The 120 Hz test generator draws high current even though the test object consumes little real power. The number that matters is winding-to-ground charging current: a large transformer presents substantial capacitance, and that current heats and sizes the generator, test transformer, cables, and switching equipment. A correctly connected shunt reactor supplies opposing reactive current locally.
Capacitive current and thermal load
During an offline phase-to-ground insulation test, the energized winding acts as one capacitor electrode. The grounded core, tank, other windings, and surrounding structure form the opposing electrode system. Current flows through the insulation capacitance rather than through a normal load path.
For sinusoidal excitation, the charging current and capacitive reactive power are:
Use the voltage actually applied across the measured capacitance. A phase-to-phase nameplate voltage is not automatically the correct value for a phase-to-ground test calculation.
This is heat, not logic. Even when dielectric losses are small, the test source must carry the full RMS charging current unless a reactor supplies most of it. Transformer winding capacitance grows with physical size and construction, so a test set adequate for a motor or small transformer may reach its current or kVA limit on a large unit.
A shunt reactor connected in parallel with the test object draws lagging current. The capacitive branch draws leading current. These currents circulate between the reactor and capacitance, leaving the source to provide the residual mismatch, losses, excitation current, and measurement-system load.
Test frequency and excitation topology
Frequency affects core flux and insulation charging current in opposite ways. For a given applied sinusoidal voltage, transformer core flux varies approximately with V/f. Raising frequency reduces flux and the risk of core saturation during an induced-voltage test. At the same time, rises directly with frequency.
At 120 Hz, a fixed capacitance and voltage draw twice the charging current they draw at 60 Hz and 2.4 times the current at 50 Hz. A higher test frequency can therefore solve an excitation problem while creating a larger reactive-power problem. Using 120 Hz and a shunt reactor together is not inherently redundant.
The connection determines which mechanism dominates:
| Test arrangement | Dominant electrical behavior | Reason for the chosen measure |
|---|---|---|
| One phase energized to ground, with specified other conductors grounded | Insulation capacitance carries charging current | Use shunt compensation when charging current exceeds the practical source capacity |
| Induced or phase-to-phase transformer excitation | Core flux and magnetizing current become major constraints | Higher frequency reduces V/f for the required test voltage |
| Combined transformer test circuit | Magnetizing current, charging current, and generator regulation interact | Use the specified frequency for flux control and calculate compensation from the complete connected capacitance |
The specified installation uses 120 Hz. Other reported transformer-test practice includes 180 Hz and older 400 Hz systems, but those frequencies are not interchangeable with 120 Hz unless the approved test procedure and equipment ratings permit the change.
Reactor and source alternatives
| Approach | Connection and purpose | Source impact | Deciding limitation |
|---|---|---|---|
| Parallel shunt reactor | Connect from the energized test bus to the return or ground reference, electrically parallel with the test-object capacitance | Reduces net reactive current and required source kVA | Reactor voltage, current, insulation, tuning range, loss, and duty rating |
| Larger test source | Supply the uncompensated capacitive load directly | Carries all charging current | Transport, generator capacity, test-transformer current, cables, and switchgear |
| Series reactor or source impedance | Place impedance in the feed path for current limiting, filtering, isolation, or resonant operation | Changes voltage regulation and circuit resonance | It does not perform ordinary shunt power-factor compensation |
| Higher excitation frequency | Change V/f to control transformer core flux |
Raises capacitive current in direct proportion to frequency | Test specification, source waveform, insulation stress, and connected capacitance |
For a source limited by winding charging current, use parallel shunt compensation. A drawing that shows a reactor “between” the generator and LV bushings is ambiguous: it may describe physical placement while the reactor is electrically connected from that bus to ground, or it may show a true series reactor. Resolve the one-line diagram before calculating a rating. The two connections serve different purposes and produce different terminal voltages.
Compensation quantities and limits
| Quantity | Calculation or limit | Where to obtain it |
|---|---|---|
Test frequency, f
|
Use 120 Hz for the specified setup | Approved test procedure and source indication |
Connected capacitance, C
|
Include every component connected at the energized node | Capacitance measurement in the final test configuration |
| Capacitive current, | 2 * pi * f * C * V |
Calculate, then compare with measured branch current |
| Capacitive reactive power, | 2 * pi * f * C * V^2 |
Calculate at maximum test voltage |
Ideal compensating inductance, L
|
1 / ((2 * pi * f)^2 * C) |
Calculate, then select from the reactor data |
| Reactor reactive power | Reactor nameplate and manufacturer data | |
| Residual source current | Vector sum of capacitive, inductive, magnetizing, and loss currents | Source and branch-current measurements during staged energization |
| Operating limits | Voltage, current, thermal duty, insulation, and frequency ratings | Generator, test transformer, reactor, cable, and switchgear nameplates |
A reported 0.5 microfarad test object at 11 kV and 50 Hz illustrates the arithmetic:
Referred ideally through an 11 kV-to-240 V test transformer, 1.73 A becomes about 79.2 A at 240 V. Rounding the high-voltage current to 2 A gives about 91.7 A, which explains an approximate 90 A estimate. Multiplying 1 A by 11000/240 would give 45.8 A, not 90 A.
Those values illustrate frequency scaling; they are not ratings for the large transformer. Measure its connected capacitance and use its specified test voltage.
Shunt-reactor selection procedure
- Classify the test as phase-to-ground applied voltage, induced phase-to-phase voltage, or a combined circuit. Record which phases, neutrals, tank, core, and other windings are energized, grounded, or left floating.
- Measure capacitance in the intended connection. Include bushings, coupling capacitors, leads, divider networks, and any other component tied to the energized node.
- Read the maximum test voltage, frequency, ramp sequence, and dwell requirement from the approved procedure. Use voltage to ground for each branch calculation.
- Calculate and . Separately identify expected magnetizing current and source losses; a reactor cancels reactive components only to the degree that their current vectors oppose.
- Compare uncompensated current with every series component rating, including the generator, test transformer, conductors, connectors, and switching equipment.
- Select a reactor or reactor combination that approaches the required inductance at 120 Hz without exceeding voltage, current, insulation, or thermal-duty ratings. Adjustable steps help accommodate different transformer capacitances.
- Confirm that the reactor is electrically in parallel with the capacitive test object when its assigned purpose is reactive compensation. Treat a true series reactor as a separate current-limiting, filtering, or resonant-network design.
- Model or calculate the complete network before energization. Near resonance, branch current and internal voltage can be much larger than source current, so low generator current alone does not prove that equipment is lightly stressed.
Partial-discharge signal integrity
The shunt reactor’s primary job is reactive-current compensation, not amplification of the partial-discharge pulse. A conventional detection circuit uses a coupling capacitor and measurement impedance to transfer high-frequency discharge components into the measuring instrument while rejecting most of the power-frequency voltage.
A stiff source does not by itself eliminate internally generated discharge activity. The measurable pulse depends on the test-object impedance, coupling network, grounding, lead inductance, bandwidth, and calibration path. Adding a reactor changes the network impedance and may change both noise transmission and pulse response, so repeat calibration in the final configuration.
Solid-state supplies can contribute switching harmonics and repetitive transients. A rotating generator connected to a capacitive network can also present regulation or excitation concerns. Separate external noise from internal discharge with a background measurement, clean single-point measurement grounding, physical separation of power and signal leads, and comparison of activity with applied-voltage phase and magnitude.
| Observed symptom | Likely cause | Diagnostic |
|---|---|---|
| Source current reaches its limit below test voltage | Uncompensated capacitance or incorrect reactor step | Measure capacitive and reactor branch currents and recalculate at the actual frequency |
| Low source current but high reactor current | Expected circulating reactive current near compensation | Compare reactor branch current and duty with its ratings |
| Terminal voltage rises sharply for a small control change | Resonant interaction or generator regulation instability | Stop the ramp and review the complete impedance and excitation system |
| Repetitive pulses remain with the test object de-energized | Source, grounding, switching, or environmental noise | Run a background scan and isolate auxiliary equipment systematically |
| Calibration response changes after adding the reactor | Measurement-network impedance changed | Recalibrate with every power and measurement component connected |
Commissioning and verification
- Verify the one-line diagram, grounding plan, clearances, interlocks, discharge path, and component ratings before applying voltage.
- Record background noise with the final leads, coupling network, reactor, and source connected.
- Inject the approved calibration signal at the specified calibration point and confirm stable instrument response.
- Raise voltage in controlled stages while recording test voltage, source current, reactor current, frequency, waveform, and partial-discharge background.
- Confirm that source current falls by the calculated amount while reactor branch current remains within its continuous or short-time duty rating.
- Check that the test voltage remains controllable throughout the ramp. Investigate hunting, abrupt voltage rise, waveform distortion, unexpected heating, or a branch-current departure from calculation before proceeding.
- Repeat the calibration check after the test configuration changes and after the voltage run, according to the approved procedure.
Acceptance requires more than reaching the target voltage. The measured current balance must agree with the calculated capacitance and selected reactor step, all component limits must remain satisfied, voltage control must remain stable, and the calibrated detection circuit must distinguish test-object activity from background noise.
Frequently asked questions
How do I know whether a transformer PD test needs a shunt reactor?
Calculate and compare the referred current and kVAr with the generator, test-transformer, cable, and switchgear ratings. Add shunt compensation when charging current, rather than real power, sets the source limit.
How do I size a shunt reactor for a 120 Hz PD test?
Measure the total connected capacitance and start with L = 1 / ((2 * pi * 120)^2 * C). Then check reactor voltage, branch current, insulation, losses, tuning range, and thermal duty at the maximum test voltage.
How do I connect the reactor to the transformer test circuit?
For reactive compensation, connect it from the energized test bus to the circuit return or ground reference, parallel with the test-object capacitance. If the drawing places it in series with the generator lead, obtain clarification because that connection serves a different function.
How do I reconcile using both 120 Hz and a shunt reactor?
The 120 Hz source controls transformer core flux through the V/f relationship, while the reactor offsets capacitive charging current. Raising frequency reduces flux for a given voltage but increases capacitive current, so both measures can be necessary.
How do I know when to stop a transformer PD test?
Stop the voltage ramp for unstable regulation, abrupt voltage rise, unexpected heating, excessive branch current, waveform distortion, failed calibration, or noise that prevents valid measurement. Discharge and ground the circuit under the approved safety procedure, then escalate the connection, reactor selection, or excitation behavior to the test-equipment and transformer manufacturers through their official support channels before re-energizing.