Power Supply voltage

JonoB

JonoB

JonoB

Associate
Joined
19 Nov 2004
Posts
248
I have been suffering from intermittent crashing on my system. It only ever happens whilst gaming.

I have suspected that the PSU (Corsair HX620) is the problem, and bought a PSU voltage tester. The results are as follows:

-12v: 11.7
+12v2: 11.9
5vSB: 4.9
PG: 320ms
+5v: 4.9
+12v1: 11.9
+3.3v: 3.3

I can see that some of these are not exactly on their "expected" value, but I have no idea if these tolerances are OK or not.

Anyone with more knowledge than me able to tell if this PSU is borked or not?
 
I hope this helps you,

http://en.wikipedia.org/wiki/ATX
Untitled2-26_zps25f39639.png
 
JonoB said:
-12v: 11.7
Click to expand...
OK, ±10% (±1.2 V), –10.8 V to –13.2 V

JonoB said:
+12v2: 11.9
Click to expand...
OK, ±5% (±0.60 V), +11.40 V to +12.60 V

JonoB said:
5vSB: 4.9
Click to expand...
OK, ±5% (±0.25 V), +4.75 V to +5.25 V

JonoB said:
+5v: 4.9
Click to expand...
OK, ±5% (±0.25 V), +4.75 V to +5.25 V

JonoB said:
+12v1: 11.9
Click to expand...
OK, ±5% (±0.60 V), +11.40 V to +12.60 V

JonoB said:
+3.3v: 3.3
Click to expand...
OK, ±5% (±0.165 V), +3.135 V to +3.465 V


Not that those readings are the be all and end all of making sure a PSU works correctly - depending on load the rails may not be able to deliver the specified voltage - but at least it gives you a fair knowledge that it is at least reasonable.
 
Thanks for the replies.

Just ran memtest and it went through one full pass without any errors.

It definitely only crashes when gaming, and only after a while (30 minutes or more), so it certainly seems load related.

When the pc crashes, the screen usually goes completely blank, sometimes accompanied by a buzzing sound from the speakers.

Somehow I still think its the psu, but I'm reluctant to fork out for a new one if its not.

Any reliable ways to test the mobo or graphics card?
 
I assume you aren't overclocking or anything at present, but it would be helpful to list your full spec, just in case it flags up anything obvious. E.g. Motherboard/CPU/RAM/Gfx/Heatsink


JonoB said:
Just ran memtest and it went through one full pass without any errors.
Click to expand...

This doesn't really rule anything other than the memory out, but combined with the next comment:

JonoB said:
It definitely only crashes when gaming, and only after a while (30 minutes or more), so it certainly seems load related.
Click to expand...

Could be power related, but perhaps it could be heat related? Heatsink not seated properly, thermal paste not applied correctly, fan not plugged in, or fan control not set correctly in the bios?

Given that Memtest runs ok, have you tried just booting the PC to desktop and leaving it idle there? does it still crash?

May be worth downloading some monitoring programs e.g. CoreTemp etc. and keeping an eye on temperatures (and if something like Motherboard Monitor is still available - even voltages).

When it does crash - does anything inside the PC feel especially hot to the touch e.g. Heatsink, Graphics card, motherboard heatsink?


JonoB said:
Somehow I still think its the psu, but I'm reluctant to fork out for a new one if its not.
Click to expand...

The PSU is not completely ruled out at this point, but I would do some further troubleshooting to eliminate everything else.

JonoB said:
Any reliable ways to test the mobo or graphics card?
Click to expand...

Graphics card can be tested in another machine if you have one - motherboard not so easily, but I would say it is probably the least likely part to have a problem.
 
Specs as follows:

- Gigabyte Z77-D3H Intel Z77 (Socket 1155) DDR3 Motherboard
- Intel Core i5-3570K
- 16Gb RAM (Kingston HyperX Blu 8GB x2)
- Gigabyte ATI Radeon HD 7850 OC Windforce 2X 2048MB GDDR5 PCI-Express Graphics Card
- Crucial 256MB SSD + Samsung 1TB
- Corsair HX-620 modular PSU
- Windows 7 (64 bit)

I suspected temperatures as well, and I have booted straight into bios right after a crash. All temperatures are completely fine, and nothing feels especially hot. I will, however, run coretemp to see if there is anything amiss.

As far as I can recall, the PC is only ever crashed once whilst left idle, but that has not happened for a long time. Every other crash has only happened whilst gaming.
 
JonoB said:
I suspected temperatures as well, and I have booted straight into bios right after a crash. All temperatures are completely fine, and nothing feels especially hot. I will, however, run coretemp to see if there is anything amiss.
Click to expand...

Check this out first, and also (assuming you are running the stock Intel cooler) it would be worth checking to see that you have all the Power Saving features turned on in the bios e.g. Speedstep / C1E etc.


JonoB said:
- Corsair HX-620 modular PSU
Click to expand...
How old is this PSU? was it carried over from a previous build - only reference I can find to this model dates back to 2006 - it could still be possible that under load it is failing (Despite the good tester voltages)
 
F*&%king hell. This thing is driving me crazy.

I ran furmark and prime95 at the same time. All 4 CPU cores maxed out, and hitting temps of around 95 degrees (stock cooler). Graphics card hitting about 65 degrees.

I ran for over an hour without any problems. System was rock solid - I could tab into email, browser coding environment, word documents all at the same time. Not a single crash.

Start playing games when everything has cooled down completely. Boom. Down she frakkin goes.
 
My Corsair HX620 was RMA'd a while back for voltage dropping under load. As it ran voltage would reduce until the system shut down.

If you can run Furmark and Prime together then it doesn't sound so much like a load issue.

What voltage is the RAM at. I have 4 x 4GB HyperX and at 1600MHz system is unstable without 1.65V.

I usually load up the XMP profile and then check voltage at the required 1.65V

Gaming is a different load than either memcheck, prime or Furmark as it pushes a lot of data around the system. I'd try memory first then if you have an OC on the system you may need to experiment with the chipset settings / voltages

AD
 
Get voltage reading when it is doing nothing
And get another set of readings under furmark and prime
At 11.7Volts in IDLE it dose not sound that good
It works until 11.4V but i will not trust a PSU under 11.9V

Big voltage drop from idle to load can do nasty problems
Use software reading if nothing else is available
 
Last edited:
JonoB said:
All 4 CPU cores maxed out, and hitting temps of around 95 degrees (stock cooler).
Click to expand...

Is that a typo? or is it actually 95 degree's Celsius. If it is, lower your clock a tad so your CPU might just survive long enough to find out what's wrong with it! What CPU is it?

This thread doesn't sound so much like a PSU issue. It seem's like one of those 'It could be anything, so it's probably memory related' problems. I know you ran prime, furmark, memtest but like everyone said we cant rule out memory/cpu and gpu just yet. I think the least likely culprits is your GPU and motherboard.

What is your memory clocked at?

I would kick it down to 1333 and lower your CPU multiplier and voltage a tad, till you are in a safe CPU temperature range. Then test it by gaming again. If you receive no problems, you had an unstable clock by either the memory being too high, the temperature being too high or more likely a mixture of both (considering individual stress tests are okay but gaming isnt). If you have 4 memory sticks, use only 2 for this. The difference in power usage between a slightly down clocked CPU is next to nothing and if your system is failing due to your PSU now, it will eventually fail whether it is down-clocked by a few tenths of a volt or not. This test will help toward ruling out your PSU or not.
 
tres_kun said:
Get voltage reading when it is doing nothing
And get another set of readings under furmark and prime
At 11.7Volts in IDLE it dose not sound that good
It works until 11.4V but i will not trust a PSU under 11.9V

Big voltage drop from idle to load can do nasty problems
Use software reading if nothing else is available
Click to expand...

Dont forget that the 12.7v in on the - 12v line, the + 12v line is reading 11.9v.

Problem with software readings as on mine i get stupid readings.

Looking at the cpu temp is a little high at 95c!
 
The temperature is 95 degrees but you will find laptop cpu working at that temperature
That is not a excuse but it dose not seem that the high temperature is the culprit
After he finds the reason for power off he should deal with the temperature
I am aware about voltage readings in software is dumb but the reading difference is what i am after
More than 0.5V is a reason for concern
 
You must log in or register to reply here.
Back
Top Bottom