3.3v and 5v are low

[walter.caorle]

Hello to all. I am a new user from Italy. I have a problem with power supply, an old strider 1000w

Amd Fx8350@5ghz+SwiftechH320
990FXA-UD7 Rev1
(2x2Gb) G.skill Eco 1600mhz @1866mhz
Asus ENGTX580 DCUII
Silverstone Strider ST 1000W
HD: 500gb+1tb+1tb



3.3v and 5v are low ....

Same PC some time ago ...


to be honest it's a bit that I had not noticed ... but since I am going to change the motherboard, too old, I wanted to solve the problem .... you tell me .. is leaving ...?

 

[EarthDog]

spec is 5%. If it's within that its good.

That said software is questionable on it's accuracy in the first place. So I woukdnt trust those values without testing with a multi meter. If you aren't having problems I woukdnt sweat it too much.

 

[walter.caorle]

I have the same values ​​even bios. I might change my motherboard that was already in my plans and see if it's ok.
thanks for the reply

 

[ATMINSIDE]

BIOS is just as flaky as software for voltage readings.
Use a multimeter.

 

[walter.caorle]

ok ... find a multimeter and try it. will update tomorrow..

 

[Bobnova]

Yup, what these guys said.
BIOS is better than software, usually, but not always.
Multimeter is the only way.
Even cheap ~$7 (whatever that might be in Euros. 5 maybe? Whatever) digital multimeter is fine. It may not be perfectly accurate, but it'll be close enough.

I've bought three cheap meters, one matches my shop's $400 meter perfectly, the other two are off by 0.3% (three tenths of a percent) or so. Pretty good for six bucks.

There's a link in my sig to the testing procedure.

Short version: Set multimeter to Volts, in the 20v (ish) range.
Black multimeter lead goes into one of the two center pins of a Molex (old style HDD power) plug. Those are black wires and are ground.
Yellow wire to the Molex should be ~12v. Red should be ~5v.
Orange, in the motherboard power connector, should be 3.3v.

 

[walter.caorle]

I am .... do not ask me how I found a multimeter at this time .. in Italy it s 3 'o’clock'!

 

[Bobnova]

lol, good work!

Time for a new PSU though I'm sorry to say. Maybe doublecheck the 3.3v again to be sure, but if it really is ~3v, the PSU has had it.

 

[walter.caorle]

i tried another rail, 3.15v in idle. under stress test 3.06v

 

[Bobnova]

It's close, but 3.135v is the minimum for the spec.
It's not likely to be a problem right where it is, but if it keeps heading downward you'll have issues. Arguably more importantly, we don't know why it's sagging, and it may be doing other unpleasant things that don't show on the multimeter.

I'd start looking into getting a new unit.

 

[walter.caorle]

unfortunately ... I think i get the corsair ax. the 760 here costs 144 € (200 $). the 860 on offer, 158 ($ 220). for the little difference i would say the 860 ... what do you say ..

 

[EarthDog]

I say save the $20 as a 760 is plenty for any single gpu/cpu combo and a lot of dual cards.

 

[wingman99]

The corsair AX 760 is plenty you can even run SLI.

 

[BobbyBubblehead]

duff 3.3v is not good to crystal oscillators. its pretty system critical because your timing will be all over the place.

and a duff 12v means your square waves are rounded thus borking up the timing also.

good power cures a multitude of pc sins

 

[Bobnova]

Crystal oscillators are run on 3v3. They are run at a far lower voltage, by whatever chip they are hooked up to.
There are plenty of things that won't appreciate low 3v3 rail voltages, namely anything powered by it directly. Anything that regulates the voltage further (many things) won't care until it gets very low.

The 12V rail has absolutely nothing to do with square waves being rounded or not rounded.
It's primarily used by HDD motors (which have active control on them. As the voltage drops they will eventually not be able to sustain their rated speed. This is not fatal, until the HDD controller decides something is wrong and shuts it down), by fans (which don't care at all), and is regulated down to power most things. As it drops the regulators have to draw more current from the rail to supply the wattage required by the devices they are powering. Eventually either the voltage will drop below the minimum required to regulate well, or more likely the current draw will get high enough that something goes BANG, or most likely the controller chip will say "Hey bro the 12v ain't even close, I'm going on strike" and shut things down. Or of course the PSU explodes and contemplates feeding 380v into the 12v rail.

There aren't very many square waves that actually have to be square in a modern computer anyway, really.


Now, all of the above is not to say that it's OK to have out-of-spec low (or high) PSU voltages.

 

[walter.caorle]

today ...... my 2 hd went off. first one. I then rebooted the computer and it turned on again. during a copy and paste then it was off the other. luckily it was not that the system ...

 

[Silver_Pharaoh]

Is that a BSOD as a background?
Also did you find a new PSU?