-1

We have a workstation machine which, for the past three days, shuts down (apparently) randomly. After it shuts down, if you press the power button, the machine won't start. However, if you pull out the power plug from the wall socket end, again put it back in, and then power-on the machine, it would start. Nothing else, but this action, would make it start.

The machine is connected to the wall power socket, through a Belkin spike protector strip. The wall power socket is connected to a common 10 kVa UPS. Other machines running in similar setups are working fine.

I have changed the machine's power cable and the spike protector strip. Same situation continues. The machine runs Windows XP SP3. The machine is not connected to the internet and is used to run a Windows application over the LAN (which is not connected to the Internet too.)

Other specs of the machine :

Intel Core 2 Duo E7500, 2.93GHz, ST3160310CS 150GB HDD, DDR2, 2GB RAM, Kingston Intel DG31PR Motherboard

Before I open the machine and check with RAM or SMPS, I wish to know what could possibly be the issue and how it is caused ?

This is not the first time I am facing this issue, though I don't remember the cause or solution from years back. :-)

Other possibly related context :

The monitor connected to the machine has these wavy forms possibly related to power. The wavy pattern is somewhat noticeable, but not hampering work by the tolerant user :-) . Another monitor connected to the same machine also had the same issue. Changing refresh rates didn't have any positive change in the issue.

Whirl Mind
  • 111
  • 7

2 Answers2

3

It's good that you mention that your (EOL since 2014) Windows XP is not connected to the Internet. However, that's not the only thing that's extremely old here. The discontinued Intel DG31PR is from Q3'07 and Core 2 Duo E7500 from Q1'09, so I'd say we are talking about at least 8yo workstation here. Do not expect it to work forever like it was new. It should have been replaced already.

Esa Jokinen
  • 46,944
  • 3
  • 83
  • 129
3

To me, those symptoms suggests that the PSU in the machine is failing: as soon as its internal control circuits detect the PSU is failing to keep the output voltages in the required ranges, they will trigger an immediate PSU shutdown to prevent damaging the rest of the hardware or letting the system run with a serious undervoltage, which might cause parts of any processed data to become silently corrupted.

The usual way to reset those control circuits within the PSU itself is to disconnect incoming power to the PSU, so that explains why pulling the power allows the system to recover.

As Michael Hampton commented, unfortunately, faults like this are often caused by aging electrolytic capacitors within the PSU, and so the fault is likely to get worse and worse over time. The motherboard of the workstation is also likely to contain some electrolytic capacitors, and those can also cause problems as they age.

telcoM
  • 4,448
  • 15
  • 25
  • We replaced the PSU with a new one, it is working fine now for a couple of days. It doesn't seem things have gotten as serious as the motherboard yet. Thanks. – Whirl Mind May 20 '18 at 06:57