Insulation and dielectric
Insulation power factor and dissipation factor testing applies an alternating voltage to insulation and measures the share of current that is lost as heat. A rise in that loss reveals moisture, contamination or aging in transformers, bushings, breakers, reactors and machine windings, often before a resistance test shows anything.
Insulation power factor and dissipation factor testing applies an alternating voltage across an insulation system and measures how much of the resulting current is in phase with that voltage, which is the part that turns into heat. Perfect insulation behaves as a pure capacitor and dissipates almost nothing. Real insulation dissipates a small amount, and the ratio of that loss to the reactive charging current is the dissipation factor, also called tan delta. Power factor is the closely related ratio of loss to total apparent current, and for the small values seen in sound insulation the two are nearly identical. The test is run with a dedicated insulation analyzer or bridge, usually at line frequency or at a selectable frequency, and it is used on liquid-filled and dry-type transformers, reactors and voltage regulators, circuit breakers and switches, instrument transformers, bushings, rotating machine windings, busway and outdoor bus structures, and on samples of insulating liquid. Unlike a direct-voltage resistance reading, it looks at the dielectric loss of the whole insulation system and responds to moisture, contamination and aging that spread through the material, which makes it one of the most informative field tests for high-value apparatus.
Insulation rarely fails without warning. Moisture entering a transformer, oil that has oxidized or picked up conducting contamination, a bushing whose condenser layers have been damaged, a winding whose varnish has aged, or an insulator with surface tracking all raise the dielectric loss well before a breakdown. A loss measurement can expose these conditions at a time when a planned outage and a repair are still possible, instead of a forced outage after a fault. The result also protects investments in two ways. At commissioning it confirms that shipping, storage and installation did not harm the insulation and it creates a baseline against which the equipment is judged for the rest of its life. In service, a series of results on the same unit shows a developing problem as a slow drift even while every single reading still looks plausible. Because it detects bulk and distributed defects that a simple resistance reading can miss, many standards list it as an optional or supplementary test on important apparatus and call for it where the cost of a failure is high.
Insulation can be modeled as a capacitor with a resistor in parallel. When alternating voltage is applied, the capacitor draws a charging current that leads the voltage by a quarter cycle, and the resistive path draws a small loss current that is in phase with it. The loss angle is the small amount by which the total current falls short of a perfect quarter-cycle lead. Its tangent is the dissipation factor, and the cosine of the phase angle between voltage and total current is the power factor. The meter measures the watts lost and the volt-amperes supplied, and in the same pass it reads the capacitance, so a change in geometry or a shorted layer shows up too. Anything that adds conduction or polarization loss to the insulation, such as dissolved water, ionic contamination, carbon or partial discharge, increases the loss current while leaving the charging current almost unchanged, so the ratio climbs. Loss rises strongly with temperature, which is why every reading is tied to a measured temperature and corrected to a reference. The loss also depends on test voltage and frequency, and surface leakage across a dirty or wet bushing adds to the loss, which is why surfaces are cleaned and why the guard circuit of the instrument is used to separate one insulation path from another.
The equipment is de-energized, isolated from every source, proven dead and grounded under the owner's lockout and permit procedures. Bus, cables, surge arresters and other connected apparatus are removed from the terminals, and bushing and insulator surfaces are cleaned and dried. On a transformer or reactor, each winding is shorted together at its terminals so that the winding behaves as a single electrode. The analyzer is then connected so that each insulation path is measured separately: from one winding to another, and from each winding to ground. The ungrounded-specimen mode routes the current of the path under test through the measuring circuit and diverts the others to the guard, while the grounded-specimen mode measures everything to ground, and the mode used is the one the equipment and the procedure call for. Bushings are tested at their power-factor or capacitance tap where one is provided and by the hot-collar method where one is not. Circuit breakers are tested with the contacts closed and again across the open poles, and oil breakers add the tank loss index. The test voltage is applied within the equipment and instrument rating, never beyond what the manufacturer allows. In a switchyard or near energized equipment, electrostatic interference can distort the reading, so the instrument's interference suppression, reversal of the test voltage or a change of frequency is used. Top oil or winding temperature, ambient temperature and humidity are noted at the time, and the same setup is repeated on later tests.
The record holds the measured power factor or dissipation factor for each insulation path tested, the measured capacitance and the watts loss that go with it, the test mode and the test voltage that was applied, and the temperature of the insulation or oil together with the value corrected to the reference temperature. It also notes the weather and relative humidity, the connection arrangement used, the condition of bushings and surfaces, whether the unit was oil-filled, dry-type or gas-insulated, and the identity and calibration status of the instrument. These items let a later reader tell a true change in the insulation from a difference in the way the test was made, and they make results from different years comparable.
No single number decides whether insulation is acceptable, and this page does not give one. Published criteria take several shapes that are used together. The first is comparison of the corrected result against the manufacturer's or factory value and against the table for the type of equipment or liquid; the governing guidance is in ANSI/NETA ATS-2025 for acceptance testing and in ANSI/NETA MTS-2023 for maintenance testing. The second is correction of every reading to a reference temperature, because loss rises with temperature and a warm reading is not comparable with a cold one. The third is comparison between like parts, such as the same bushing position on three phases or identical units in the same installation, where an outlier points to a local problem. The fourth is trending against earlier results on the same equipment, where a steady rise matters even if the figures stay inside published guidance. The fifth is judging the capacitance alongside the loss, since a shift in capacitance can indicate shorted layers. Results are best read together with the resistance readings and the test conditions. For the actual values and tables, refer to your own licensed copy of the standards named above, which remain the authority for acceptance.
Acceptance and maintenance tests serve different purposes. At acceptance the test sets the baseline for new or repaired equipment, so it is made carefully under the best conditions available. In maintenance the useful question is how the result has changed since the last one, so repeatable conditions and a consistent method matter most. The equipment class changes the details. On liquid-filled transformers the winding insulation, the bushings and a sample of the liquid are measured separately. On dry-type units and reactors the winding insulation is measured to ground and between windings. On circuit breakers and switches, readings are taken with the contacts closed and open, and on oil breakers the tank loss index is added. On rotating machines the winding is measured, often at stepped voltages to see how the loss rises, a related check described under the power factor tip-up test. On instrument transformers and coupling capacitors the capacitance and loss are read per section of the stack. Many of these measurements are optional in the standards, which makes them a decision of the owner and the testing agency. They complement insulation resistance and dielectric withstand work and do not replace them.
The equipment classes and standard sections that call for this test, and the UltraDb forms that carry it.
Both describe dielectric loss. Dissipation factor, or tan delta, is the loss current divided by the charging current, while power factor is the loss current divided by the total current. For the small values of sound insulation the two are practically the same number, and they diverge only when the loss becomes large.
Dielectric loss changes strongly with temperature, so a warm winding and a cold winding of the same sound unit give different readings. Correcting to a common reference temperature, using the method in the governing standard or the manufacturer's data, makes results from different days and seasons comparable.
Yes. A resistance reading is dominated by leakage along the best path, while a loss measurement responds to moisture, aging and contamination spread through the material. For that reason the two tests are used together, and a rising loss on a unit with acceptable resistance is still a warning.
It is a comparison of the loss measured with the breaker open and closed, intended to reveal losses in the tank lining, the liquid and the lift rods that a single measurement would hide. Its interpretation follows the manufacturer's guidance and the published standards for oil breakers.
UltraDb references ASTM, ANSI, IEEE, NETA, NFPA, NEMA, FAA, TIA, and IEC standards descriptively, to identify the published methods its forms are built from. UltraDb and Verlecta are not affiliated with, endorsed by, certified by, or licensed by NETA, ASTM, ANSI, IEEE, NFPA, NEMA, FAA, TIA, IEC, or Megger. Standard names and section numbers are the property of their respective organizations. Always test to the edition your contract specifies. Where a standard is sold through the ANSI webstore, Verlecta may earn a commission on a purchase made through our link, at no extra cost to you. A commission never decides which standards we cite or how a form is built.