CT off by 26% 4

[stevenal]

Tripped on bus differential for a through fault last week. When troubleshooting, we found one of the 1200:5 CTs was actually 1200:3.7. That explains the trip, but what would cause such inaccuracy? Demagnetizing did not help. All taps on the multi-ratio CT had the same error. Has anyone seen this?

 

[dpc]

Was this on a phase that saw fault current?

 

[Skogsgurra]

This is getting very interesting. Exciting, I would say. No pun meant.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[scottf]

Stevenal-

The factory actually thinks magnetization is causing a 26% ratio error? Wow.

Also, if it is a protection core, I'm not sure taking the secondary up to 10A is going to work to demag it.

Perhaps they mean take the excitation current up to 10A? Do they have you apply voltage across the secondary with a variac and then raising it until there is 10A flowing? That sounds a bit on the high side. Also, once you hit the level you want, make sure you lower it back down slowly.

 

[stevenal]

Scottf,

Yes, excitation current. Tried the variac method before. Once you get above the knee, a slight change in voltage by the variac causes a huge change in current and it is easy to trip the breaker on the test set. Our new unit is automatic, set the parameters and start it. But it maxes out at 5 amps. This amount is well into the saturation area and should be sufficient. Don't know why they want ten. Five looks almost like ten anyway on a log log scale.

dpc,

Yes, this CT saw fault current from the line to ground through fault. Relay records indicate no other through faults involving this phase. Other identical CTs on this phase test okay.

Skogsgurra,

We're a sick bunch, aren't we? Those that find this stuff exciting.

All,

Will a turn to turn short show up in an excitation test by lowering the knee voltage detectably?

 

[scottf]

Yes...knee point voltage will lower and Iexc at knee-point will increase.

As for the test method...I was surprised by 10A, because 2-3A Iexc is deep saturation (probably) and is plenty to demag core.

 

[slavag]

Stevenal.
What is a ratio ( multi-ratio). of this CT.
How much cores?
Are you check also metering core?
Regards.
Slava

 

[stevenal]

Slavig,

The taps are ANSI standard for a 1200:5.
X1-X5 1200:5
X2-X5 1000:5
X3-X5 900:5
X4-X5 400:5

The other possible ratios are derivable from the above. We are using the full winding for the differential protection.

An identical CT sits above this one on the same bushing, that is used for metering. It tests fine.

 

[stevenal]

The tests were very interesting. Both the CTs on the bushing had no knees. All was linear up to 5 A. On a hunch we reran the test on one CT with the other's 2ndary open circuited. Now we reached the 500V voltage limit without going above 2A on the C400 CTs. They seem to be sharing a magnetic circuit. I'm beginning to believe the CTs were damaged from a known lightning incident a few years back. Following the strike, the opposite pole from the one in question faulted to ground, and had to be replaced.

 

[Skogsgurra]

Say that again. Sharing..?

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[stevenal]

Okay, sharing. Instead of two individual two winding transformers with their primary windings in series, we seem to have a three winding transformer where the state of one secondary affects the other. Acts like they are sharing a common core. We can dump 5 amps into the 6Y secondary without developing the expected voltage, or getting the nice hum from a normal excitation test. With the primary open circuited, the only available path is through the shorted 6X secondary under this hypothesis. Open circuit the 6X, and the excitation test on 6Y looks and sounds normal.

Or maybe they are sharing the secondary circuit, as in a short circuit or crossed wires between the CTs and the the terminal stips in the breaker cabinet. No continuity foud yesterday with an ohmmeter, though. Forgot the megger.

Going for a visual inspection today, as well as further testing.

 

[slavag]

Stevenal.
Why you don't replace this CT?
Regards.
Slava

 

[cranky108]

Several years ago we had a simular problem wth the outer most CT on a GE gas breaker. The explanation we recieved from GE was the shroud was effecting the magnitics. There answer was to use plasitc, or something like that, bolts to hold the shroud over the CT's.

Hope this helps.

 

[slavag]

I would like see reaction of Scottf to this recomendation of GE.
Slava

 

[scottf]

Hmm...sounds like the shroud was somehow making a completed loop around the CTs. Obviously, the breaker design would have measures taken to prevent that.

Something like that would also explain some of the issues in this case.

Also, could be some secondary wiring errors, i.e. crossed wires between the 2 windings.

Another, more remote scenario, would be an insulation failure between adjacent CTs, where the outer most layer of one becomes shorted to the other.

I highly doubt that the 2 secondaries are wound around a common core. Would be highly unusual and wouldn't really work.

 

[BJC]

I vote with GTstartup.
I have seen the note "Short unused CT terminals" on enough drawing and specs to know there are some people who do it know matter what.
If your connecting X1-X5, X2,X3 and X4 should be open.

 

[slavag]

Yes, BJC you are right, but short of unused CT terminals in multi-ratio type of CT are not damage it. You see lower current, in Stevenal's case we see 1600 instea 1200.
Regards.
Slava

 

[BJC]

slavag
They don't necessarly damage it but if shorted the currents can affect the accuracy.

 

[stevenal]

Scottf,

I didn't mean to suggest the CTs were actually wound around a common core; they're just acting that way. The shroud extends down to the CT support, which is made of plastic.

This morning we pulled the shroud off to get a look. No obvious problems. We had a meggar today, so began testing with it. 6X and 6Y were both shorted to ground. 4Y was okay. Thinking we may have a problem in the conduit, we pulled the leads out. With the leads going directly to the CTs, we again read a dead short to ground. No ground leads go to the CTs, and they appear to be on non-conductive spacers. We propped up the 6X CT on some rolls of electrical tape and retested it. We now had some measurable Meg-ohms on it while 6Y below still read a direct short. We removed the CTs from the bushing and put them directly on a metal grid. Meggaring between the grid and winding, again showed a direct short. Raising the CTs above on lineman gloves yielded acceptable results. After reconnecting the ground lead to the wrapping of the CT, we again read a direct short from winding to wrap. We clipped the meggar leads across some of the spacers that we assumed were meant to be insulators. Instead we found zero Meg-ohms.

At this point we left the station, and brought the CTs back to our warm shop for further testing. Here they test fine, passing all ratio tests and meggar tests. We also have respectable Meg-ohm readings between windings and wrap.

This station is close to the beach, and experiences salt fog conditions. The salt gets everywhere, and attracts any available moisture. I suspect there may be some salt under the wrappings of the CTs, since they do not appear to be sealed in any way. The shroud can keep the rain off, but not keep the fog out. Don’t know why only one bushing of one breaker of one station would be the only one affected, though. The weather has been clear during all of our outdoor testing.

 

[scottf]

Stevenal-

Hmm....

Normally internal low meggar readings on BCTs is due to moisture intrusion. The moisture soaks into some of the insulation material and brings down the dielectric strength.

Normally salt deposits alone wouldn't do that and if they did, you would expect to see it in lab conditions as well.

Definitely a head scratcher.

With the relatively low cost of wrapped BCTs, you may want to consider just replacing them to be safe. (disclaimer...I work for a company who makes BCTs

 

[stevenal]

scottf,

What insulation materials are used that are subject to moisture intrusion and resulting dielectric strength reduction? And is there a better material for this application? Or should we be specifying sealed units, rather than wrapped? I don't mind replacing the CTs, just would like to think we're fixing the problem rather than just postponing it.

We'll be setting the CTs outside this evening. Under cover, but subject to the morning dew. We'll re-test 1st thing tomorrow.