Ground fault indicator gets bright during fault

[CFTech]

Hello, first post. Long story short, ground fault lights are referenced to ground and normally go dim or go dark if ground fault occurs. I have one now that gets brighter when 20 horse cooling tower motor is turned on. Phase to phase arcing within the motor?

3 phase ungrounded Y

Thanks,
Don S.

 

[davidbeach]

One light getting brighter would mean a higher voltage on that phase; a higher voltage that would have to have a corresponding dip on the other two phases. Could be some sort of phase-phase fault in the motor; could be that you're not seeing exactly what's happening with the lights (the human eye doesn't do a good job of recognizing small differences in the amount of light). In either case, it would be a good idea get some recording meters in use and get a much better idea of what is happening.

 

[rbulsara]

What happens (or happened) to the motor or its controls?

 

[LionelHutz]

If you have a light from each phase to ground then if one phase is grounded it's voltage will be 0V with respect to ground while the other two's voltage will be 1.7 times higher wrt ground compared to a system with balanced voltages.

With an ungrounded system you typically see phase to neutral voltage between each phase and ground. but. If one phase is grounded then you get phase-phase voltage when measuring between the other two phases and ground (ground is now the same as one of the phases).

 

[rbulsara]

How did you reference the light to ground on an ungrounded system?

 

[waross]

Considering the resistance of the light bulbs, if this is a typical old style ground detector setup with no grounding other than the grounding lights, then you have high resistance leakage to ground.
Consider the normal current through the lamps and the capacitive charging current of your system. Now consider how much this current must change to cause the lamps to brighten or dim noticably. This change in current is about the level of your leakage current. eg; Very low.
With one lamp going bright, the problem is on the other two phases. Look for single phase loads such as control transformers with damp windings, or one damp winding in a motor with delta connected windings. To megger you may have to disconnect the motor leads and check each winding individually.
yours.

 

[CFTech]

We generate our own power here at a natural gas compressor station. 3-312 KVA generators, old Ingersal PVG-8 prime movers. The switch gear and MCC is all new (1999). The 480 ground detection is pretty simple in that it has a 120 volt light per phase isolated by it's own transformer referenced to the ground buss.

I haven't checked in to it yet since I am working a week of midnight's this week. When I do dive in I'll remove the motor from the circuit and Megger both ways.

Thanks all for the fuel for thought.

Regards, Don