Power supply possible problem

R

rpm

n00b
Joined
Sep 24, 2011
Messages
12
Hello everyone, this is my first post in this forum , I registered to share with you my problem and, hopefully , understand what is wrong with my machine.

Specs :

CPU i7 2600k
Mobo Asus p8p67
8 GB Corsair Vengeance low profile 1600 Mhz v1.5
Corsair TX 750
SSD Crucial m4 64 Gb.
Scythe Ninja 3
GTX 460

I kept only the GTX 460 from the old config and bought all the rest brand new.

Problem :

When I play a game , if I tab it the computer reboots once every 9-10 times. No BSOD.
I am able to replicate it in 2 ways :

1) OCCT Power Suplly test. As long as CPU kicks in with 100% usage + GPU 100% usage, system reboots.
2) Kombustor GPU burn in test, 1920x1080. As soon as i launch it , system reboots. ( It does not happen with res lower than that , does the GPU needs more power at higher res? ).

I have already updated Bios on the motherboard, all settings are DEFAULT. It happened also with other settings and OC CPU , but it happens at default too so it must not be a problem of settings.
I updated the M4 firmware to 009. but as long as i know , ssd should not be the problem.

I am able to run prime95 for hours and there will be no restart.
I can run Kombustor at low res for hours , no restart.
OCCT power supply leads to a reboot.
Kombustor 1920x1080 leads to a reboot.

What do you guys think?

I think my PSU may be faulty , and maybe I should RMA it.
but before doing it , i need to be reasonably sure thats the problem as i work with my pc and i will need to buy a new one before rma'ing it and then sell it when it comes back.
A few questions :

- Can a faulty PSU give NO problems at startup and lead to a system reboot when under a certain amount of load?

- Why does it crash with kombustor at 1920x1080 but not lower res? with OCCT it just crashes aftter the CPU kicks in ( after 2-3 mins i guess ) , so it's possible it's a problem of power.

- Is it possible my GPU is the problem ? it worked perfectly until i changed the rig

Btw CPU temps are 24-57 under load ( stock ) and 24-65 OC @4.3 ghz.
GPU temps are 28-70

Thanks in advance for your opinions.
 
Are there chances the GPU is the problem? it shouldn't be caused from the CPU or the memory afaik and from tests i've done.
 
That's a strange behavior for any VGA related problem.. what I''d bet on is a massive voltage drop in high load situations and that means your PSU has regulation issues.. however, you need to check the system with a different PSU first, or if you have a multimeter you can check the voltages yourself while duplicating that high stress scenarios when the crash happens.
 
Ok i've got no multimeter , it seems i will need another PSU to test with the system which is gonna be a PITA as i have none atm and i wouldnt like to buy one for a test!
Can you think of other causes which may cause the reboot ?
ps : thanks for the answer, Profumo
 
know someone with a scope or multimeter? expensive way to go, buying a new psu for the sake of a suspicion
 
rpm said:
Ok i've got no multimeter , it seems i will need another PSU to test with the system which is gonna be a PITA as i have none atm and i wouldnt like to buy one for a test!
Can you think of other causes which may cause the reboot ?
ps : thanks for the answer, Profumo
Click to expand...

Not really, maybe some BIOS issues(post some screens with the CPU/GPU/Memory power settings if you can), but I'm inclined to think it's the PSU by the way you describe the behavior(resets and not crashes and no BSOD), do you get artifacts or other visual effects when your VGA is in stress?

And you don't need to buy a new PSU, just borrow one from a friend for a couple of hours..
 
Bios is @default settings now and no , i dont get artifacts. in fact i can play if im lucky for 1 hour and have no crash , as i guess games dont imply 100% cpu and gpu together ( at least the one i tried since i formatted , which is warhammer 40.000 space marine ).

so cpu power settings are default , and vcore is pretty high actually . when vcore was low ( during some tests to fine-tune the system , i had a bsod anyway not a reboot ).

and gpu is @max voltage with a frequency similar to default , so it should be uber stable.
 
Internal PLL Enabled
LL Calibration enabled, Medium
PLL overvoltage Auto
Load Line Calibration Medium

I didnt specify any CPU PLL tho, i will look whats the setting in bios.
 
I hate this Autom/Medium/non nominal settings..

Sandy Bridge CPUs degrade very fast under high voltage(pretty poor silicon), you said you have the VCore pretty high, it should in any circumstances be at or below 1.5v otherwise you will damage your CPU, manually set the PLL to 1.6v max(auto is probably 1.8).

But honestly, the most likely cause is the PSU :)
 
EDITED

what do you mean by degrade?
I dont know what cpu PLL actually is , it's on Auto ( 1.8 during bios check ).

My vcore is high coz now it's on Auto and i didnt use any offset. I can reach 4.3 stable with around 1.250, which is pretty decent.
When sitting on desktop , cpu is @ 1.6 Ghz and vcore 0.990-1 v.

So i should set Cpu PLL to 1.6? What does it do? i'm not sure I understood what it is and how it works.
 
CPU PLL coresponds to the clock generator voltage, it will help with BCLK overclocks(irrelevant on SandyBridge).
Yeah, 1.25 is excellent:)
When the CPU degrades it will ask for more voltage to run stable at clocks previously achieved with lower voltages, eventually it will die.. it's a way to kill Sandy Bridge CPUs and RMA it for another sample, is something overclockers do in search for better performing chips(binning with style as some would call it;p), you set the VCore to 1.65 and run a stress test for a couple of hours;)
 
Ok , thank you . I dont think this is somehow related to my crashes anyway , is it ? I'm saying that coz i dont think kombustor fails because of it , it would make no sense to me.
As far as i understood , my 1.8 CPU PLL is ok , is it?
 
More or less, 1.6 would simply be safer and would not make the slightest difference in performance:)

Yeah, with 1.25 VCore you're more than ok.. as I said initially, my bet is on a faulty PSU and until you get confirmation or exclusion of that, there''s really no point in playing with the BIOS in hopes for stability.
 
Ok im running a 650w chinese PSU ( named ENERGON , never seen a sh*it like this one ) but this is all i could get in a few hours.

Running the same system without 2 HDs , i dont think they can make any difference.

I will let you know , gonna OCCT the system right now.

UPDATE : 30 mins of OCCT + kmobustor 1920x1080 together, no crash yet. it seems we may have found the faulty component.
 
Last edited:
Energon is a PSU line from Inter-Tech, and they are low end CWT units afaik :)

So essentially you need a new PSU :)
 
yup this is a temporary one , gonna RMA mine with corsair and hope they are fast!
for the time being this energon is doing fine :)
 
Ok i boight a new one , a modular one , and gonna send the old one in RMA...

one quick question : is it normal to have OCCT power supply test with some errors on it ?
 
rpm said:
Ok i boight a new one , a modular one , and gonna send the old one in RMA...

one quick question : is it normal to have OCCT power supply test with some errors on it ?
Click to expand...

Depends on the error/s.
 
Upload a screen with what you get, I haven't used OCCT in a while and I'm not sure we're talking about the same thing.. also, they used to have a forum iirc so you might wanna' check there too.
 
You must log in or register to reply here.
Back
Top