PSU instability test

miker2808

Honorable
Aug 23, 2016
49
0
10,530
hello, lately my PC is acting very very strange. (for a long time I had many weird stuff.. but now its kinda very slow [I had huge GPU instability, crashing all the time and I decreased frequency to fix that])
Now. I checked a topic which showed a guide how to check for a PSU instability. and it showed that I have to check my voltages and see if it is too low or too high
now I did that and found a very very weird thing.

Here is it what I got: (Using HWMonitor)
CPU VCORE Value 1.160V Min: 1.144V Max: 1.312V
VIN1 Value: 1.856V Min: 1.832V Max:1.864V
+3.3V value: 3.168V Min: 2.992V Max: 3.184V
+5V Value: 4.752V Min: 4.488V Max: 4.776V
+12V Value 8.252V Min 8.085V Max: 8.252V
VIN5: Value: 1.488V MIN 1.376V Max 1.488V
VIN6: 1.024V 1.000V 1.32V

so.. for +12V it is far from the minimum it should be (about 11.6 I believe)
now I dont know for sure if this is actually bad.. maybe its not calculated well and etc..

MB: Asustek P5KPL-AM
Intel Core 2 Quad Q8400
Windows 10 x64 home
2x2 ram dual channel 333mhz (400 and 800)
Radeon RX460 2GB
WDC WD2500KS
PSU: Zalman 450W

Would like to have some guide with what to do
 
Solution
HW Monitor seems to be way off. I would throw those results away. What I do see common from all monitors is the 5V is too low. 4.75 is the lowest allowed value for the 5V rail, and if speedfan and others are showing results <4.75V that's your issue. Using a Voltmeter to double check the values would be my first step. But it does seem like you need a new PSU.
The 3.3 V is 9% below rated voltage
The 5.0 V is 10% below rated voltage
The 12.0 V is 30% below rated voltage

The readings pon every rail are unacceptable. The allowable variation per the ATX spec is 5%. At worst, I look for 2.5% for a light duty office or web box ... down to < 1.0 % for a high end gaming box or overclocked workstation.

Two possibilities:

1. The PSU is indeed outta whack and should be replaced. A voltage meter would answer this question.

2. Sometimes the utility misreads the sensor, doesn't recognize the sensor or the sensor is faulty

3. The utility is out of date. I find HWiNFO to be the most reliable and by far the most complete / exhaustive utility available. I have used Speeedfan in the past but in this decade, it's usefulness has dimmed as most boards today come with their own utility and a Speedfan support profile maybe doesn't get developed for that board. You can find the temp sensors that Speedfan supports here:

http://www.almico.com/forumsensors.php
 

miker2808

Honorable
Aug 23, 2016
49
0
10,530

Thanks for the replay I tested other utilities (in total 3. HWM, Speedfan and the Bios itself)
here are the results.

Test-1 HW Monitor
V5= 4.752
3.3V= 3.168
V12=8.252
----
test-2 SpeedFan
V5=4.76 - .66
3.3V= 3.15v
12V= 12.25 - .30
----
test-3 Bios
Core- 1.320v
3.3= 3.136v
5v= 4.736v
12v= 12.249v

Here are the results.
now that it sums up, I would like to get the advice. PSU is really that far from minimum voltage and because of it the PC is malfunctioning? (randomly freezing, especially in-games) (GPU is crashing at base clock. works only under 30%)

I thought about getting new Parts. SSD, MB (CPu and ram with it)
in conclusion:
Just tell me what to do
Thanks in advance and of course for the fast response
 
HW Monitor seems to be way off. I would throw those results away. What I do see common from all monitors is the 5V is too low. 4.75 is the lowest allowed value for the 5V rail, and if speedfan and others are showing results <4.75V that's your issue. Using a Voltmeter to double check the values would be my first step. But it does seem like you need a new PSU.
 
Solution
Well it's not like you have a possible warranty issue as those numbers show that all the rails on the PSU are essentially within the ATX specification. Your reading of 4.74 (4.376) is well within the margin of error for reading voltages so it won't cut any ice on a warranty claim. There is no way to justify a claim that 0.01 voltage variation might be causing some problem. If it was new, you would be able to return it ... possibly. But if in use, no shot.

As I said above, a volt meter is the way to go if you want to verify but not a $12 model from your local hardware store. Borrow one from somebody who works in the electronic / electrical field.

For whatever reason, HWM is not picking up the correct data for the 12v rail .... is it the latest version ? I really can't give you an assessment of other utilities, and how they have kept up update wise, as have only been using HWinFO for the last few years as it measures everything you can think of and then some.

Zalman 450 is kinda like saying "Blue Chevy" ... w/o the model number, it's impossible to assess any quality concerns

ZM450-FX is made by Sirtec, non-modular, safe to 432 watts @ 12v, 3 year warranty, Bronze
No reviews from reliable sites available

ZM450-US (USA only) is made by CWT, non-modular, safe to 372 watts @ 12v, 3 year warranty, Not Certified
HardOCP gave it a fail rating

ZM450-P was made by Andyson, non-modular, safe to 360 watts @ 12v, 3 year warranty, 80 Plus
No reviews from reliable sites available

ZM450-GS is made by CWT, non-modular, safe to 360 watts @ 12v, 3 year warranty, Not Certified
No reviews from reliable sites available

In short, I wouldn't put any of those in a build. Zalman does sell / has sold some decent PSUs (primarily HP series) but there's none that I'd use in the 450 watt size.

If you are in the US, the Seasonic S12 / M12 in either 520 , 620 or 650 watt sizes are a steal at current prices.


S12 - 430 ($37)
https://www.newegg.com/Product/Product.aspx?Item=N82E16817151074&

S12 - 520 ($40)
https://www.newegg.com/Product/Product.aspx?Item=N82E16817151094

S12 - 520 ($45)
https://www.newegg.com/Product/Product.aspx?Item=N82E16817151096

M12 - 520 ($45)
https://www.newegg.com/Product/Product.aspx?Item=N82E16817151093