Ground Fault mis-operation 1

[living2learn]

I installed a main circuit breaker 1600A on the secondary of an existing 1000kVA 12.47kV delta:480Y277V with ground fault protection. I received a call stating the circuit breaker failed the ground fault test because it did not pass the no trip test. It was set at 320A and was tripping at 160A. The testing firm is reputable and the breaker is brand new. No testing reports were supplied with the new breaker unfortunately. I asked if the impedance across the neutral disconnect link was measured and was told yes. Unfortunately I didn't ask what it was because I looked at the test reports and found out it was .1 ohm's with the link removed.
The contractor has ordered a new circuit breaker and has concluded this one is broken. I believe we have parallel paths because the neutral is obviously bonded more than once which the circuit breaker would see the current twice.

Is my conclusion correct that if the neutral was bonded more than once we would have parallel paths and the current would be seen twice?

Thoughts?

 

[rcw retired EE]

With a setting of = 320 Amps and tripping at = 160 Amps during the "No Trip" test I would suspect the neutral CT is wired backwards since trip = 50% of setting. Or it could be the test set up had the neutral current going through the neutral CT inthe wrong direction.

The no trip test circulates current through two phases of the breaker or through one phase and the neutral. Circulating current of the proper polarity (In "A", Out "B" or In "A" Out "N") should not trip at high currents. It simulates a large load.

If the neutral CT is wired backwards or the test current polarity is wrong, the signals add instead of cancel and cause a trip at 50% of setpoint.

If the test current was run through the breaker, and not the neutral, an internal phase CT may be backwards.

 

[living2learn]

Thanks! They did two different tests. The fist was when phase C (IN) was connected to N (OUT) and it tripped at 50%. They took the neutral CT out of play and utilized the residual method inside the breaker and ran 160A's (50%) through it (C (IN) was connected to N (OUT)) and it still tripped - even without the neutral CT in the equation!

So my thoughts about a multi-bonded system are not accurate?

 

[rcw retired EE]

In my experience, main breaker ground fault tests inject primary current from a test set through the breaker and the neutral CT. The injection test works even if the neutral has multiple grounds. A separate part of the GF test checks for grounds on the load side of the neutral CT by measuring neutral bus to ground impedance with the N-G bonding jumper open.

Test set output is low voltage, high current and is usually ungrounded and isolated. Test set current only flows between the two test lead connections. There is no parallel path for current flow with the isolated test set output. If the test connections did have a parallel path, the tripping point would be higher than setpoint due to the shunting of the parallel path.


Your results still don't make sense unless there is a CT or rating plug error. Pushing current through just one breaker phase should cause a GF trip right at setpoint, not 50%.

 

[living2learn]

"A separate part of the GF test checks for grounds on the load side of the neutral CT by measuring neutral bus to ground impedance with the N-G bonding jumper open." They did that and it came back as .1 ohm's.

I agree completly with all your comments! Thanks again - I have been scratching my head pretty hard on this one and all I could come up was multiple N-G bonds recirculating.

 

[living2learn]

I found a generator that was running during the test that had its neutral connected to ground with a 3 pole switch that was serving loads during the time of the test.