PDA

View Full Version : Low ATX 12 Volts rail



blanco
01-11-2012, 05:02 AM
Been looking at a problem on a friend's desktop. Freezing (hanging)
for long periods and I know that there are no viruses or corrupt files
Checked the RAM modules by substitution and the sockets.
Convinced that the HDD is ok and suspected another hardware problem.
Running the diognostic SIW software shows the ATX PSU 12 Volts rail to
be hovering around 7.6 Volts. Haven't measured this but the figure shown
is surely a problem as all other Voltages are within spec.
I believe the 12V would be regulated at +/- 5% and it's well below that,
so do you think that would be the cause of the symptoms described?,
assuming that there is nothing dragging down the 12V rail.
Have advised him to purchase a direct replacement ATX PSU.

Frank_sumbody
01-11-2012, 06:35 AM
It is very possible that one rail might of dropped compared to the others, the Volts is measured from one output and is feedback into the primary circuit to correct the output, So that if the house mains supply drops the feedback rail to the primary circuit drops, this in turn adjusts the mark-space ratio to drive the primary circuit harder to provide the output higher to the correct voltages.

HOWEVER:
This works fine while the supply is new, but as the power supply unit ages, the parts like Electrolytics fault, their prime killer been 'heat', these Electrolytics can be checked with a device called a ESR (Equivalent Series Resistance) meter, when Electrolytics are new the ESR is low, but with the help of heat the chemicals dry out pushing the ESR up, Sometime you can just replace Electrolytics and this might cure the power supply problem, but then the time and effort to replace all the Electrolytics it might be better to replace the power supply, they are cheap today

I used to fix those when 286 first came out and the power units cost a arm and a leg, most of the time they where worth repairing, but not now it might be beter just change the power supply.

If you have done Physics two, you may of done a experiment to find the internal resistance of the battery, and once again as the battery goes flat or ages the internal resistance also goes up to the point where a car battery will not start the car. so what I am saying is the internal resistance of a battery and ESR of Electrolytics are very similar.

Agent_24
01-11-2012, 10:26 AM
Running the diognostic SIW software shows the ATX PSU 12 Volts rail to
be hovering around 7.6 Volts. Haven't measured this but the figure shown
is surely a problem as all other Voltages are within spec.

Measure it from the PSU with a real voltmeter - A lot of those software monitors get the voltages way off.
If the 12v rail really was at only 7.6v the PC wouldn't even start up!


I agree with Frank, it's more likely bad capacitors and too much noise on the output causing the lockups. Swap it for a known good unit and if that solves the problem then repair or replace the PSU before it ruins the rest of your hardware.

blanco
07-11-2012, 09:00 AM
It was the internal hard drive! Measured the 12 Volt rail
and it was fine, despite the low value shown by SIW.
Has worked perfectly for 3 days since changing the drive
and nothing else.
Thanks guys.

Frank_sumbody
07-11-2012, 04:46 PM
So now it has a new hard drive, what value on the 12 Volt rail does the SIW software give?

You can in fact in the future run the power supply without the computer attached, but it is better to have a load on the supply rails such as 12 Volt car light bulbs, of course if you have a oscilloscope you can check the waveform on the supply rails (to see if the electrolytics are doing there job) , you dont have to buy a oscilloscope you might find oscilloscope software and run it on your portable laptop and attach the probe via the sound card, if you are really lucky you might even find free oscilloscope software on internet.

Of course checking the power supply out of computer saves time waiting for replacement power supply only to find the fault is still there.
Google "testing atx power supply without motherboard" to read more.

Agent_24
07-11-2012, 04:53 PM
Pure PC-based oscilloscope software is useless for testing a PSU because the sound card is limited to audio frequencies, and the noise on a PSU output is often at much higher frequencies.

blanco
07-11-2012, 09:21 PM
Frank_sombody: The SIW prog still shows 7.6 Volts but Speedfan
and others show 12.06 Volts, confirmed with digital multimeter.
The same version of SIW shows correct Voltage on my system,
so I find that strange. Everything has worked perfectly since the
hard drive was changed a few days ago and the new ATX PSU my
mate ordered from ebay is now an unused spare unit.

Agent_24
07-11-2012, 09:51 PM
Frank_sombody: The SIW prog still shows 7.6 Volts but Speedfan
and others show 12.06 Volts, confirmed with digital multimeter.
The same version of SIW shows correct Voltage on my system,
so I find that strange.

Some software just doesn't read the sensors on some boards correctly.

blanco
08-11-2012, 12:26 AM
Or possibly a faulty sensor circuit on the motherboard

Agent_24
08-11-2012, 12:36 AM
Or possibly a faulty sensor circuit on the motherboard

Doubt it, especially if the other software is correct. I've run many different hardware monitor programs on many different boards, none of which were faulty, and had varying readouts between them. There are many sensor chips and many ways to interface with them. Not every program seems to be able to account for all the possible permutations, I would blame the software first before the hardware, when you get a dodgy reading.

Quite likely if the sensors were reading the voltages wrong the thing wouldn't start up at all, as it would be fooled into thinking there was a PSU problem.

The best software to check the hardware with will be the sensor readouts in BIOS setup, if it has one. That will be programmed to read the sensors correctly, if it's correct then the problem lies with whatever software you're using.

If the BIOS readings differ greatly from readings taken by a known good voltmeter then I guess something could be up with the sensors, though I've never seen it happen before.

blanco
08-11-2012, 09:42 AM
Agent_24: there are no Voltage readings in his BIOS - I looked in
there before running ay hardware monitor software. Anyway, thanks
for your interest on the subject, I have it all running sweetly on a
recycled 250 GB drive untill his new 1TB drive arrives and I copy
everything over from the syst image.
Time to close this thread, unless there is more to say.

PENTIUM
08-11-2012, 03:42 PM
In my experience SIW sometimes gives misleading voltage reports, variations can occur, too, between mobo sensors.