Ground Fault detected by UPS

[boater bill]

I am having a ghost of a problem and could use some help.
We did a renovation of a facility that had several floating neutrals on transformers we found and bonded properly. One panel serving the UPS for the voice data system is showing an intermittant ground fault. To date we have:
1.tried a different UPS with the same result, so the UPS is working properly.
2.plugged the UPS into a different panel system and the problem goes away.
3.lift the iso ground wire off of the iso bar, and connect to the panel ground. The problem cleared for about 15 minutes then came back.
4.the mini split AC unit on the panel is wired to B and C phases, I changed the 3 circuits that we are plugging the server equipment in to, all to A phase with their own neutrals, this caused the problem to clear for about a day, then came back the next morning.
5.at first, shutting the mini split off at the breaker caused the problem to clear, but not anymore (according to the maintenence manager).
6.putting any of my meters on the line doesn't show any problems.

We are going to try to move the ground connection from the main switchboard to a ground rod in the electrcial room may clear the fault. I am thinking there are parallel neutrals somewhere causing problems that the UPS is picking up as a ground fault.

Any other ideas?

 

[Buck Parrish]

I vote for #2

 

[Pierre C Belarge]

Bill
I cannot say for sure, but this sounds like a classic example of a "high resistance ground fault".

Try to isolate the circuit affected. Once you isolate the affected circuit, check for ALL connections of the ungrounded conductor and the ground fault current path. The ground fault current path may be the EGC or the raceway/wiring method.
Remember the Effective Ground Fault Current Path includes both the ungrounded and the EGC as the ground fault circuit.
The longer the circuit is, the harder it will be to locate the high resistance issue, especially if the high resistance is created due to a number of poor connections.


I am curious what you find, please keep us updated to what you find.

 

[boater bill]

Found it.
One of the bonding jumpers in the transformer serving this and other panels wasn't torqued. Just laying in there, it would have been better if it was not connected so we could have found it sooner.

torques it down and all of the ground fault errors went away.

Thanks