blow a blood vessel.

did you get it from OCUK? if so perhaps they would check it all out without having to go back to a manufacturer

 

[Plec]

I was suspecting the PSU may be at fault, because when I run OCCT, the voltage readings on there only show the 12v as 8.5v... infact every single piece of software I have that give voltage readings all give me wrong readings.... I was told that this is nothing to worry about, and it is just the software reporting false readings....

I really don't want to spend out another £100 for a decent PSU only to find this has nothing to do with my problem... I don't want to waste anymore money on this rig to be honest!!!

And I don't know anyone else in my area with a PC that has a good PSU installed... just cheapo generic things, that would struggle to power my graphics card nevermind my entire system!!!

Unfortunately, until you test a new psu you're a bit limited as to what you can do next - as you've tried almost everything. Checking your voltages via software doesn't really give you any indications that your psu is funtionong properly - unless one of the rails is consistently showing that it's below the 10% tolerance.

Have you got a local pc shop? You could explain your problem to them - buy a new psu and if it works keep it and return your hyper for a refund. If the replacement doesn't work then return it to the shop and pay a minor restocking fee for their trouble - agree a price before you buy. (you would need to agree all this with them first - but they usually don't mind.). It would save you all the hassle of sending everything back - if it proved to be the psu.

The above would probably cost you £10 for peace of mind - and it may prove to be the problem.

Let me know how you get on - i'm intrigued to know what it is.

Good luck.

 

[Popov]

This thread piqued my inerest as you have an almost identical build to the one I'm putting together myself. If I had to point at anything being the problem I would definitely go for the PSU as a cause. I followed a link to an Astek psu power draw app and it came out at pretty much bang on 650w draw (when overclocked)

I must admit not knowing much about the reliability of your brand of psu but I'm pretty sure it's not one people usually go for on these forums which leads me to believe it may not be the best. I would definitely follow the suggestion above and try to get a test psu with a higher wattage to see if it really is the problem.

 

[AWPC]

http://forums.overclockers.co.uk/showpost.php?p=14117832&postcount=31

Try one of the beta bioses as GB are working on some issues with Ram timings all the time.

 

[DavyBoy]

I just wish that I knew someone with a decent PSU so I could test this, but I dont know anyone...

And all the local PC shops are garbage to, they only stock cheap and nasty PSU's... the most powerful one is only 500w...

I did plan on upgrading my PSU, but after seeing how high my temperatures get under full load with my Zalman cooler, I thought that a new cooler would be more important... But if you guys really do think that it is the PSU that is at fault, then I guess I should just go ahead and use the money for a new PSU...

I can get a new cooler anytime, and my current Zalman is fine for stock frequencies... I just want my rig to be stable!!!

Can anyone suggest a decent PSU that doesn't cost an arm and a leg???

@AWPC, I personally dont think it is a BIOS issue, not when I have already tried three different BIOS'es all with the exact same results....

 

[Plec]

I did plan on upgrading my PSU, but after seeing how high my temperatures get under full load with my Zalman cooler, I thought that a new cooler would be more important... But if you guys really do think that it is the PSU that is at fault, then I guess I should just go ahead and use the money for a new PSU...

Disclamer: I'm not positive about anything - but it does look like a power problem your experiencing.

You could buy a new psu and find it doesn't solve your problems, which is why i was so keen for you to test one first. But the Hiper does look the most likely candidate as the weakest link...

The only security you have, is that if you buy online you have 7 days to return it if doesn't prove to be the problem - you'll just have to pay the return postage and make sure it's in the condition you bought it. (Not an ideal method - but you have to work with what you've got).)

Let us know how you get on.

 

[DavyBoy]

Thanks for all your help Plec, it is much appreciated man...

I thought that I would give Hiper a call to see what they had to say about this, but the guy I talked to didn't really say much, he just asked me to send him an email listing everything that is inside my case so that he could look into it in more detail for me...

I will await his reply to see what he thinks about this...

 

[Plec]

Nps, DavyBoy - i know from, all too many, personal experiences how frustrating these problems can be.

Hopefully, you're on the right track...

 

[AWPC]

Get your PC working stable with the minimum connected to the PSU and only use 2GB Ram. If you have an older gfx card which draws less power then that would be a good idea as well. This way you can isolate if its the PSU or not.

But like I said Gigabyte have several bioses as some Ram is troublesome.

Never overclock until you have a stable system in all applications otherwise you risk damaging your hardware if something is not right.

 

[DavyBoy]

I have already tried connecting an old 8500GT, and I ran each RAM module individually, and I still experienced the same problems...

Thanks for the suggestions anyway man...

 

[Plec]

What have you decided to do, DavyBoy?

 

[DavyBoy]

What have you decided to do, DavyBoy?

There really is nothing that I can do at this very moment, as I dont have the funds to purchase a new PSU... I wont have that kind of money for another 3 weeks...

But I dont want to wait all that time to get a new PSU only to find that this isn't the cause of the problem!!!

I really dont want to RMA anything, but it might be the only option that I have if a new PSU doesn't resolve these issues...

I have never had so much hassle like this with a new rig before, and all my previous systems have been cheap and cheerful... this is the first time I have spent a serious amount of money on a build, and it runs like crap!!!

Can anyone suggest anything else I could try to see if I can get this thing working, or do you all think that it is defo the PSU that isn't capable of powering this hardware???

 

[Plec]

You're reached an impasse really - you've tried everything with the components you have to hand - but, unfortunately, it really needs testing with a different psu just so you can rule it out.

My last throw of the dice would be to suggest you clear the CMOS and set all the voltages and timings manually - set the vCore and the memory voltages slightly above stock and relax the memory timings.

See what happens...

 

[DavyBoy]

Thanks for the suggestions Plec, I already tried playing with all the timings and voltages, setting the voltages slightly above stock and loosening RAM timings...

It has been these times that I have been playing around with timings that I have actually managed to get it stable with the CPU at stock frequency....

I have had the machine successfully run large in-place FFT's in Prime95 for 11 hours without any problems... but I would go back to the PC a day later with the BIOS still set the EXACT same, and then I goto run Prime torture test and it starts crashing on me again....

Here are some screens of successful Prime95 runs I have done with this rig...

11 hours:

5 hours:

and a 2 hour 15 minute run @ 3.2GHz:

I just wish that it would stay stable all of the time... there is definitely something wrong I think....

 

[AWPC]

Think it maybe your PSU as its combined 12v rails are only 36A. GTX280 draws around 34-35A of those so does not leave much for tolerances.

To prove this to yourself you need to put in your 8500GT then without changing anything else run that same Prime Test again and see what happens.

Failing that its possible 1 tiny little bios setting is not right as I7 are very sensitive to some settings and can fail easily even when set to auto as some bios are buggy and even though you set it to auto the voltages do not conform.

Try the latest beta bios from that URL I gave earlier but ensure you default all bios settings & clear the DMI data then exit the bios (so it saves those default settings) and re-enter the bios. Also, I7 are very sensitive to Ram timings as well. Only 1066FSB is officially supported on the X58 chipset and some ram has more issues than others.

Are you using the XMP Ram profile in the bios as that is unstable most of the time!

Its also easy to corrupt your bios on I7 when overclocking (I have done this many times) and sometimes a corrupt bios is hard to overwrite!!!