My PC makes the USB connect/disconnect sound when I turn on the light???

[Bicurico]

Hi,

My main computer is doing somethinig weird and only now have I understood when it happens. It was doing the USB connect/disconnect sound many times over the day, without me doing anything, much less attaching or removing any USB device.

So I just found out ot happens when I turn on any light in the room (desk light, ceiling light, etc. In fact it happens when I switch on ANY consumer. And it is not a faulty outlet or 6-in-1 plug (don't know the english name). Even the ceiling light will cause it.

I suspect it is the PSU of the computer that is not filtering the ripples correctly any more, perhaps due to dried capacitors. Would that make sense?

Other computers which I operate next to it don't have these symptoms.

There is no apparent device being disconnected, either.

Opinions?

Regards,
Vitor

[Bicurico]

Update: I am pretty sure the problem was (?) a faulty USB hub with subsequently attached SIM card reader. I removed the SIM card reader and the problem seems solved. Of course I will need to replace the USB hub, but right now I still need it. Anyway it seems that the USB hub is/was at the verge of power consumption (card reader + 3DConnexion SpacePilot) and any spike in my houses power grid caused the card reader to disconnect. If I connect this device directly to a PC USB port, it behaves normally.

All a bit strange...

[Ian.M]

Throughout the history of the USB standards, USB leads for faster than Low Speed are required to be shielded and the shield is required to be connected to each plug shell

Unfortunately most cheap USB leads don't have the shielding (if even present) connected to the plug shells, which significantly increases their susceptibility to EMI. 

They also often use too low a wire gauge for the power wires, so result in excessive voltage drop under load.

Therefore before blaming your PSU, or hub check all detachable USB leads for end-to-end shield resistance, and if possible, with the device connected and powered,  check Vbus with respect to local ground at both ends of the cable, or set up a jig to measure loop resistance of the cable's power wires (The jig is a good quality type A socket with the power pins shorted and a breakout board for whatever 'flavour' of B socket the cable needs to access the Vbus and 0V pins)  to see if the cable is introducing excessive voltage drop.

[Bicurico]

At the end it was a small USB GPS receiver!

I opeend it and it had a tiny battery cell (about 5mm diameter), which had leaked and was oxidized.

I think this battery acts as a backup for the GPS data for fast reconnection. I removed it but still the device would act erractic, probably consuming too much power or whatever.

Since I removed it, my computer is no longer acting weird on any consumer being switched on.

Interestingly the GPS receiver is apparently working when connected to an unpowered USB hub, but won't be recognized if connected directly to the USB port of the computer.

I ordered a replacement GPS receiver (they cost around 8 Euro) and this one gets tossed away (not really, it just gets to the parts bin).

Regards,
Vitor

PS: I don't have any "USB test jig" and at this point I am not investing in one...

[SiliconWizard]

That usually comes from a faulty - or just badly designed - USB peripheral. It can happen with a peripheral that is simply too sensitive to EMI.
For instance, I have a (now) relatively old Brother laser printer (which still works fine) connected via USB, that exhibits a similar behavior. Sometimes when plugging in some device in the same plug rail as the printer, it will trigger an usb re-enumeration by itself.