"Isolated" variable power supply killing my devices with capacitive coupling?

[jamesglanville]

I have a maplin power supply (http://www.maplin.co.uk/p/100w-slim-bench-power-supply-n93cx) which I have now realised has killed several of my devices, including a few AVR UARTs, a papilio voltage regulator, and maybe more.

The problem is that the negative output is "kind of" isolated - there is significant capacitance to earth. I noticed tiny sparks when connecting it to earth-referenced devices (eg to a FTDI cable connected to USB ground). I have now "fixed" it by connecting the output negative lead to earth through a 10k resistor (there is a grounding terminal on the front panel). The capacitance is significant enough to damage devices, and when I connect a voltmeter between earth and the negative output I have measured >150V! This does not quickly drop through my multimeter input impedance, which leads me to think the capacitance is far too high. With the 10k resistor the output rests at ground potential.

My question is, which of these if any is the "answer"?:

Crappy PSU with high capacitance to ground.

The ground point on the front of the unit is supposed to be used for this purpose.

This is normal with isolated PSUs

[Fsck]

A) Crappy PSU with high capacitance to ground.

is the closest choice, however there are other environment-related possibilities.
and yes, the ground terminal is there for such problems, though more for the fact that floating power supplies float from earth.
ie: common mode current noise, agilent recommends a 100kR resistor and a 2200pF capacitor in parallel between negative and chassis ground for an E3631.

[T3sl4co1l]

A linear supply need not have much capacitance, but a switcher is guaranteed to have several nF as part of EMI filtering.  Beyond that... who knows?  I've seen antique equipment with 0.1uF (wax paper -- eww!) from chassis to ground... or even neutral.

Tim

[Kjelt]

Are you sure that is the cause?
I killed a few components with these cheap SMPS because when they switch on they sometimes output 3 times the voltage it is set at. So if you for instance use it to power a 5V digital circuit and you switch it on, for a few ms sometimes voltages up to 15V are directly output from these SMPS's thereby destroying the circuit. You can only use these for directly powering your circuit if the circuit it self has a voltage regulator on board. Just my experience.

[rr100]

I've had the same trouble precisely against the same: "PC" ground (I assume "USB ground" comes from a PC). If this is the case then the PC might also be at fault (well the SMPS is guilty for not being disconnected enough and the PC for supplying the "frying shock").

Just to be on the safe side I made a point of always using a cheap netbook on battery (also not connected to ethernet or anything else, therefore as floating as it gets) when interfacing with "non-consumer" electronics for tests or programming.

[Kjelt]

If the USB is a problem get an USB isolator, for instance from Olimex. We had some problems with PoE icw USB as well, white smoke 

[rr100]

Interesting, I wanted to say that is more expensive than both my netbook and the usual devices I work on but it is actually decently priced. No word on protection for actual potential between the usb pins (they only mention short circuit protection) but still will cover 99.99% of the scenarios. On the other hand so will a battery powered laptop.