Insulation-resistance (IR) measurement is a useful diagnostic and cleanliness check, but it is not the IEC power-frequency withstand test. A high megohm value cannot certify dielectric withstand, and a low value is not automatically failed insulation until voltage, temperature, humidity, connected equipment, surface leakage and charging time are understood.
This guide explains how to apply DC insulation-resistance testing safely to MV switchgear during manufacture, FAT, commissioning and maintenance. It covers test boundaries, selection of DC voltage, charging/absorption/leakage current, guard terminals, phase-to-earth and phase-to-phase tests, environmental correction, PI/DAR limitations, discharge, trending and escalation to dielectric or condition-diagnostic tests.
Executive conclusions
- Use IR as a diagnostic screen and baseline; do not substitute it for IEC 62271 routine dielectric withstand.
- Define the connected test object before selecting voltage. VTs, surge arresters, sensors, electronics, heaters and secondary circuits can be damaged or distort results.
- Take the test voltage and acceptance from the manufacturer/project procedure; one universal “1 kV per kV” rule is unsafe and unjustified.
- Measure at a defined time after voltage application because current changes as capacitance charges and dielectric absorption decays.
- Temperature, humidity, condensation, contamination and surface condition can dominate results.
- A guard terminal can divert surface leakage from the measuring channel, but guarded and unguarded results answer different questions.
- PI and DAR can be misleading on compact mixed-insulation switchgear; use them only where the asset/manufacturer method supports interpretation.
- Discharge and prove the circuit safe after every test. Automatic instrument discharge is a control, not permission to touch immediately.
- Trend comparable phase/path results and investigate abrupt changes; do not chase one generic megohm threshold.
- Document all isolations and restore them with independent verification before energisation.
1. What IR measurement actually measures
An insulation tester applies DC voltage and derives resistance from measured current:
R(t) = Vtest / I(t)
The current after voltage application is not a single pure “leakage” component:
- capacitive charging current, initially high and rapidly decaying;
- dielectric absorption/polarisation current, decaying more slowly;
- surface leakage current, strongly affected by moisture/contamination and geometry;
- volume/conduction leakage through the insulation;
- connected-device current through arresters, indicators, RC circuits or electronics.
Therefore a reading at 5 seconds and one at 60 seconds are not equivalent. Record the time and observe the curve, not only the final display.
2. IR versus dielectric withstand
| IR diagnostic test | Power-frequency withstand test |
|---|---|
| DC, generally lower energy/voltage selected for diagnostic purpose | Specified AC test voltage, duration and configuration under IEC product standard |
| Reports resistance/current versus time | Demonstrates no disruptive discharge/breakdown under specified stress and acceptance |
| Sensitive to contamination, moisture and connected devices | Formal routine/type dielectric evidence |
| Useful before/after transport, cleaning or maintenance | Required where applicable by the contracted routine-test framework |
| No universal conversion to AC withstand capability | Cannot be replaced by a high megohm value |
IR can be a prudent pre-check before a formal withstand test, but passing IR does not guarantee the withstand test will pass. Conversely, an anomalous IR result should trigger investigation before higher stress is applied.
3. Define the test boundary
- complete main circuit phase to enclosure/earth;
- phase-to-phase path with other conductors earthed/isolated as defined;
- bus section only;
- feeder/cable-side primary circuit;
- breaker pole or switch component separately;
- VT primary/secondary winding only under its own approved method;
- auxiliary/control circuit group under a lower suitable voltage;
- specific bushing, insulator or contaminated compartment.
Draw the test circuit and list every connected component. The same panel can give radically different results with a VT, surge arrester, voltage-detection system or cable connected.
4. Connected equipment that needs special treatment
| Item | Risk or influence | Control |
|---|---|---|
| Voltage transformer | Creates winding path; test voltage may be inappropriate | Isolate/fuse/disconnect per manufacturer procedure |
| Surge arrester/RC snubber | Conducts or clamps DC; can be damaged | Disconnect only as approved; record/restoration |
| VPIS/VDS/sensor/LPIT | Electronic/capacitive path and voltage limitation | Follow device instructions; never assume MV withstand |
| Protection relay/meter/PLC | Low-voltage electronics vulnerable to test voltage | Define auxiliary test boundary and disconnect safely |
| Space heater/thermostat | Intentional resistance to earth or circuit path | Isolate and test separately as required |
| Power/control cable | Adds capacitance, absorption and leakage | State whether included; use cable procedure if separate |
| Temporary grounds/test links | Can short the test or alter path | Control through a written link/ground register |
Isolation itself creates risk: a forgotten disconnected arrester or VT can leave the commissioned plant unprotected. Use a numbered restoration checklist and independent verification.
5. Selecting DC test voltage
- Use the switchgear/component manufacturer’s approved diagnostic procedure.
- Respect the lowest-rated connected component, especially electronics and sensors.
- Distinguish main-circuit, secondary/control-circuit and cable test voltages.
- Do not derive DC voltage automatically from rated system voltage.
- Use the same voltage for trend comparison unless engineering assessment states otherwise.
- Ramp where the instrument/procedure requires it and stop on abnormal current/instability.
- Confirm instrument maximum energy/current and discharge capability for the test capacitance.
Typical commercial instruments offer several selectable voltages; this flexibility does not mean every setting is safe for every switchgear path. The approved method—not instrument availability—sets the voltage.
6. Safety controls
- Isolate all sources, prove dead and apply safety earths under the switching programme.
- Define how safety earths are temporarily repositioned for the test without exposing personnel.
- Fence the test area and use high-voltage warning/remote operation where appropriate.
- Control stored mechanical energy and remote controls.
- Use leads, probes and gloves rated for the test voltage and environment.
- Connect the earth/return lead first and remove it last.
- Stop if unexpected external voltage, unstable current, discharge or noise appears.
- Allow automatic discharge, then apply an approved manual earth/discharge device where required.
- Measure residual voltage and prove safe before touch.
- Maintain earth for sufficient time to address dielectric absorption/return voltage.
A capacitive circuit can retain hazardous energy:
Estored = ½CV²
Higher voltage increases energy with the square of voltage. Treat discharge as a planned test step, not an afterthought.
7. Preparation and environmental recording
- Inspect for moisture, condensation, dust, carbon tracking, insects, salt/cement/metal contamination and damaged insulation.
- Record ambient temperature, relative humidity and representative insulation/conductor temperature.
- Verify anti-condensation heaters and storage history.
- Clean/dry only with manufacturer-approved materials and methods.
- Allow solvent/moisture to evaporate before test.
- Close/open doors and covers in the defined configuration.
- Document cable/VT/arrester/sensor connections and earth-switch positions.
- Use the same test points and duration as the baseline.
Surface moisture can cause orders-of-magnitude change without bulk insulation failure. That does not make the result irrelevant—it identifies a service-readiness problem—but the corrective action differs from replacing an insulator.
8. Phase-to-earth and phase-to-phase configurations
- Phase to earth: connect the tested phase(s) to positive/high terminal as specified; bond non-tested phases and enclosure to return/earth.
- Phase to phase: test one phase against another or a bonded group according to the approved circuit, with enclosure safely earthed.
- Three phases tied: useful for overall insulation-to-earth trend but can hide one weak phase; follow with individual phases where required.
- Open breaker/disconnector: can separate bus and cable paths; verify isolation distance and connected components.
- Closed breaker: includes breaker and both panel sides; ensure safe/test method permits it.
Photograph or sketch the configuration. “L1 measured” is ambiguous unless the status of L2/L3, earth switch, breaker, VTs and cables is known.
9. Timed readings, DAR and PI
Timed ratios are commonly defined by the selected procedure, for example:
DAR = R60s / R30s
PI = R10min / R1min
- Confirm the manufacturer/procedure uses the same time definitions.
- PI is most established for certain rotating-machine insulation systems; it is not automatically an IEC acceptance criterion for MV switchgear.
- Compact dry insulation may charge quickly, producing a ratio near 1 despite sound insulation.
- Multiple components/surface leakage can dominate the curve.
- A very high 1-minute value can make the ratio numerically uninformative.
- Use raw time-current/resistance behaviour and absolute/trend context, not ratio alone.
10. Guard terminal: when and why
A guard terminal diverts selected surface leakage around the instrument’s measuring channel. It can help distinguish bulk insulation current from leakage over a contaminated surface or along a bushing. Official Megger information for its BM5200, for example, describes the guard terminal as minimising surface-leakage effects.
- Place the guard only according to a validated test diagram.
- Keep guard lead insulation and creepage suitable for the applied voltage.
- Do not interpret a high guarded result as proof that surface contamination is acceptable in service.
- Record both configurations if used and label them clearly.
- Do not exceed the instrument guard-current capability.
Guarded and unguarded readings measure different current paths. They should not be mixed in a trend database.
11. Step-by-step procedure
- Approve test voltage, circuit diagram, acceptance/trend rule and safety boundary.
- Identify lineup/panel/phase and all connected/isolated devices.
- Inspect/clean/dry and record environment.
- Verify tester model/serial/calibration, leads and voltage setting.
- Apply safety isolation/earths; configure test path using a controlled link register.
- Connect return/earth, high-voltage and guard leads with the source off.
- Warn/clear the area and apply voltage.
- Record voltage and resistance/current at specified times; observe stability.
- Stop on breakdown, unexpected current, noise or instrument limit.
- Remove voltage, allow automatic discharge, apply/verify approved earth and prove residual voltage safe.
- Repeat for remaining phases/configurations using comparable conditions.
- Restore all devices/links/grounds and independently verify.
- Evaluate phase, baseline and environmental trend; raise NCR/diagnostic action where needed.
12. Interpreting common patterns
| Pattern | Possible causes | Next checks |
|---|---|---|
| All phases low after storage | Condensation, dirt, connected device, cable capacitance/leakage | Inspect/dry, isolate sections, confirm configuration |
| One phase much lower | Local contamination/damage, VT/arrester, cable, tracking | Segment bus/feeder/breaker; inspect phase |
| Rises steadily with time | Normal charging/absorption behaviour | Compare curve and final value with baseline |
| Unstable/jumping | Discharge/corona, wet tracking, loose lead, external noise | Stop safely; inspect setup and insulation |
| High guarded, low unguarded | Surface leakage dominates | Clean/dry; verify creepage/service environment |
| Much lower after site cabling | Cable/terminations/additional equipment or damage | Separate switchgear and cable paths |
13. Temperature correction and trending
Insulation resistance generally changes strongly with temperature, but the correction factor depends on insulation system and manufacturer guidance. Avoid applying a generic “halve/double per 10 °C” rule to a mixed switchgear assembly without validation.
- Trend at comparable temperature/humidity where practical.
- If correction is approved, preserve raw value, corrected value, curve/factor and temperature.
- Compare identical phases and panels tested at the same time.
- Use logarithmic plots where values span decades.
- Investigate abrupt departures from baseline even above a broad minimum.
- Correlate with visual inspection, partial discharge, dielectric testing and service events.
14. Acceptance strategy
- Manufacturer/project minimum at specified voltage, time and temperature.
- Stable curve without breakdown/flashover or abnormal current.
- Reasonable phase/panel balance for identical configurations.
- Comparison with factory/commissioning baseline.
- No unresolved moisture, contamination or damaged insulation.
- Formal IEC dielectric test completed where required; IR never replaces it.
If no approved limit exists, label the test diagnostic rather than creating a contractual pass/fail number. Engineering review can still classify the condition and decide whether cleaning, drying, segmentation, withstand testing or replacement is needed.
15. Record and restoration
- asset/panel/phase and exact test circuit;
- breaker, disconnector and earth-switch positions;
- connected/disconnected VT, arrester, cable, sensor and electronics;
- test voltage, timed readings/current curve and duration;
- guard configuration;
- ambient/asset temperature and humidity;
- instrument/serial/calibration and lead set;
- acceptance/trend basis and result;
- cleaning/drying/retest/NCR history;
- residual-voltage/discharge completion;
- signed link/ground/device restoration checklist.
Common mistakes
- Calling IR a dielectric withstand test.
- Applying maximum tester voltage without reviewing connected equipment.
- Leaving VTs, arresters or sensors connected unintentionally.
- Comparing readings at different times/temperatures/configurations.
- Using one universal megohm threshold for every switchgear.
- Relying on PI/DAR without asset-specific interpretation.
- Using guard to hide unacceptable surface contamination.
- Touching before discharge/return-voltage verification.
- Forgetting an isolated protective device during restoration.
- Reporting only “> instrument range” with no voltage/time/path.
Official standards and primary references
- IEC 62271-1:2017+AMD1:2021 — current common dielectric/routine-test framework for high-voltage switchgear.
- IEC 62271-200:2021+AMD1:2024 — current metal-enclosed MV assembly dielectric requirements.
- IEC 62271-100:2021+AMD1:2024 — current circuit-breaker insulation/test requirements.
- IEC 62271-200:2021 base edition — base text associated with the current consolidated amendment set.
- Megger BM5200 official information and technical resources — practical instrument reference for timed IR, automatic discharge and guard-terminal use; not an IEC acceptance criterion.
- ISO/IEC 17025:2017 — competence/calibration framework where applicable.
Engineering note: Select voltage and interpretation from the switchgear/component manufacturer and project procedure. Where the product standard requires power-frequency withstand, IR remains complementary diagnostic evidence only.