PSU Upgrade

quyeno

quyeno

quyeno

Associate
Joined
11 Feb 2009
Posts
40
Hi guys and gals

I'm looking for advice and would really appreciate some help. To cut a long story short my pc has been resetting recently when playing games, figured it down to the PSU....but back to this later!

I used Thermaltake's PSU calculator and it recommends for my system 754W:

i5 750 OCd to 4GHz
GTX580
4x DDR3 DIMMS
1 DVD_RW
High End Mobo
Corsair H50 cooler
1x 140mm fan
5x 120mm fan
3 USB devices

Would a 850W provide a safe margin for operation for yrs (since most PSU come with long warranties now) or would it be best to go for a 1000W?

Now the strange thing which I don't understand and would appreciate some light on. I've been running this system now since April (the psu Corsair TX650W was bought back in dec 2009 though) and the system has been fine. I've been playing mega sessions of Battlefield 3 and no issues what so ever until recently where my pc would reboot. I've ran OCCT and memtest and all my components seem fine but the PSU test fails in OCCT. But what I don't understand is, if the recommmended wattage for my setup is 754W, then how did my system manage on 650W? The 650W was recommended by the Corsair configurator back then.

Any help and suggestions would be appreciated

Thanks
Q
 
650w is fine for your system, when I had my overclocked i7 920 system with a 580 GTX in I was getting around 430w from the mains at full load.

PSU Calculators seem to over compensate.

This could still say a PSU fault but what is OCCT detecting to show PSU fault?
 
Hi mate, thanks for the reply. OCCT runs a linpak test and also its GPU test simulataneously which test for errors and designed to run the GPU hot using various shaders (kind of like furmark). Not sure exactly what OCCT is detecting to pick up errors but my guess is that the PSU test is designed to draw as much power as possible for a length of time and maybe analysing voltages supplied to the CPU and GPU nd also testing for compute errors in the CPU and GPU.

I've ran OCCT cpu error test sperately and it found no errors after an hour, also ran linpak cpu heat test seperately for an hour and was stable and also ran the gpu test seperately for an hour and no errors. But if I run the PSU test (linpak and GPU test at the same time) my system reboots after 2 minutes. So i guess the PSU must have a droopping 12v rail and is prob not what it used to be.

All my temps are fine, CPU at max load is about 68C, GPU 78C, SYS 60C.

Last night the rest of my pc booted up but the GTX580 wasn't powered up. Fans weren't spinning and got a POST beep (one long, four short beeps or something). I could even hear my pc going into windows. So i dug out my old GTX275 and everything seems to be working but haven't tried playing BF3 yet, too busy trying to find a replacement PSU before I RMA the TX650W. Tried the GTX580 in my mates PC and it was fine, so it's not a faulty GFX card.

I've read it could be a mobo fault but I dont think so...but then again how would you test if it's a faulty mobo?

At the moment I'm short listing a Corsair AX850, OCZ 1000W and even considering OcUK 850W. But wondering what's the OEM for the OcUK PSU?
 
Last edited:
Avoid the Ocuk version.

The Corsair AX850 will be fine, with finding out if your motherboard is at fault then you need to start swopping parts out until you have exhausted all possible means.

Daft question - Are you testing with your Overclock or I have missed that bit!

Oh- Do you own a multimeter?
 
Hey man

Yeah I'm testing with my i5 OC'd to 4GHz, have been running it at this speed since day one. My mobo is an Asus P7P55D.

I don't have a mulitmeter but could pick one up. How would I test the PSU with it?
 
Drop the overclock to default settings and retest.

When you stress your PC fine a spare molex and see what the voltages read, you might have make some tails and connect them to the multimeter so your hands are free and you can monitor.
 
Just wondering, why reset back to default CPU settings?

I did that when I started getting issues with my pc rebooting and tried playing BF3 with stock settings and still got random rebooting. The pc seems fine running the OC settings while just non demanding stuff like browsing, downloading, etc.
 
Just to rule out a possible issue with the overclock least then it may point towards a iffy PSU if your pc still fails under load.

Just trying to think of any other issues which could course this to happen!
 
Yeah, that's what I thought but I can't test the system as it was before since it won't power the GTX580 and now using a GTX275. I'm ordering the AX850 from a well known book retailer so it should then be easy to return the PSU if it turns out not to the problem. :P I think all in all though that everything seems to be pointing to a failing PSU, it is over 3 years old and running a heavily overclocked system. Great thing that Corsair gave a generous 5yr warranty :D
 
Hi

I got my new PSU install, Corsair AX850 and using the GTX580. It's running whereas the last TX650 did even power the fans. But unfortunately i'm still getting random crashes when playing Battlefield3. Is there a program which will constantly track and save system temps, cpu temps and GPU temps? Got a feeling may be the GPU is over heating and shutting down the system.
 
Cool, was using HWMonitor but unforunately it doesnt continuosly log to file. I've noticed with HWMonitor that my GPU temp is about 48C at idle. CPU temps are 26-31C (i5 now running at stock speeds). System temp isabout 30C. EVerything looks ok i guess but the gpu seems a bit high for idle.

Just wondering, would a faulty PCI-E slot cause reboots when playing games?
 
You must log in or register to reply here.
Back
Top Bottom