User Tools

Site Tools

Action disabled: source

wiki:hardware:troubleshooting:psu

When troubleshooting gets creative

Old blog post

This article was originally published on October 29, 2008

Why this works

By unplugging the machine in question and waiting ten minutes, the capacitors in the machine's power supply have enough time to drain any residual charge, which may be interfering with the normal function of the machine's components (both the motherboard and other internal components).

Pressing the power button before reconnecting the power cable can, depending on the hardware, discharge the power supply's capacitors, but this isn't a given.

I tend to be the go-to person for tech support among my friends, family, and neighbors. A few nights ago, one of my neighbors asked me to help him with his custom-built machine. His USB ports had mysteriously quit working. I messed around, and the same thing kept happening. His Microsoft USB wireless mouse worked just fine, but his HP all-in-one inkjet printer and his SanDisk thumb drive didn’t work. Plugging them in resulted in a “one of the devices connected has malfunctioned” error from Windows. His thumb drive appeared to work just fine in Ubuntu Linux.

We removed all the USB host items from the Device Manager and reinstalled them, to no avail. Then I tried disabling the USB enhanced controller item (this is generally what lets Windows know that a USB controller is capable of USB 2.0). This worked, but kept his devices from working at USB 2.0 speeds. We even reinstalling Windows completely and still had the same problem.

I started Googling around and found this article1) regarding my exact problem – the SiS PCI to USB Enhanced Host Controller mysteriously quit working, and took all USB 2.0 devices down with it. This blogger’s solution was to shutdown and unplug the computer, wait for about half an hour, hold in the power button for a minute, and then restart the machine.

And, like magic…it worked.

I can only begin to guess why. Like this other blogger, my neighbor’s machine had a Gigabyte-brand motherboard. My theory is that something was getting corrupted or otherwise mistranslated by the motherboard regarding the USB 2.0 controller, and whatever was breaking things was being remembered across shutdowns and reboots.

By letting the machine sit for awhile, all the capacitors on the motherboard and power supply emptied out, thereby completely cutting power to the machine. This apparently reset whatever dark magic was causing the problem in the first place.

It’s likely that in Matt’s case, Linux was allowing the ports to work, but only at 1.1 speeds. Since we didn’t test how quickly file transfers took place in Linux, I can’t confirm this theory, but it would fit with my above assumptions.

At any rate, now we have our answer. The next time this happens (and I imagine it will), I’ll try the fixes listed on usbman.com first, so we can try to cure the problem once and for all.