Question Weird power issues, possible PSU failure

[CraigParton]

Hey guys, I'm needing a bit of advice. I've been having this odd issue after moving. When my washing machine kicks on, my USB hub will power off for a second. I use the hub for my wired keyboard and mouse, so they both power off as well. It's like you just unplugged and plugged in the connector. That's very annoying, especially when gaming. I know washing machines are usually powered by their own breaker, but it seems this one is on the same circuit as my PC. At first I thought this was the issue, however I also noticed that if I turn on the ceiling fan in that room it does the same thing. A fan doesn't pull anywhere near the amperage a washing machine would. So now I'm thinking this has more to do with the electric motor startup.

But this is what really gets me. The USB devices work off the 5V supply. If the 5V supply momentarily dropped enough to kill my hub, wouldn't I notice more issues than that? Wouldn't that power off the logic for the HDDs and other devices on the motherboard? But nothing else quits working, I just have to wait a couple of seconds for my keyboard and mouse to reconnect and I'm good to go. This particular hub has no external power, USB only, and it also has a built-in SD card reader and a blue LED backlight that flickers off when this happens. The only things I have plugged into it are my keyboard and mouse. They keyboard is a Razer Chroma, so it may use a bit more power than a standard keyboard, but nothing crazy.

So now I'm thinking my PSU may be failing. Just before I made this post my PC powered off without warning, then restarted on it's own. That's the first time that's ever happened. I'm thinking the USB issue might be due to the slight voltage dip when the electric motors on my ceiling fan and washer are starting up. Power supplies are supposed to handle those with no issues, but if mine is failing perhaps it can't. Does that sound far fetched? Or do I have deeper issues with my home wiring? Any advice would be greatly appreciated!

 

[LTC8K6]

Maybe it is the surge from arcing switches or relays?

Is your computer plugged into a UPS or at least a surge protector?

You may also have a grounding issue with your outlets.

 

[LADave]

I just replaced a PSU yesterday, after a couple BSODs reporting memory faults while waking up from "sleep", plus a reboot while unplugging a USB external drive with its own power supply. No more problems since the swap. This is on a Dell Optiplex 780 mini tower. The replacement PSU came from an Optiplex 620. Same wattage, same connectors.

I have a no-name PSU tester that the various connectors plug into, lighting green LEDs to indicate proper voltages. It gave no indication of anything amiss. That wasn't a useful way to identify a PSU that was intermittently flakey under conditions the tester didn't simulate.Testing with a DVM has more or less the same shortcoming, especially if anomalies are of short duration, and I just don't have enough hands to run three DVMs, back-probing three different pins, while disconnecting a USB.

I downloaded and tried HWMonitor and Open Hardware Monitor, but neither one showed voltages on the Opti 780 or the 380 I tried, nor on an HP 8000 Elite. Apparently none of these PCs' motherboards have voltage sensors for software to access. However since all important voltages (3.3, 5, 12) are on the PCI bus, I'm surprised I couldn't google up a PCI card that collects data on these voltages. I seem to recall such a card for the old ISA bus, circa the 1980s.

P.S. all three voltages are nominally present on SATA power connectors, which suggests a more universal monitoring point, that should also work on laptops.

 

[CraigParton]

That's good info, thank you for your reply! I have a four channel DSO that I use at my shop (I'm an auto tech that specializes in wiring issues). Seems easy enough to fabricate a male plug for the SATA connector to monitor the voltages. I should do that and start the washer to set what happens. I could also monitor the AC plug and see what happens to that voltage.

Another interesting one to watch would be the POWER_GOOD signal from the supply. I think that may be what caused the shutdown. I'll try to test it later this week and post the results.

 

[LADave]

Unfortunately my 'scope is a very basic one-channel Heathkit, however there's a $15 Arduino board with no fewer than six ADC channels, with USB for power and communication with a mother ship. It seems very possible to breadboard up an interface with voltage dividers and diode protection, and to write software or modify existing. Sampling rate would be somewhat limited, however I suspect all the caps in a PSU buffer against a huge dV/dT.