A window CT (zero-sequence or single-phase ground CT) installed over shielded medium-voltage cable, such as 15 kV MV-105, reads correctly only if the shield current through the window is controlled. The shield does not block the magnetic field. It adds its own current to the CT's measurement, and that current decides whether a 50/51G relay trips on external events or stays silent for a real cable fault.
What does the operator see when shield current reaches the ground CT?
Two opposite screen-level symptoms come from the same cause.
| What the operator sees | Cause at the CT |
|---|---|
| 50G/51G element picks up during an external event (utility transmission fault, arcing in an upstream interrupter) with no fault on the protected feeder | Ground current returns on the cable shields, both shield ends are grounded, and the shield current passes through the window without cancellation |
| Several feeder breakers trip together on a single upstream event | Shield wires terminated incorrectly back through each zero-sequence CT, so every CT sees returning shield current |
| A large motor trips on the 50G element after a partial failure of an upstream vacuum interrupter | Ground bond between the two substations is not intact, so ground current takes the path through the cable shields past the motor's GFCT |
| Cable fault exists, but the ground relay does not operate until the shield burns open | Shield is grounded at one end only and its current passes back through the window, opposing the conductor current |
The first three are nuisance trips (false positives). The fourth is a blind relay (false negative). Diagnose by asking which way the shield current flows through the window relative to the conductor current.
Does the shield stop the window CT from working?
No. The CT responds to the net current through its window: the phasor sum of every conductor that passes through the core. The copper tape shield is a nonmagnetic conductor, so it does not screen the magnetic field of the phase conductor. It only matters when it carries current.
- Shield carries no current: the CT sees only the conductor current (or, for three phases through one window, the residual current).
- Shield carries current through the window: the CT adds it to the conductor current as a phasor sum.
- Cable fault current: the fault current leaves on the conductor and returns on the shield, so it is often close to 180 degrees out of phase with the conductor current. The CT sees the difference, which can be near zero.
The shield's purpose is a uniform electric field around the conductor. That has no bearing on whether the CT's magnetic coupling works. The concern is current, not field shape.
Is the copper tape shield strong enough to affect relay operation?
Yes. On 15 kV MV-105 cable the tape shield carries enough current, in either direction, to change what a sensitive ground relay measures. A common ground-fault scheme uses a 200:5 or 50:5 CT into a 50/51G relay with a low pickup for fast clearing of downstream ground faults. A low pickup on a low-ratio CT means small shield currents are a meaningful fraction of the setting.
- Both shield ends grounded: an external fault or normal ground current can circulate through the three shields. If the shield currents pass through the window without cancellation, the sum can exceed pickup.
- One shield end grounded: the shield carries the whole fault current for a cable fault. If that shield passes back through the window, the CT sees conductor and shield currents cancel and the relay does not sense the fault until the shield wires burn up.
Which shield termination approach works best?
Three arrangements are possible. Only two give a relay that responds to the conductor current alone.
| Arrangement | Shield current in the window | Result |
|---|---|---|
| A. Shield passes through the window as part of the cable, ground lead goes straight to ground on the far side | Full shield current, same direction as the cable shield | Blind to cable faults, or nuisance trips from external ground current. Most common installation error. |
| B. Shield passes through the window, then a made-up shield lead returns back through the core in the opposite direction before grounding | Shield current cancelled (goes through once each way) | Relay sees conductor or residual current only |
| C. Shield grounded on the cable side of the window so shield current never crosses the core | None | Same result as B when the ground point and cable geometry allow it |
Recommendation: use B when the shield already passes through the CT with the cable and the termination geometry puts the ground point beyond the core. Route the shield ground lead back through the window in the opposite direction before it lands on the ground bus. Use C when the termination hardware lets you ground the shield before it reaches the CT. B needs no cable re-termination, and the cancellation works for both ground-fault current on the shield and circulating current from an external event.
How do you route the shield lead through the CT?
- De-energize, lock out, and prove dead. Verify the CT secondary is shorted or connected to the relay burden before disturbing primary conductors.
- Identify every point where each shield is grounded (both ends of the cable run, plus any intermediate bonds) and record which shields pass through the window.
- Trace the direction each shield travels through the core. For three single-conductor cables through one zero-sequence CT, treat all three shields separately.
- For every shield that passes through the core, make the shield ground lead long enough to reach back through the window. Feed it through the core in the direction opposite to the cable shield, then land it on the ground bus.
- Confirm the lead does not touch the cable shield or other grounds before it passes through the core, so no current bypasses the loop.
- Confirm no second ground connection exists on the shield between the window and the termination that would let shield current avoid the returning lead.
- Check that the ground bond between substations (main sub to distribution sub) is intact. A broken bond forces ground current onto the shields.
What do external-event trips at 13.8 kV and on a 12000 HP motor teach about shield paths?
Two separate 13.8 kV industrial plants had every feeder breaker trip on ground-fault relays during a transmission line fault at the adjacent utility substation. Utility fault current returned via the cable shields, and the shield wires had been terminated incorrectly back through the zero-sequence CTs. Each relay saw the same shield current and picked up.
A 12000 HP motor tripped on the 50G element after a vacuum interrupter on a feeder breaker at the main substation partially failed. The arcing produced noise and spikes on the power system. The ground bond between the main sub and the distribution sub was not intact, so ground current looked for a return path to the main sub. The motor sat near the distribution sub, and the path through its cable shields and past the GFCT was enough to pick up the 50G element.
Two lessons follow. Shield termination through the CT and the station ground bond are one system. A correct CT loop does not fix a broken ground bond, and a good bond does not fix a shield terminated through the core the wrong way.
How do you verify the ground CT ignores shield current?
- With the circuit de-energized, apply a low-voltage test current through the shield ground lead path only (shield and returning lead) and measure the CT secondary. Expect zero within instrument resolution. Any reading means the returning lead is not cancelling the shield current.
- Apply the same test current through the phase conductor only and measure the secondary. Expect a value that matches the CT ratio (for example, 200:5 or 50:5 as installed).
- Reverse the shield lead loop direction on a test basis. The reading should change from zero to a nonzero value, which confirms the loop is doing the cancelling and is wound the right way.
- Record the shield ground points and the direction each lead passes through the core in the as-built drawings.
- With the feeder energized, clamp-measure current on each shield and on the ground lead. Compare with the relay's 50G and 51G pickup settings, and check the relay event record after any upstream disturbance for a residual-current trace.
What happens if the shield passes through the window CT and is grounded straight on the far side?
Shield current is summed with conductor current by the CT. During a cable fault the shield carries the return current in the opposite direction, so the CT sees the difference and the 50/51G relay may not operate until the shield wires burn up.
What happens if both ends of the cable shield are grounded and the shield passes through the CT uncancelled?
External ground current or a nearby system fault can circulate through the shields and the CT adds it to the reading. A low-pickup 50G element on a 200:5 or 50:5 CT can trip even though the protected feeder is healthy.
What happens if the ground bond between substations is not intact?
Ground current seeks a return path and uses the cable shields, so the shield current passes through nearby GFCTs. A 12000 HP motor located near the distribution sub tripped its 50G element for exactly this reason.
What happens if the shield ground lead is looped back through the CT in the wrong direction?
The lead adds to the shield current instead of cancelling it, so the relay sees shield current at twice the effective coupling. Verify by injecting test current through the shield path only: the CT secondary must read zero, and a nonzero reading means the loop direction must be reversed.