4
\$\begingroup\$

I have a workstation that's becoming a little long in the tooth, as it were. It's a rock solid system, but about a year ago it had a faulty power supply that I replaced. During the past 11 or so months, it's run just fine, but recently it again had a "hiccup" where it just died (no panic) and then rebooted. I suspect the power supply again, but I just don't know. (Unfortunately the power supplies are not cheap for this server.)

It got me to thinking: Is there a way to test power supply input and output to find out if and when a power supply fails or is failing?

It seems as though there should be a way to measure both of these in/out lines (realizing that DC out on the inside is a little more involved) to monitor such issues...

\$\endgroup\$
  • \$\begingroup\$ How many supply rails? Most modern computers use +12, +5, +3.3 Vdc rails. There is also a "power good" signal that you should monitor. Please update your question with the requested information and we can give you guidance. \$\endgroup\$ – Dwayne Reid Nov 8 '18 at 17:12
  • \$\begingroup\$ Find some software to log the SM (System Management Bus) values (+12, +5, +3.3 voltages, etc.) Analyze the log over the course of a few days and especially leading up to an event. Could be the supply, or even bad electrolytics on the motherboard. \$\endgroup\$ – rdtsc Nov 8 '18 at 18:01
2
\$\begingroup\$

Is there a way to test power supply input and output to find out if and when a power supply fails or is failing?

There is a way if you had an oscilloscope, you could monitor the voltage rails and the PG signal.

In my mind there are two problems that could happen:

1) The power supply has some kind of fault and drops out 2) A component in the system shorts out the supply and the voltage drops, then the supply turns off momentarily

You may be able to monitor the rails with a scope and trigger if the voltage falls below the recommended value. If the voltage falls and you know what rail its on this could help you determine if its the supply or a component. I would think that if it's a component the rail would fall below its recommended tolerance, then you would see the PG signal fall. Knowing which rail may also help, It is my understanding that most supplies have separate regulators for each voltage rail, so a hard drive or fan's voltage rail (like +5V) would most likely be separate from the motherboard

enter image description here Source: https://www.lifewire.com/power-supply-voltage-tolerances-2624583

The ATX specification defines the Power-Good signal as a +5-volt (V) signal generated in the power supply when it has passed its internal self-tests and the outputs have stabilized. This normally takes between 0.1 and 0.5 seconds after the power supply is switched on. The signal is then sent to the motherboard, where it is received by the processor timer chip that controls the reset line to the processor.

The ATX specification requires that the power-good signal ("PWR_OK") go high no sooner than 100 ms after the power rails have stabilized, and remain high for 16 ms after loss of AC power, and fall (to less than 0.4 V) at least 1 ms before the power rails fall out of specification (to 95% of their nominal value).

Source: https://en.wikipedia.org/wiki/Power_good_signal

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.