The importance of testing your psu with a multimeter

[John]

It is crucial that you know your TRUE voltages on the 3 main voltage rails, present on all ATX PSU's. Many people still rely on means of software, such as MBM, Speedfan, and the BIOS, but these are all very unreliable. I have been able to prove this, and what it brought to my attention, that those who trust software, are potentially putting their software in danger, particularly those who have adjustable rails in their PSU.

How to guides:

Exteme Overclocking PSU Testing Guide
Driver Heaven PSU Testing Guide
How to test additional voltages on the DFI Lanparty NF4 mainboards -- Thanks Fresh Daemon


You want the three main voltage rails to be within 5% of the regulation range.
http://www.formfactors.org/developer/specs/ATX12V_PSDG_2_2_public_br2.pdf

I decided to pick up a $20 Digital MultiMeter from Sears and test my OCZ 520w PowerStream.

My Rig specs:

OCZ Powerstream 520w Powersupply
(5) 80mm Panaflo L1A (21dB) Case Fans
AMD A64 3200+ Winchester @ 2.6GHz 1.5v
Thermalright XP-90 HSF w/ 92mm Stealth & AS5
DFI Lanparty NF4 Ultra-D 623_3 bios
2 GB (2x1GB) GigaRAM DDR466
BFG 6800GT 256 MB PCI-E (400/1100)
Western Digital Raptor 74 GB SATA
Seagate 7200.8 250GB SATA
Black Lite-On 16x DVD
Black Lite-On 16x DL DVD ± R/RW

Rail
+3.3
+5
+12

BIOS Reports
3.30
4.97
11.96

DMMeter Reading while in BIOS
3.32
5.08
12.09

DMMeter while running Prime95 torture test
3.32
5.08
12.09

Sandra 05 SP2 (pure software reading)
3.26
4.95
11.54

As you can see the OCZ 520w PowerStream is solid as a rock, even under 100% load there are no rail fluctuations! Now you know that it is not a good idea to rely on software readings. By testing your power supply you'll know whether it is handling your system like it should or if you need to look for a better one.

 

[imported_Kensai]

 

[Kaspian]

Those meters are relatively cheap for what they do. Some times you can get them at harbor freight tools as low as $2.99 new.

 

[Suspicious-Teach8788]

Home Depot, Radio Shack has cheap as hell voltmeters.

11.54 is not that great... I sitll find Antec a little better on voltage regulation.

 

[John]

Originally posted by: DLeRium
11.54 is not that great... I sitll find Antec a little better on voltage regulation.

Please keep in mind that the Sandra 2005 voltage readings are software results. The voltages tested with the meter are within ±2% of the ATX spec.

 

[Suspicious-Teach8788]

Originally posted by: John

Originally posted by: DLeRium
11.54 is not that great... I sitll find Antec a little better on voltage regulation.

Please keep in mind that the Sandra 2005 voltage readings are software results. The voltages tested with the meter are within ±2% of the ATX spec.

Oh oh oh lol I thought you were running Sandra stress test or something and then took a volt reading... LoL. Nevermind then.... Stupid me =P

 

[rise]

agreed.

heres my 12v showing how inaccurate software reads are.

 

[John]

Originally posted by: rise4310
agreed.

heres my 12v showing how inaccurate software reads are.

 

[imported_Tick]

Yeah, every electronics geek should have a multi anyway.

 

[bendixG15]

And what instrument do you use to test the accuracy of a cheap multimeter ???

 

[Varun]

Originally posted by: bendixG15
And what instrument do you use to test the accuracy of a cheap multimeter ???

No kidding. A $10 multimeter will likely have low input impedance, and could be as inacurate as the software.

 

[skisteven1]

Originally posted by: Varun

Originally posted by: bendixG15
And what instrument do you use to test the accuracy of a cheap multimeter ???

No kidding. A $10 multimeter will likely have low input impedance, and could be as inacurate as the software.

touché

 

[John]

Originally posted by: Varun

Originally posted by: bendixG15
And what instrument do you use to test the accuracy of a cheap multimeter ???

No kidding. A $10 multimeter will likely have low input impedance, and could be as inacurate as the software.

That's why I splurged and spent $20. I'll borrow a $150.00 Fluke DMM and compare the results w/ my $20 Craftsman. The only options that the higher end models seem to offer is true rms, auto ranging, auto power off, temp, higher ac/dc/amp ranges, etc., but the accuracy for all of the DMM's appears to be ~0.5%.

 

[rise]

lol John, mine was ~30 i think.

 

[John]

300w LPM2-20 Power Supply (Generic)
Athlon 1GHz
512MB
20GB HDD
MSI K7T Pro 2
DVDRW
DVDROM
Floppy
SB Live
10/100 Nic

Tested during a Prime95 Torture Test

$140 Fluke 73-III (DC accuracy ±0.3% +1)
3.25
4.82
12.58

$20 Sears Craftsman Model 82140 (DC accuracy ±0.5%)
3.25
4.82
12.61

The rails dropped about a little bit (maybe 1-2%) when I started the Prime95 Torture test.

So as you can see there is a slight variance in the +12v reading (.03v) between the $20 Craftsman and the $140 Fluke. Personally I don't think that it's enough to be concerned about. I suppose if you were testing higher end equipment then it would make sense to spend the extra cash on a high end DMM.

 

[rise]

thanks for that John.

i just got my dmm a few months ago but i think i'll take it down to the local electronics store and see if they can test/calibrate it next week.

 

[ShadowBlade]

Originally posted by: rise4310
agreed.

heres my 12v showing how inaccurate software reads are.

is your print screen key broken?

 

[rise]

kinda hard to use the print screen and include the dmm :roll:

 

[The Pentium Guy]

"kinda hard to use the print screen and include the dmm "
. Nice monitor .

 

[rise]

Originally posted by: The Pentium Guy
"kinda hard to use the print screen and include the dmm "
. Nice monitor .

thanks man, they're pretty damn nice for the $$. worst support in the world though

 

[hurtstotalktoyou]

The power supply is either adequate or it's not. Any cheap PSU will do. It may die some day sooner than a more expensive PSU, but it will almost certainly last for several years, which is as long as most PCs continue operation.

 

[John]

Originally posted by: hurtstotalktoyou
The power supply is either adequate or it's not. Any cheap PSU will do. It may die some day sooner than a more expensive PSU, but it will almost certainly last for several years, which is as long as most PCs continue operation.

Unfortunately your "advice" may have been sufficient for the old generation (P2, P3, K6, etc.), but with 'current' hardware (ex. NF4, Athlon 64, 6800GT, 7800GTX, etc.) you should use a quality power supply with a big +12v rail, or dual +12v rails. The power supply is often the most overlooked component because people don't understand their role. It's the component that powers your entire system. A cheap 450w psu is not the same as a quality psu. The el-cheapos take their measrements at a low temp to boost the numbers. When cheap power supplies blow they usually take a few extra components with them.

 

[hurtstotalktoyou]

Originally posted by: John
Unfortunately your "advice" may have been sufficient for the old generation (P2, P3, K6, etc.), but with 'current' hardware (ex. NF4, Athlon 64, 6800GT, 7800GTX, etc.) you should use a quality power supply with a big +12v rail, or dual +12v rails. The power supply is often the most overlooked component because people don't understand their role. It's the component that powers your entire system. A cheap 450w psu is not the same as a quality psu. The el-cheapos take their measrements at a low temp to boost the numbers. When cheap power supplies blow they usually take a few extra components with them.

While it is technically possible that a cheap power supply might surge and kill a motherboard, the odds of that happening are extremely small. In fact, every case I've heard about where a PSU is accused of ruining a system, it's possible--often probable--some other factor is really to blame.