USB power supply sparks USB shield/ground to mains earth via y-cap

[megaoetzwuchst]

Hello all!
I found that two of my power supplies (the above Anker PowerPort +1) and another Anker 24W dual-port emitted sparks when I touched grounded equipment with the micro USB plug outside shield (at the position that the shield would touch the grounded item). I measured the voltage from USB (-shield/ground) to mains earth with an oscilloscope - which showed about mains AC voltage - although I could not really draw any significant current.

I was quite surprised as these are generally high quality supplies - but other power adapters I have do not have the same issue.

Finally I got round to disassembling one of these power supplies (the PowerPort +1), and find that when I remove the capacitor that bridges the primary rectified (-) side to secondary (5V USB -) , the sparking does not occur anymore. When testing the capacitor (see attachment) It shows a capacity of ~900pf - would that be within reason?. It measures resistance > 20MOhm so seems intact.





What other reasons could there be for the sparking / the capacitor conducting so much HF current? Is this expected or could it be that both of my USB devices are faulty? If so - what would be a reasonable capacitor to get (I probably won't use this one again anyway as they don't go apart easily).

Thanks for any input - would be interesting to get to the cause of this!
Best, moewu

[T3sl4co1l]

It's doing what it's supposed to...

It's legal as long as the AC mains current (through the capacitor) is under 3.5mA or so (depending on standard followed).

If you don't want sparking, find a grounded one (and don't complain that it costs more!).

Tim

[megaoetzwuchst]

Ah yes thanks a lot! 
I was thinking this would be somehow in the design - I guess they use a larger cap than other makes. Makes a lot more sense now!
I was actually quite happy having a floating power supply for measuring ... I guess I could use a non-switching supply - there must be other floating options that don't have the issue?

Best,
Daniel

[Jeroen3]

You've just removed the Y-capacitor. It's there to prevent high frequency switching noise from the switchmode converter on the secundary side, and prevent fast dv/dt on the transformer.
By removing it you now have high frequency noise on the secondary side, instead of 50 Hz hum.

Please put it back.

If you do not like the <1mA leakage current, ground the appliance.

[amyk]

It's doing what it's supposed to...

It's legal as long as the AC mains current (through the capacitor) is under 3.5mA or so (depending on standard followed).

If you don't want sparking, find a grounded one (and don't complain that it costs more!).

Tim

UL60950 is the relevant standard and 3.5mA is the allowed earth leakage for grounded equipment... not the touch current for exposed parts, which is limited to 250uA maximum.

3.5mA is going to be painful for most people, and even deadly in the right circumstances. bigclivedotcom has a video where he deliberately takes 1mA through a finger and even for him it's not pleasant at all.

[megaoetzwuchst]

Please put it back.

No worries I was not planning to operate it without the cap

[Jeroen3]

It can get worse if you have many of these appliances on a 3-prong power bar that isn't on a grounded socket.
Unfortunately this is possible in europe.