Current spike when bonding separate grounds?!

[timecrystal]

Hi there,

This has totally stumped me and there must be a gap in my knowledge.

I have two completely isolated systems with different grounds. When I bond the grounds (and ONLY the grounds) between them I am seeing a quick pulse in current flow from one ground to the other.

I tested this by putting a resistor between the two grounds and running the Oscilloscope over the resistor when making the ground connection. Whenever I make the ground connection a short spike of voltage is measured over the resistor.

I understand this can happen if the two isolated grounds have a potential between them, e.g. because of static electricity buildup, but this happens on every consecutive connection between the two systems. I would imagine that static electricity buildup could discharge on the first connection but not every consecutive connection. I don't understand what would cause a current spike to flow from one ground to the other consistently like that.

Can anyone explain why this is happening? I'm completely at a loss. Thank you!

Time Crystal

[timecrystal]

Nevermind, I found the ground loop. In other words, these systems actually weren't isolated.

[Ian.M]

What is your definition of "completely isolated"?   If it isn't: 'self-contained battery powered, with no other external connections, in a sealed high quality insulating box' (e.g. a perspex, HDPE or polystyrene sheet box supported by teflon or other highly insulating rods), then something else may be going on, likely involving leakage currents, but if the circuits have high voltage sections, possibly involving ionisation of the air.

<LOL> too late, I *knew* they weren't properly isolated . . . .

[timecrystal]

<LOL> too late, I *knew* they weren't properly isolated . . . .

Hehe, Yeah. Big D'Oh moment. The sad part about it all is that it was a ground loop I had never considered throughout my design. I always assumed this other connection (that was causing the loop) was isolated as well. I'm pretty late in the design process so it's either a "back to the drawing board" or a big caveat to my users. The ground loop is somewhat justified (it's a cheap USB device), it just feels so dirty.

[Ian.M]

If USB is involved, the default assumption is there is no isolation unless you explicitly add it for *EVERY* USB device and USB derived power source involved.  Trusting any 'USB charger' PSU to be isolated and floating is extremely unwise.  Even if it isn't pure chinesum with 1mm or less creepage distance from primary to secondary, there will be a Y capacitor from primary to secondary of several or even tens of nF with approx. half the line voltage on the primary side of it.

[WattsThat]

If computers are involved with the usb side of things, you can usually count on a desktop to connect the negative usb supply to earth by way of the three wire line cordset. A laptop with an external two wire cordset will leave the usb power floating. 

[Ian.M]

'Floating', although technically accurate has misleading connotations of relatively placid motion.  The real situation is, (with up to 0.75mA of line frequency leakage current  from the non-grounded laptop PSU Y capacitor), better described as bouncing up and down at line frequency like a cigarette boat at full throttle leaping waves !