Bad Power Supply?

Status
Not open for further replies.
M

modenapsu

Posts: 11   +0
I just signed up after googling a problem I was having and I found a forum post here that recommended a way to check what that problem might be.

I download, installed and ran Speedfan to check the voltages of my power supply. See the attached image for the screenshot. I am suspecting that my + and - 12V and my + and - 5V should be telling me that my power supply is bad.

I am also getting random computer shut downs (just completely turns off) while doing anything on the computer from browsing the internet to playing a game. It happens even when I am not using the computer and it occurs every few hours.

The power supply in question is a KINGWIN Mach 1 ABT-1220MA1S 1220W

Think my power supply is bad?

Since I cant link a picture the readings are:
+12V: 5.12V
-12V: -1.17V
+5V: 5.08V
-5V: -7.17V
+3.3V: 3.38V

I posted a pic at
modena135(dot)squarespace(dot)com/storage/PSU(dot)jpeg
 
If your PSU voltages are outside of these values then yes you need to replace it.

Voltage Rail........Tolerance.........Minimum Voltage.......Maximum Voltage
+3.3VDC............± 5%...............+3.135 VDC..............+3.465 VDC
+5VDC...............± 5%...............+4.750 VDC..............+5.250 VDC
+5VSB...............± 5%...............+4.750 VDC..............+5.250 VDC
-5VDC (if used)...± 10%..............-4.500 VDC...............-5.500 VDC
+12VDC.............± 5%...............+11.400 VDC.............+12.600 VDC
-12VDC..............± 10%..............-10.800 VDC.............- 13.200 VDC
 
Speedfan is known to show low 12v readings. I'd probably confirm by running another monitor- Everest is usually pretty close-here is the latest free build http://majorgeeks.com/EVEREST_Ultimate_Edition_d4876.html Any build newer than this is trial and pay.
What do your BIOS voltages read ?
While your voltages might be off I would have thought that a reading of 5.12v wouldn't even power up the board- I'd suspect a software glitch.

Personally I'd probably get the PSU tested-and probably start by buying a voltmeter (they should be relatively inexpensive at your hardware/electronics store), and if the BIOS hardware monitor shows low voltages then that would be definite.
Hopefully you're covered by warranty.

EDIT:
The disclaimer at Speedfan acknowledges that, system dependant, the readings aren't foolproof
http://www.almico.com/sffaq.php
And what is probably an extreme example here...
http://www.dslreports.com/forum/r19381458-12V-readings-very-low
 
I have a power supply tester from Frozen CPU: FrozenCPU ATX 2.0 Ultimate LCD Power Supply Tester (20/24 pin ATX, SATA, P4, Xeon, PCI-E, Floppy, 4 pin)

I cant post a link but that is what it is called there, maybe search for it there and let me know if that is what you mean.

What else could account for my computer just turning off like that? That seemed like a dead ringer to me for PSU failure. Any logs or anything on my computer I could post to help figure this out?

<grabbing everest now. havent used it in a while, what test or options should I be using in everest to check my psu?>
 
It is as you say more than likely a PSU problem- but if you were certain, why post ?
We are just offering a troubleshooting guide to verify the problem (if possible).

The Everest screen has a side menu on the left hand side
Click to expand "Computer" and click "Sensor" (I think) and it should show all relevant voltages.
Checking your BIOS's voltages (Phoenix-Award BIOS's for instance has this under "PC Health Health Status" -if you can't locate it then post your motherboard model and I'll post back on how to reach the correct section) is probably the quickest and surest voltage values you're likely to find. Everest shouldn't be too far out.
 
Dividebyzero - hey I am sorry for my tone, I didn't mean to sound like I knew it was a PSU problem. My tone was meant to be like "scratching my head in confusion". I posted because I saw some of the stuff you guys had posted here on your boards and it seemed like those that frequented this forum were very knowledgeable. Again, sorry for the tone if I implied that I kinda know that it is my PSU, because I don't know that it is my PSU, it just seems that the PSU is the culprit.
 
From Everest:

Voltage Values
CPU Core: 1.31V
+3.3V: 3.38V
+5V: 5.08V
+12V: 12.29
VBAT Battery: 3.25V
DIMM: 1.65V
GPU1: GPU VRM: 1.04V
GPU2: GPU Vcc: 3.27V
GPU2: GPU VRM: 1.04V

Current Values
CPU: 29.02A (42.50A when moving a window around)
GPU1: GPU VRM: 4.35A
GPU2: GPU VRM: 4.35A

Power Values
CPU: 39.66 - 59.10 W (thats the high and low range I have seen it go through while typing this)



So do all those look ok? Is there a way I can put my PSU through a test?
Anything else I can do in Windows7 test wise to figure out any other things that might be the problem or any logs I can post that could aid in figuring out what it could be that is causing my computer to just randomly shut down?

My CPU Specs:
Core i7 920 Bloomfield @ 2.66Ghz | Gigabyte GA-EX58-UD3R Rev 1.0 - Bios F8I | Nvidia GTX 295 Video Card | OCZ Reaper HPC 6GB (3x2GB) DDR3 1866 PC3 1500 | KINGWIN Mach 1 ABT-1220MA1S 1220W | Western Digital VelociRaptor - 150GB 10,000RPM 16MB Cache SATA 3.0GB/s [WD1500HLFS] | 2x Western Digital Caviar Black - 1TB 7200RPM 32MB Cache SATA 3.0Gb/s [WD1001FALS] | WesternDigital Caviar SE16 - 750GB 7200RPM 16MB Cache SATA 3.0Gb/s [WD7500AAKS] | Western Digital Cavier Blue - 500GB 7200RPM 16MB Cache SATA 3.0Gb/s [WD5000AAKS] |Razer Lycosa/Lachesis | Windows 7 Pro x64


The PSU tester I have says that the PSU is within proper working limits.
 
All good.No offence taken.
I'm sure you wouldn't want anyone here to give you a definitive YES/NO answer based on limited information. So we like to tread carefully and eliminate possible causes of problems as well as targeting the obvious. It can save problems down the road, and possibly uncover faults in related hardware at the same time.
I usually supply a lot of info - but I tend to think of others that will follow with similar problems or questions. A step-by-step troubleshooting guide for the next person who may need the help and reads the thread may save them some valuble time or lead them to solve the problem themselves.

Post back when you have a better idea of the voltages.
The PSU tester you linked to looks like it will do the job nicely- certainly more upmarket than the one own.
 
The voltages look fine from Everest.
I would test with the voltmeter as well.
You can test the PSU with a program called OCCT -find it here http://www.ocbase.com/perestroika_en/index.php?Download
OCCT also has CPU and Graphics Vram tests-should you ever need to use them.
And this thread has some info about it's implementation and the graphs the test produce.
https://www.techspot.com/vb/topic141283.html
It's is a good software test for PSU voltage ripple that can cause instability in the system - you'll notice that it measures the ripple in addition to showing the voltage draw.
The ATX specification (where HellfireXD got the tolerances in his post) also calls for a ripple of no more than 120mV (0.12 volts) on the +12v rail/s -anything outside that is bad juju.
 
So I used the System Stability Test in Everest. The system ran the CPU test for 1.5 minutes and I stopped it. No problems, CPU has at 100C. Ran the FPU test, the comptuer 'turned off' in the first 10 seconds. Turned computer back on and ran the System memory test, the computer again 'turned off' at about the 40 second mark.

I do have the computer OC'd to 3.2GHZ and idle CPU temp is around 58C. I have a ton of system mod stuff to add, like nicer cables, new fans, etc but I want to solve this problem before I do anything like that. I also have a new heatsink and fan for the CPU because when I bought mine, it was highly rated but now it is known (a year later) that this heatsink/fan is hardly better than the stock heatsink/fan. Again, I want to solve this problem before I put that in.

Doe that above turning off during those tests point to anything? Think re-setting the RAM could help or anything like that?

The PSU tester I have says that the PSU is within proper working limits.
 
Your board uses the Award/Phoenix BIOS- and as your board is similar to my own, getting around it should be fairly easy.
If you want to check voltages in the BIOS:
On the main screen, on the left hand side you'll see "PC Health Status"- select this and the section will show you all the voltages (Vcore, RAM, +3.3, +5, +12) as well as fan speeds (if connected to the motherboard), tempretures, and the ability to enable warnings and smart fan. Navigate from tthis page using the "Menu level" in the right hand pane.
You can also access the systems voltages and temps etc. through the EasyTune 6 software that came with your board (if you have installed it from the utilities disc).
Happy testing.
 
Here is something new: when I tried running OCCT, it said my CPU was too hot, I checked the temp in CoreTemp64 and it was north of 80C at idle OC'd to 3.2ghz. It used to be at 58-60C at that OC. So I put the CPU back to normal, 2.66ghz and it is def running hotter than it was 2 weeks ago, at idle at 2.66ghz, I am getting a CPU temp of 45C.

Any thought?
And I really appreciate you helping - I would be flipping out if I was going at this alone.
 
OCCT - even when my CPU is set at no OC - wont run.

It stopped on the Power Supply test at 1 min 45 sec and a 'CPU too HOT' message. It did give me 8 different images, do you want to see any of them?

I did hear a fan really pick up some speed during the test, not sure if it was the fan on my CPU or GPU though.
 
Ah.
Yes. I think the OC'ing probably has a lot to do with it.

#####EDIT - I just saw your post READ THIS FIRST ###########

Check the cooler to make sure all four "leg" pins are firmly latched. I'd probably lift the cooler in any case -wipe the generic grease off both the cooler and CPU heatspreader with some Arcticlean/Isopropyl and lint free cloth and reapply a good TIM and reseat the cooler.


Your CPU temp is way too high -the CPU will be throttling at those kind of temps.
I'd get into the BIOS and set the default settings (2.66GHz) until you ditch the stock cooler (check the end of the post)
Then run the PSU test in OCCT to check the ripple measurement
The board should be stable at default settings- but I'd run CPU stability testing just to make sure. Memtest86 for RAM, and your choice of OCCT (good all-rounder), IBT and/or LinX (will produce the max. heat from the CPU), Orthos and Prime95 (both accepted benchmarks for stability but they should ideally run 8-24 hrs) and good old Crysis - it has the uncanny knack of finding instability where nornal stress testing doesn't.
What testing did you do to the system to make sure the overclock was stable?

At the end of testing you are happy with the voltmeter's readings and OCCT's ripple measurements then the prime culprit is likely your CPU temp and/or OC'ing instability- in which case you might like to start up another thread regarding system testing for stability and/or OC'ing.
 
Here are the images for the power supply at a 6 minute test (just wanted it to survive the test!)

1.png

2.png

3.png

4.png

5.png
 
I posted some images of the OCCT Power Supply test - ran a 6 minute test just to ensure that it would finish the test.

the images are at

Hey I have images from OCCT, I ran a 6 minute power supply test (i made it that short so that it would survive the test).

Mind glancing at them quickly? thanks so much in advance!

the images are at
modena135(dot)squarespace(dot)com/storage/1.png

there are 10 pics - 1.png, 2.png, etc, 10.png
 
The images:
Temp is too high. Anything over 80C and you'll end up killing the CPU. The Intel whitepaper (specification ) for the 920 is 67.9C as max operating temp- a little unreasonable, but I wouldn't run the CPU over 75C.
Did you check the cooler mounting ? At stock Vcore your idle temps should be very low 30's.
I'd say that the cooler isn't seated properly- the stock Intel cooler for LGA1366 is a PITA -you'll find numerous posts around the web regarding what looks like a good fitting cooler that isn't.
For example: a 3.2GHz CPU speed with stock Intel cooler should give you an idle temp of approx 35C and 55-60C under full load in a moderate-good tower chassis
The OCCT 12v reading is borked. OCCT isn't picking up the motherboard sensor at it's correct voltage level. Discount the result at present.
The +3.3 and +5 voltage look fine- as they were reported earlier.
The GPU temp is way too high -GTX295's run hot anyway, but airflow through the chassis can't be very good. What chassis? is the cooling standard ? More to the point why are you stress testing the graphics card ?

I'm assuming that you reset BIOS to default. Is this the case ? ( Vcore, QPI/Vtt voltage, IOH Core and DRAM voltage should all be set to auto if it has)
Under PC Health in your BIOS
What is Vcore reported as ?
Is the +12v within the normal range ?

If your BIOS is set to default and the cooler is mounted correctly and you can only get through 6 minutes of OCCT then you have something seriously wrong with your system.
Go through the thread and make a checklist of the things you need to do. If you work through them methodically you'll save yourself a lot of grief down the track.
 
I will get back to you on this tomorrow.

2 things:
I am not running stock cooler now, I am running this: Thermaltake V1 AX CPU Cooler CL-P0508
I have in my possession a Cogage True Spirit and Yate Loon fan that I plan to use instead of the Thermaltake in the near future.
Also, the GTX 295 does not even start to throttle down until over 120C. Going to be adding more cooling to my system, I already have the parts, just trying to get this all solved before I move ahead with those upgrades. I currently have above adequate air cooling in a Lian-Li case.
 
Status
Not open for further replies.

Similar threads

Shawn Knight
Lian Li introduces T-shaped power supply for dual-chamber PC cases
Replies
7
Views
293
mbrowne5061
M
NukePunk
New PC this year, any help welcome :)
Replies
9
Views
1K
AlaskaGuy
AlaskaGuy
Vilpted
12v@60a =24v@30a Opinions?
Replies
2
Views
2K
wiyosaya
wiyosaya

Latest posts

Back