Acceptance Testing a Relocated 1500 kVA Dry-Type Transformer

Turns ratio

• 28.80 / 28.75 = 1.0017 → +0.17% high. Field practice for a two-winding transformer is agreement within about 0.5% of nameplate (tighter if the nameplate ratio is exact). Acceptable.

Winding resistance (75 °C corrected)

• Average = (0.982 + 0.981 + 1.164) / 3 = 1.042 Ω
• Imbalance = (max − min) / average = (1.164 − 0.981) / 1.042 = 17.6%
• H3-H1 alone is +11.7% from the average.
• Commonly applied criterion is on the order of 2% between phases (confirm against the manufacturer's data and the current NETA MTS table). Fails by a wide margin.

Excitation current

• Average = (1.9 + 1.8 + 3.4) / 3 = 2.37 A
• H3 = +43% from the average. Typical acceptance is within about 10% between phases. Fails.

Insulation resistance of 5 GΩ is good but is a go/no-go check, not a condition indicator.

Verdict: two independent tests point at the same phase. Do not energize.

Candidate faults, most to least likely given the shipping history:

1. A poor connection in the H3 circuit - a not-fully-seated bushing stud, damaged lug, or a terminal that was re-landed without proper torque. This raises that winding's resistance directly. On a delta primary it can also distort the excitation current, because the applied voltage per phase is no longer balanced as the test set sees it.
2. A tap or connection problem - the three phases not on the same tap position, or a tap link not fully seated on H3.
3. A shorted turn in the H3 winding - this is the one that must be excluded, and it is consistent with high excitation current. A few shorted turns do not necessarily move the ratio much.
4. A damaged bushing or a contaminated creepage path, which would show on insulation resistance - and this one is clean at 5 GΩ, so it is lower on the list.

Next steps, in order:

• De-energized visual and mechanical inspection of H3: bushing seating, torque, tap link position on all three phases.
• Repeat winding resistance with a Kelvin bridge or a four-wire micro-ohmmeter, stabilized, with lead-length compensation, and compare all three pairs again after re-torquing. If the reading normalizes, it was the connection - document and retest excitation.
• Repeat excitation at reduced voltage (e.g., 50–70% of rated) and compare phases; a shorted turn tracks with voltage and produces an abnormally high magnetizing current on one phase.
• Perform SFRA and compare the three phases against each other; a winding deformation or turn short shows as a shift in the high-frequency response.
• Core-to-ground and winding-to-winding insulation tests, and a check that the core is properly grounded at one point.
• Confirm the tap changer position is identical on all phases and matches the nameplate ratio used.

Do not apply full voltage until the resistance and excitation anomalies are explained.

The ratio test is a coarse, low-sensitivity test.

• A 13.8 kV delta primary has hundreds of turns per phase. A shorted turn - even a handful - changes the effective turns ratio by a fraction of a percent, which lands inside the measurement resolution and inside the 0.5% acceptance band. The ratio can pass while a winding is actively shorting.
• A shorted turn forms a closed loop that only draws significant current under excitation. The ratio test applies a low voltage and measures a voltage division; it does not stress the turn-to-turn insulation or reveal the circulating current in a shorted loop.
• A high-resistance connection also barely moves the ratio, because the ratio test draws almost no current - a few milliamps - so a joint that looks fine at milliamps can fail under load.
• Conversely, ratio is good at catching gross problems: open windings, reversed sections, wrong tap, wrong winding connections. It is a screening test, not a condition assessment.

This is why acceptance uses ratio, winding resistance, excitation current, insulation resistance, and (where warranted) SFRA together - and why the decision here rests on the resistance and excitation results, not on the ratio.