Question PSU Testing Shows High Voltage

Jun 9, 2019
8
0
10
I'm troubleshooting an ASUS P5B motherboard that won't post, starting with my Power Supply. Using a multimeter, and referencing this guide, this pinout chart, and these tolerance specifications, here are my results:

First, the power cord from my receptacle is giving 122.3VAC, which is out of the -6% to +10% tolerance for 110VAC (103.4VAC to 121VAC). So my PSU is getting slightly too much voltage, and the story continues with the 24-pin connector. I tested a few other receptacles and they all put out around 122VAC.

All of the +5VDC pins (#s 4, 6, 9, 21, 22, 23) are putting out 5.3VDC, which is above the tolerance maximum of 5.250VDC. The +3.3VDC and +12VDC pins are all within tolerance.

Is a PSU supposed to condition the incoming voltage before outputting to the motherboard? Since mine is passing higher voltage than spec, does this mean the PSU is bad?

Thanks!
 

clutchc

Titan
Ambassador
First, which multi-meter are you using?
The residential line voltage of 122VAC you are reading may not be out of range. Residential voltage is actually supposed to be a nominal 120VAC in most areas (US). Mine is 122.10VAC, in fact, as measured with my trusty Fluke 117.
 
Last edited:
Firstly, if your PSU is one of the modern PFC units, its input voltage tolerance is anywhere from 100VAC to 240VAC, with a line frequency from about 43Hz to 63Hz. The input voltage can vary between those two extremes, and the PSU just isn't going to care all that much about it. (Personally, I prefer to run my PSU on 230VAC, because it loads the utility KW/h meter more favorably.)

Secondly, PSUs do not "condition" the mains voltage. You need a power line conditioner to do that, and the best and easiest way to achieve that is to install a UPS between the power cord of the PSU and the power mains connection point.

Thirdly, within the United States of North America, you will typically find the nominal "110VAC" mains voltage varying between about 109VAC to above 125VAC, and all is well--though your incandescent lamps will show a bit of variation in brightness.

Lastly, you really should post the exact brand and model of the PSU you are testing. If it's a cr@p unit, you'll probably hear about it, and that would explain the poor output regulation that you are seeing. It wouldn't hurt, also, to mention the specific digital multimeter that you are using for tests, because a low quality meter would also explain spurious measurements.

As implied, earlier, you will never go wrong when using a Fluke for such tasks.
 
Jun 9, 2019
8
0
10
What exact power supply make/model do you have? What is the voltage on pin 8?

The PSU is a "WT520q" 520W, and a quick Google search leads me to believe this is a cheapo unit. The PC was given to me, so I can't complain too much!

Pin 8 (gray wire) is described as "Power On", so I didn't measure it. I'll have to go back and do that tomorrow.
 
OK...that meter has 0.5% of full scale accuracy, which is going to translate to 0.1V of potential error on the 20VDC scale. That means that your specific measurements may not necessarily be out of spec.

As TJ mentioned, you should check the Power_Good terminal on the 24-pin connector (grey wire at pin #8). If it's at +5VDC, then the PSU thinks that it has stable output rails. If the Power_Good terminal is at, or very close to, 0VDC, then the PSU is saying that things are internally "out of whack", and the motherboard will refuse to post.
 
Last edited:
Jun 9, 2019
8
0
10
Pin 8 is is the 'power ok' signal. It's used for the PSU to tell the motherboard of the output voltages are good (i.e. within spec).

I can't find any info on that PSU, which almost always bodes poorly for the quality of the unit.

Pin 8 is showing 5.29VDC. I also tested pin 20 (since it was the other one I skipped before) and it shows -4.87VDC.
 
Jun 9, 2019
8
0
10
As TJ mentioned, you should check the Power_Good terminal on the 24-pin connector (grey wire at pin #8). If it's at +5VDC, then the PSU thinks that it has stable output rails. If the Power_Good terminal is at, or very close to, )VDC, then the PSU is saying that things are internaly "out of whack", and the motherboard will refuse to post.

I'm seeing 5.29VDC at pin 8. Does this information rule out a bad PSU? What can I look at next?
 
Jun 9, 2019
8
0
10
Well that PSU is probably still junk and should be replaced, even if it isn't the issue here.

What are your full system specs? What happens when you try to start the PC? Nothing at all? Fans and lights go on but no display? Can you reach the BIOS at all?

If I can rule out a bad motherboard, I'll get a new PSU. Otherwise, I'm going to refrain from throwing more money at this rig until I diagnose the root cause of the problem.

Here are the specs;
PSU: WT520q 520W
Motherboard: Asus P5B Green
CPU: Intel Core 2 Duo 6600 @2.4GHz
Graphics Card: ATI FireGL V3400 PCIE 128MB
RAID Card: Syba SY-PCI40010
RAM (DDR2 SDRAM): Samsung M378T5663QZ3 2GB and Kingston KVR667D2N5K2 2GB

Symptoms;
Even when the rig was "working", it would occasionally shut down on its own, which I believe was hardware related as it was running Linux. I would also get messages about "BIOS Settings Changed..." during reboots and startups sometimes.

Now, I can't even get the system to post. A few days ago, I was able to get booted up after clearing CMOS and moving a single stick of RAM around the 4 different slots. Turned the box off, connected power to my two SATA HDDs and it never posted or booted since.

Troubleshooting steps;
I've disconnected everything from the PSU and Motherboard except the Graphics Card, SATA Card (although no power to the connected HDDs) and 1 stick of RAM (alternating between the two listed above).
I installed a new CMOS battery.
I disconnected IDE and power from CDROM and OS HDD.

What next?
I have two brand new sticks of Samsung M378T5663EH3 2GB RAM, but I don't want to open the packages until I've exhausted other testing of the Motherboard.
 
Last edited: