The three test modes on a power factor test set are not interchangeable. GST, UST, and Guard mode each measure a different current path through the test specimen, and applying the wrong mode to a given measurement either gives you a meaningless result or combines the insulation you are trying to evaluate with something else you are not. Understanding what each mode connects to is the foundation of setting up a test correctly.
A power factor test set has three terminals that matter for understanding the modes: the high-voltage (HV) output, the measuring (low-current) input, and the guard terminal. The HV output applies the test voltage. The measuring input is the ammeter side — it measures the current flowing back through the return path. The guard terminal provides a third connection that can be used to divert unwanted current away from the measurement.
The test set also has a safety ground connection that bonds the instrument chassis to substation ground. This is separate from the measurement connections. The safety ground is always connected. The measurement connections change with each mode.
In GST, the specimen being tested is grounded — meaning the low-voltage end of the insulation is connected to the ground of the test set. The high voltage is applied to the high-voltage terminal of the specimen, and the return current flows from the low-voltage terminal through the measuring circuit to the test set return, then to ground.
GST is the default mode for most transformer winding tests. On a CHL test, the H winding is energized at high voltage, the X winding is connected to the measuring input, and the transformer tank (which is grounded) completes the circuit. The current that flows through the insulation between H and X, and between both windings and the tank, is measured. This gives the total insulation current for the combined winding-to-winding and winding-to-ground insulation paths.
The limitation of GST is that it measures all current paths between the HV terminal and ground simultaneously. If there are multiple insulation sections in parallel between the HV conductor and ground, GST sees their combined parallel admittance. You cannot distinguish between the winding-to-ground insulation and the winding-to-winding insulation from a single GST reading alone.
In UST, the low-voltage terminal of the specimen is connected to the measuring input but the specimen is not grounded. The test measures only the current that flows between the HV and the specific terminal connected to the measuring input, rejecting any current that flows to ground.
This is how you separate individual insulation sections in a transformer. To test only the H-to-X winding insulation (CH test), the H winding is energized, the X winding is connected to the measuring input, and the tank is left floating — not connected to either the return or the guard. The current that flows from H to X through the inter-winding insulation is measured directly, without including the H-to-tank or X-to-tank insulation in the result.
UST is also used for bushing test tap measurements. When testing C1 of a condenser bushing in UST, the test tap (which connects to the outermost foil layer) is connected to the measuring input and the flange is left floating. The current measured flows only through the insulation from the center conductor to the test tap — that is, only through C1. The C2 section from the test tap to the flange is not grounded and does not contribute to the measurement.
The distinction from GST is significant: UST measures a specific insulation path between two conductors, excluding ground. GST measures everything from the HV terminal to ground. When you compare CH and CHL measurements on a transformer and see a large difference between the two results, that difference contains information about the X-to-ground insulation condition that CH alone does not show.
Guard mode adds the guard terminal to either GST or UST to eliminate a specific stray current path from the measurement. The guard terminal is connected to a surface or conductor that carries leakage current you do not want to include in the result. Current flowing into the guard terminal is diverted around the measuring circuit — it does not register in the power factor reading.
The most common application is guarding out surface leakage on bushing porcelain or on transformer bushings during damp conditions. If a bushing exterior is wet or dirty, current flows along the porcelain surface from the HV terminal to ground. Without guarding, this surface current is included in the measurement and inflates the apparent power factor of the bulk insulation. With the guard terminal connected to a clean dry conductive band or clamp on the bushing surface below the contamination, the surface current flows into the guard and not into the measuring circuit. The result reflects only the bulk insulation.
Guard mode is also used in three-winding transformer tests to exclude one winding from a measurement between the other two. If you are testing H-to-X insulation (UST) but the tertiary winding Y is floating nearby and capacitively coupled to both H and X, the stray coupling current from Y adds noise to the measurement. Connecting Y to the guard terminal diverts that coupling current away from the measuring input.
Guard should be used deliberately, not as a default. Applying the guard incorrectly — guarding out a current path that is actually part of the insulation you intended to test — produces a result that is lower than the true condition of the specimen. A result that looks good because the test was set up wrong is worse than a result that correctly shows a problem.
The three standard transformer test configurations — CHL, CH, and CL — use different mode combinations and connection sequences. CHL is a GST measurement that includes both winding-to-winding and winding-to-ground insulation in one reading. CH is a UST measurement that isolates the inter-winding insulation. CL is another GST or UST measurement depending on the specific test protocol, covering the X winding to ground insulation.
The reason for running all three is arithmetic: the readings from CHL, CH, and CL are related, and calculating the expected values from two of the three and comparing to the measured third is a consistency check. A result that does not add up arithmetically suggests that one of the measurements picked up stray current from a connection error, a grounding problem, or external interference. That inconsistency is as diagnostic as the actual power factor value.
Manufacturers’ test procedures and Doble application guides specify the exact connection and mode for each test on each equipment type. Following the procedure as written is more important than improvising a connection that seems equivalent. Slight differences in which terminal is energized and which is measured produce different current paths, and the interpretation of the result depends on knowing exactly which path was measured.
Power frequency interference from adjacent energized equipment is the main source of measurement error in field testing. The test set applies 60 Hz voltage to the specimen, and the return current it measures is also at 60 Hz. Energized equipment nearby generates stray 60 Hz currents that couple into the test leads and specimen, adding to or subtracting from the measured current. Depending on the phase angle of the interference relative to the test voltage, it can make the power factor appear higher or lower than its actual value.
Modern test sets indicate the level of interference in the reading and flag results where interference exceeds a threshold. Some sets offer a second test frequency (50 Hz in a 60 Hz system, or a slightly offset frequency) and derive the true 60 Hz result by filtering out the interference mathematically. When interference is flagged, do not accept the reading without taking steps to identify and reduce the source. Shortening the test leads, re-routing them away from energized conductors, and ensuring all connections are secure are the first steps.
A reading that looks unexpectedly good or unexpectedly bad should prompt a mode and connection check before it goes into a test report. Test results that cannot be explained by equipment condition often can be explained by a connection that is different from what was intended.
Our field crews run full transformer and bushing power factor test programs across Florida and the Southeast. Send us your equipment list and we’ll respond within one business day.