Insulation and dielectric
Insulation resistance testing applies a direct voltage across insulation and measures the small current that results, reported as a resistance. It shows whether moisture, contamination or aging has weakened the insulation on transformers, cables, motors, switchgear and breakers, and it is a routine check before energizing and on a maintenance schedule.
An insulation resistance test applies a steady direct voltage between a conductor and ground, or between two separate windings or phases, and measures the current that flows through and across the insulation. The instrument divides the applied voltage by that current and reports the result as a resistance, usually a very large one. Healthy insulation passes only a tiny current, so a high reading is good news and a low or falling reading is a warning. The test is used on power and distribution transformers, power and control cable, motors and generators, switchgear and busway, circuit breakers, and many other assemblies whose job is to keep energized parts apart from each other and from ground. Done as intended it is non-destructive: the voltage is chosen to stress the insulation without approaching its breakdown strength. The reading is not a property of the material alone. It reflects the whole insulation system, including its surface condition, its moisture content, its temperature and the size or length of the equipment under test, which is why results are always read together with the conditions in which they were taken. It is among the most widely performed electrical tests in the field because it is quick, repeatable and needs only a portable instrument.
Insulation usually weakens gradually before it fails. Moisture ingress, surface contamination from dust, oil or salt, thermal aging, mechanical damage and chemical attack all lower its resistance, and the change shows up in a reading long before the equipment trips. Finding it during a planned outage is far cheaper than finding it during a fault, when a breakdown can destroy a winding or a cable, start a fire, or put people at risk. For that reason the test is run on new equipment before it is energized, to confirm that shipping, storage and installation did no harm; after repairs or long periods of storage; and on a recurring schedule for equipment in service. The value of the schedule comes from the record it builds. A single reading says something, but a series taken under comparable conditions says much more, because a steady decline points to a developing problem even when every individual reading still looks acceptable. The test also screens equipment ahead of more demanding tests, since there is little sense in stressing insulation that is already visibly wet or dirty.
When a direct voltage is applied to insulation, the current that flows is the sum of three parts. The first is the capacitive charging current, a brief surge as the insulation charges like a capacitor, which dies away quickly. The second is the absorption current, which comes from the slow realignment of charge within the insulating materials and decays over a longer period. The third is the leakage current, a small steady flow through the insulation itself and across its surfaces. Because the first two parts shrink with time while the leakage part stays roughly constant, the total current falls and the resistance calculated from it rises for as long as the voltage stays on. Clean, dry insulation shows a steady climb over the duration of the test. Wet or contaminated insulation has a large leakage component from the start, so the reading is low and flattens early. Ratios of timed readings summarize how much the reading climbs. The dielectric absorption ratio compares two readings taken early in the test, and the polarization index compares readings taken over a longer span. Because both are ratios, they depend little on the size of the equipment and less on temperature than a single reading does, which gives a way to judge condition. Temperature matters a great deal: insulation resistance falls sharply as temperature rises, so a warm reading cannot be compared directly with a cold one unless both are corrected to a common reference temperature. Humidity and surface condition act on the leakage path in the same way, which is why they are noted with every result.
Safe preparation comes first. The equipment is de-energized, isolated from every source, proven free of voltage and grounded, following the owner's lockout and work-permit procedures. Connections that could carry the test voltage into other equipment, such as surge arresters, electronic devices or cables to remote loads, are disconnected or otherwise protected. The surfaces of bushings and terminations are cleaned and dried so that surface leakage does not mask the condition of the insulation underneath. The tester is connected between the part under test and ground, with the remaining windings or phases grounded, or connected to the instrument's guard terminal when the aim is to measure one insulation path at a time. The guard terminal diverts surface leakage, or the current of paths that are not of interest, around the measuring circuit. With the leads connected, the temporary ground is removed from the part under test only, and the other parts stay grounded. The test voltage is selected from the governing standard's table for the voltage class of the equipment, never by rule of thumb, and the instrument is confirmed to be in calibration. The voltage is then applied and held while timed readings are taken, and the values are written down as they are observed, not reconstructed afterward. When the test ends, the insulation is discharged through the instrument and then with a grounding hook, and it stays grounded for a period related to how long it was charged, because large capacitive equipment can hold a dangerous charge. Ambient temperature, equipment temperature and relative humidity are recorded at the time of the test, and the same instrument and voltage class are used on repeat tests so that results can be trended.
The record of an insulation resistance test is short and general. It holds the resistance readings at each timed interval, the ratio derived from them where one is used, the voltage class that was applied and how long it was held, the temperature of the equipment and of the surroundings, the relative humidity, and the identity of the instrument together with its calibration status. It also notes the as-found condition of the equipment, for example clean and dry, freshly removed from service, or visibly contaminated, and any connections that were lifted to isolate the item. These details matter because a resistance reading has meaning only alongside the conditions that produced it. Months later, a person comparing a new result with the stored one needs to know whether temperature, voltage and setup were comparable, or whether the difference they see may be explained by the way the test was done.
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 an absolute minimum that steps up with the voltage class of the equipment; the governing tables are found in ANSI/NETA ATS-2025 for acceptance testing and in ANSI/NETA MTS-2023 for maintenance testing, and they differ by equipment type. The second is correction of the reading to a reference temperature before any comparison, because warm insulation reads lower than cold insulation. The third is comparison between like parts, such as the three phases of a winding, where an outlier points to a local problem. The fourth is trending against earlier results on the same equipment, where a steady decline matters even if the readings stay above a published minimum. The fifth is the set of timed-reading ratios, the dielectric absorption ratio and the polarization index, which have their own published guidance and are judged on the shape of the climb. Manufacturer data and nameplate information come into play as well. Readings are best judged together, with the test conditions in view. 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 differ in purpose. An acceptance test establishes that new or repaired equipment is sound and creates the baseline for later comparison, so it is performed carefully under the best conditions available. A maintenance test checks the condition of equipment already in service, where the useful question is how the reading has changed since the last one, so repeatable conditions and method matter most. The equipment class changes the details. On rotating machines the ratio of timed readings and the winding temperature carry more weight, and the machine is usually tested with the other windings grounded and with a longer series of timed readings. On cables the length and the condition of the terminations affect the reading, and shields and neighboring conductors are handled according to the cable type. On transformers the combinations tested, windings to ground and windings to each other, and the oil-filled or dry-type construction shape the interpretation. On low-voltage gear and breakers, readings across open poles and from pole to frame are common, and the test voltage follows the low-voltage rating of the gear, with solid-state trip units, meters and surge protection disconnected or isolated before the test. In every case the procedure follows the governing standard for that equipment class.
The equipment classes and standard sections that call for this test, and the UltraDb forms that carry it.
The charging and absorption parts of the current fade as the insulation becomes polarized, leaving mostly the small steady leakage current. With the current falling and the applied voltage constant, the calculated resistance climbs. Clean, dry insulation shows a clear climb; wet or dirty insulation shows little, because leakage dominates from the start.
Insulation resistance drops as the temperature of the insulation rises, so a warm reading and a cold reading of the same sound equipment can look very different. Correcting each result to a common reference temperature, using the method given in the governing standard, lets results from different days and seasons be compared fairly.
When the test voltage is matched to the equipment rating and the procedure is followed, the test is considered non-destructive, because the voltage is chosen to stress the insulation gently and not to approach breakdown. Insulation that is already wet, contaminated or badly aged can still be harmed, so inspection and cleaning come first.
In the tables of the governing standards, for example ANSI/NETA ATS-2025 for acceptance work and ANSI/NETA MTS-2023 for maintenance work, which depend on equipment type and voltage class. Use your licensed copy of the standard, and compare against earlier results and manufacturer data as well.
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.