A weird AC phenomenon - do I have a ghost in my lab?

[LaryPant]

Hello everybody,

I have had this weird phenomenon in my lab (desk) for quite a while now. It only happens every so often, but I cannot seem to figure it out.

What happens is: Sometimes when I turn off my lab PSU my USB DAC cuts out, but only if I use the line output and have my speakers connected.

The long story:
I have two power strips mounted under my desk and they feed my O-scope, 2 PSUs, Signal generator, two laptop chargers, a monitor and some minor appliances. Both of them are connected to the same outlet.
I have also mounted a USB hub right next the power strips and the 230 V AC cord runs right by it - see attachment. I have a keyboard, a mouse, a USB DAC and two other cables - which go to nowhere at the moment - connected to the hub. The hub is not externally powered but plugged right in to my laptop.

The USB DAC is a Behringer UCA202 and I sometimes connect the lineout to my speaker system at the other side of them room. They are powered from their own outlet.

Sometimes when I turn off one of my PSUs - RnD Lab 320-kd3005d - and the speakers are connected the audio cuts out for 1-2 seconds and then comes back. According to the OS on my laptop, the USB DAC itself gets disconnected as well. The DAC itself is located far away from the power lines, but it is plugged in to the USB hub which is close by. The rest of the things plugged in to the USB hub are not being disconnected.

My guess would be, that turning off my PSU somehow creates a power spike or something which triggers something in the USB hub, but this seems really unlikely to me as their is no physical connection between those and the AC cable should be shielded appropriately. If this is the case, should I be worried? Could this cause problems in other devices but which just haven't shown itself yet?

I have attached a picture of how close the AC cord is to the USB hub.

I would really like an explanation of why this happens! I can live with it, but the phenomenon has really peaked my curiosity.

Thank you in advance,

Lary

PS: I do not believe in ghosts

EDIT: Fixed spelling mistakes.

[amyk]

My guess would be, that turning off my PSU somehow creates a power spike or something which triggers something in the USB hub, but this seems really unlikely to me as their is no physical connection between those and the AC cable should be shielded appropriately.

I guess it's the power switch arcing on turn-off, which creates a brief pulse of noise at high enough amplitude to get picked up by and disturb the USB. The fact that it doesn't happen all the time is indicative of that --- if you turn it off around the zero crossing, there's little to no arc produced, but if you turn it off near the peak, there's a much higher transient.

I've never seen a shielded mains cable.

[cgroen]

Is that gray frame made of metal  ??
I would NOT do that......no wonder a spark can find its way around when not switched at zero crossing.....

[Ian.M]

Check continuity of *ALL* your USB cable shields, plug shell to plug shell.  Many don't meet the USB 2.0 spec for  shield resistance (under 0.6 ohms), and cheap cables may not even have the shield connected!

[2N3055]

You could have grounding problem in your sockets or power strips.. Try taking long cable and connect everything (including speakers) to same power strip and see if problem goes away..

[LaryPant]

Thanks to everyone who replied!

I guess it's the power switch arcing on turn-off, which creates a brief pulse of noise at high enough amplitude to get picked up by and disturb the USB. The fact that it doesn't happen all the time is indicative of that --- if you turn it off around the zero crossing, there's little to no arc produced, but if you turn it off near the peak, there's a much higher transient.

That sounds quite plausible, I am going to have a look at arcing.

I've never seen a shielded mains cable.

Ah, I didn't mean shielded as such, but isolated. I was under the impression that these cables should be isolated good enough to prevent something like this from happening.

Is that gray frame made of metal  ??
I would NOT do that......no wonder a spark can find its way around when not switched at zero crossing.....

That is a good question, it is from Ikea.
I just checked, the outside of it is not conductive, but that might just be a coating. The inside is most likely a metal or alloy of some kind.

Check continuity of *ALL* your USB cable shields, plug shell to plug shell.  Many don't meet the USB 2.0 spec for  shield resistance (under 0.6 ohms), and cheap cables may not even have the shield connected!

Thanks, I will check.

You could have grounding problem in your sockets or power strips.. Try taking long cable and connect everything (including speakers) to same power strip and see if problem goes away..

Yeah, I most likely have grounding issues as well. I am unfortunately not using three pin power strips, but why would a longer cable help me out?

Thanks again for the replies. The consensus appears to be that the problem stems from a high noise pulse being generated when I switch off the supply which is able to arc over to the USB hub it self.
Possibly due to poor grounding and/or the metal frame that holds the power strips and USB hub.

If I understand it correctly, the solution would be to get rid of the metal frame and increase the distance to reduce arcing and better the grounding?

And more importantly: If I didn't change anything, could this damage my other equipment over time or is the effect negligible?

[jonpaul]

Bonjour a tous:

I think the PSU mentioned is a lab supply with a linear bulk AC><DC, transformer/rectifier/filter cap.
The iron core 50/60 Hz transformer has inrush current and core  stores energy depending on the point of AC cycle at turn-off.

There is no EMI or transient filter on the  PSU, so the transients cause a surge into the mains.

No consumer power strip, "transient protector" or filter will be effective, and some energy is radiated as well as conducted.

The storage of a several kG iron core transformer can be many joules.


Finally, Uri Behrringer is a fine audio designer, but all low cost consumer products are made cheaply in China.

Thus any non-pro  audio products will be susceptible to transients conducted or radiated as they do not meet the high level EMI and transient regulations.

Finally no harm should come from repeated resets, as the MPU and digital line are received the interference but are somewhat internally protected.

Just the ramblings of an old retired EE (1968)


Bon Chance,

Jon

[2N3055]

You could have grounding problem in your sockets or power strips.. Try taking long cable and connect everything (including speakers) to same power strip and see if problem goes away..

Yeah, I most likely have grounding issues as well. I am unfortunately not using three pin power strips, but why would a longer cable help me out?

Thanks again for the replies. The consensus appears to be that the problem stems from a high noise pulse being generated when I switch off the supply which is able to arc over to the USB hub it self.
Possibly due to poor grounding and/or the metal frame that holds the power strips and USB hub.

If I understand it correctly, the solution would be to get rid of the metal frame and increase the distance to reduce arcing and better the grounding?

And more importantly: If I didn't change anything, could this damage my other equipment over time or is the effect negligible?

First you misunderstood arcing. Arcing happens inside switch on your PSU, and that creates lots of EMI energy..
There won't be arc between two pieces of equipment, you don't have 100000V to jump. Pulse from self-induction of trafo would be seen on mains voltage as voltage spikes and that could confuse something connected to it.

Do you have anything grounded? It is not about longer cable, but having all equipment connected to same electric circuit and common grounding point. 
If you don't have grounded sockets, equipment with switching PSU can inject small currents into grounding connection and inject voltage into grounding pin, and everything connected to it.

As I said, find properly grounded socket in your wall and connect everything into one proper socket strip with proper grounding. Then test. There should be no potential between any grounds on any equipment connected together, their grounding (reference points) must be equipotential..

[Syntax Error]

If from your picture, the metal frame is not touching the PCB and parts, causing random activation, you can only conclude that you have a 5G ghost in your machine. It goes something like this...