What could be causing low voltage readings on motherboard?

doveman · Post by **doveman** » Sat Mar 20, 2010 7:54 am

My Video card output on my P4 system went blank whilst I tried playing a game the other day. It was the first time I've tried since replacing the previous card that died with a second-hand 2600XT, which has been working fine with MediaPortal for the last few months.

If I test the voltages coming out of my DC board on a Molex, the 12v is OK at 12.05v and the 5v is about 5.07v. However, Hardware Sensors Monitor (the only program that seems able to correctly report the 12v line on this board) shows 11.61v, 5v and 3.15v, which if accurate seems likely to cause problems.

So is there anything on the motherboard that I might be able to fix that's causing these low voltages?

lm · Post by lm » Sat Mar 20, 2010 10:24 am

In other words, -3.25% on 12V and -4.5% on 3.3V lines.

Now, according to ATX specification 2.2, section 4.1.4 Voltage Tolerances, both lines allow +/- 5%, so the numbers are fine. Why do you say they are likely to cause problems?

Your actual problem lies elsewhere. Forget about the voltages.

doveman · Post by **doveman** » Sat Mar 20, 2010 5:00 pm

It's only 'cos ryboto reported in another thread that he'd been getting 11.83v at idle, which dropped to 11.4v with the video card loaded, at which point he had random issues.

He said he solved that by shortening the cable from the DA-2, which raised the voltage a bit and solved his problems.

However, as I measured 12.38V coming out of the DA-2 I couldn't see that I needed to hack the cable and thought if there is a voltage problem it must be further down the line, such as the motherboard. You might be right that there's some other problem with the motherboard or something else that's nothing to do with voltage. I haven't got a clue what it could be though.

Vicotnik · Post by **Vicotnik** » Sat Mar 20, 2010 9:02 pm

You should start looking at bottlenecks. If you have 12.38v at the brick but only 12.05v at a molex then you have a little drop there. I had problems with a low +12v rail when I first set up my PicoPSU setup. Back then I used the barrel connector that came with the Pico. I have since then skipped that one. I don't use any connector at all right now, and if I would like to make the system a bit more modular later I would probably use a Mate-N-Lok 15A connector. I also let the 4pin +12v connector draw power directly from the brick, not from the PicoPSU.

doveman · Post by **doveman** » Sun Mar 21, 2010 4:24 am

I'm already running the CPU 4pin +12v and the GPU supplemental +12v directly from the brick.

I tested the voltages last year but I just tested again and with the PC powered down, there's 12.4v at the GPU power connector (coming straight from the brick). With the ATX connector disconnected from the motherboard, when I power up the PSU there's 12.38v and 5.06v on the Molex and the ATX connector, which also reads 3.27v and -11.95v, so I can't see any problem there.

Vicotnik · Post by **Vicotnik** » Sun Mar 21, 2010 5:13 am

Then it's like lm says, that your problems doesn't have anything to do with the voltages.

doveman · Post by **doveman** » Sun Mar 21, 2010 9:05 am

Surely all I've shown is that the power supply is working properly. There could still be a voltage problem on the motherboard, as ryboto had when the video card was under load and the voltage dropped from 11.83v to 11.4v.

Can you think of anything else that would cause the video card to shutdown whilst playing a game but work fine the rest of the time?

I wouldn't have thought it would be temps, as it's still using the original noisy fan, which spins up when gaming. I had trouble with this motherboard and the previous graphics card, which completely died eventually. It seemed very fussy about how it was aligned in the slot and sometimes would boot into windows but then crash when I did something graphical, such as playing a video or loading a game.

I'd imagine that was due to some physical fault in the card rather than the slot though, as obviously the card gets knocked around a lot more, what with fitting Acceleros and so on. This doesn't seem the same anyway, as I had no problems getting the new card to work after dropping it in and I had the game running for about 45mins before the display went blank.

lm · Post by lm » Sun Mar 21, 2010 12:49 pm

I guess you could try with a regular PSU, just in case.

Also try another video card on that system and that video card on another system?

doveman · Post by **doveman** » Sun Mar 21, 2010 1:06 pm

Yeah, that's a bit tricky unfortunately as I haven't got a working spare PSU, a spare AGP card (not one that can handle games anyway) or another AGP motherboard.

Anyway, surely I've established that there's nothing wrong with the PSU haven't I?

doveman · Post by **doveman** » Sat Mar 27, 2010 7:44 am

bump, has everyone given up on me

Wibla · Post by **Wibla** » Sat Mar 27, 2010 9:43 am

Well, you've kinda isolated that its not a power supply issue, right? then there's just the whole rest of the computer to troubleshoot...

BTW having dodgy voltage readings in bios/"HW monitor apps" isnt unusual, i've not come across too many motherboards who actually reports all the voltages 100% correctly.

doveman · Post by **doveman** » Sun Mar 28, 2010 11:14 am

Thanks, I just wanted someone else to confirm that my tests confirmed the PSU was fine.

Guess I'll ignore what the HW monitor app says then if they're not accurate. Just have to try and find out where the problem is now....

doveman · Post by **doveman** » Wed Mar 31, 2010 11:41 am

I wonder if a difference between the 12v at the AGP slot and the 12v at the 6-pin GPU supplemental power socket could cause problems?

As the latter is coming straight from the brick, it's about 12.4v, whilst if the BIOS readings are correct the AGP/motherboard 12v could be around 11.6v.

Would the graphics card be able to tolerate that amount of difference between it's two power sources?

BlackWhizz · Post by **BlackWhizz** » Wed Mar 31, 2010 12:04 pm

Those sensors aren't calibrated. Thats the biggest problem of em. You can fix it with a BIOS update (if it fixes it). I always neglect the bios en hook my multimeter up to a molex

doveman · Post by **doveman** » Wed Mar 31, 2010 4:47 pm

That's fine for checking the PSU, but I don't think it shows anything about the voltages at the CPU/GPU, as the motherboard has various voltage related components that I imagine could become faulty.

It's a shame motherboards don't have a handy breakout post-these components so that we could easily check the voltages with our meters.

Vicotnik · Post by **Vicotnik** » Wed Mar 31, 2010 10:49 pm

doveman wrote:It's a shame motherboards don't have a handy breakout post-these components so that we could easily check the voltages with our meters.

Usually it's possible to find the relevant points where to stick the multimeter probes though.

doveman · Post by **doveman** » Sat Apr 03, 2010 8:48 am

Yeah but the points are so small I run the risk of shorting something!

I was just watching an AVI with MediaPortal and it shut down on me, by which I mean the screen went black and the audio stopped although the fans continued spinning. One of the log files that MP creates seems to have been interrupted half way through a line, so it seems that probably all processing did stop.

Any idea what could cause this sort of fault? Obviously the GPU wasn't doing anything strenuous at the time, so I'd be surprised if it was to blame.

doveman · Post by **doveman** » Sat Apr 03, 2010 6:56 pm

Actually it's occured to me that it might be my old Kerio 2.1.5 firewall that's crashing the PC, as I found it was causing BSOD on my main PC.

Quite why I'd be getting a blank screen rather than a BSOD on my HTPC I don't know, but I've replaced the firewall now so I'll see if that does the trick. Damn shame no-one's brought out one as simple as Kerio though.