Intermittent loss-of-phase

[eag]

Background - 3-phase, 2000A service fed from 3 single-phase pole mounted transformers. Switchgear is NEMA-3R, manufacturer unknown (may be GE), with push-to-trip, handle-actuated draw-in mechanism, with 1600A fuses.
Problem - There is loss-of-phase occuring intermittently (anywhere from 3-7 days apart recently) on the C phase. Analysis so far shows that voltage downstream from the main is normal up to to point of phase loss, then upon loss-of-phase falls to between 0-10V for anywhere between 30 minutes to an hour before returning back to normal voltage. What could cause this sort of intermittent loss-of-phase in the system described?

 

[augie47]

Since you have an analysis, I assume you are measuring the voltage. My first thought would be to measure on the line side of your disconnect in case it's a POCO problem..quite possible, Secondly I would recommend an IR scan. A lost phase would indicate a loose connection which should show on the scan.

 

[masterinbama]

Is this an open Delta set up? Are you sure that your loss of phase is not actually an unintentional grounding of a phase?

 

[augie47]

good point !! a far greater probability.

 

[masterinbama]

Where in Huntsville is there a service that size fed by pole mounts?

 

[mivey]

Where in Huntsville is there a service that size fed by pole mounts?

That is big but not unheard of. Probably can see these in an old industrial section or old urban areas. We usually stop at 250 kVA units (this would give you 2kA at 208V) for OH and the last set I recall had them mounted on a platform between poles. At any rate, Howard makes above 250 kVA including 333, 500, and 833 kVA single-phase units.

 

[masterinbama]

That is big but not unheard of. Probably can see these in an old industrial section or old urban areas. We usually stop at 250 kVA units (this would give you 2kA at 208V) for OH and the last set I recall had them mounted on a platform between poles. At any rate, Howard makes above 250 kVA including 333, 500, and 833 kVA single-phase units.

I am intimate with the entire area of Huntsville Alabama having worked almost all of my 32 years in the trade in the area.

I cannot think of any building left in this area with a service that size fed by pole mounted transformers.

 

[mivey]

I am intimate with the entire area of Huntsville Alabama having worked almost all of my 32 years in the trade in the area.

I cannot think of any building left in this area with a service that size fed by pole mounted transformers.

In all fairness to your observations, the service may be 2000 A but the pots are probably much smaller, like a set of 167s or such so you might not have noticed the disconnect.

 

[eag]

Since no button to edit my post, here is an update: (This problem is in Athens, GA) - augie47 hit on part of my first inclination for troubleshooting. I proposed to the maintenance contractor that they put a PQ analyzer on the line and load side of the main switch to measure voltage and narrow down to POCO or customer side when the outage occurred again. Of course, IR was another good thought and would most likely confirm my main suspicion. Beyond that suggestion to start, because of the nature of the outages, my main suspicion was the knife switch. Knowing that the door of the 3R enclosure was rusted and hanging loose for quite some time before it was repaired, I postulated the gear was exposed to excessive moisture inudation, thus probable corrosion. So in my mind, if the switch had been operated in the past, it could be that the suspect phase may not have fully engaged in its slot upon draw-in. If the knife was just resting at the slot entrance, this would most likely cause a high-resistance (loose) connection. The metals could heat up to the point that their expansion could result in either mostly or completely a loss of connection. Then, as the metals cooled down, re-establish connection until heating and expansion of the loose connection again caused all or part phase loss. At last count, the contractor was able to determine upon tripping the switch that the A and B phases did not disengage, then could see with a cover above the switch removed that the C phase was indeed not fully engaged in the slot. The switch is in the process of being replaced.

 

[Jraef]

Sounds as thought you have a draw-out main? Zog is a good resource on this sort of stuff, hopefully he checks in. But you said you are not sure of the mfr of the gear. A couple of them have had issues with their "tulip" connectors, the set of spring loaded copper tabs on the switch assembly that grab onto a post on the gear assembly to make electrical contact.


Tulip assembly.

One of them was Siemens on the initial release of their WL units, I believe ABB or GE may have been another one that had problems with their tulips.

 

[G._S._Ohm]

(anywhere from 3-7 days apart recently)

To be 95% certain that a fix worked you'll need 9 to 21 consecutive trouble free days.