An insulation resistance reading is a single number that answers a question with too many causes. Five megohms can mean dry and healthy insulation with a small area, or wet and dying insulation with a large area. The polarisation index separates those two cases without any extra equipment, using only time.
The method is to apply a DC voltage, read the current after one minute, keep the voltage on, and read again after ten minutes. The ratio of the two resistances is the index.
The arithmetic and the physics
| Quantity | Value | What it reflects |
| Test voltage | 500 V DC for low voltage equipment, higher for medium voltage | Must be below the insulation rating |
| First reading | At 60 seconds | Charging current has decayed, absorption is still running |
| Second reading | At 600 seconds | Absorption is nearly complete |
| Index | R at 10 min divided by R at 1 min | A ratio, so it is independent of size |
| Typical pass | 2.0 or higher | Industry convention, not a universal clause |
| Suspect | 1.0 to 2.0 | Moisture, contamination or ageing |
| Failed | Below 1.0 | The reading is falling, which means conduction is dominant |
The physics is absorption. When DC is applied to a dielectric, dipoles align and charge migrates into the material, and that movement looks like a current that decays over minutes. Dry insulation shows a large decay, so resistance rises and the ratio is high. Wet insulation conducts through the water, the absorption component is swamped, and the ratio collapses toward 1.
Why the ratio is more useful than the absolute value
Absolute resistance depends on the area of the insulation, the temperature and the test voltage. A large machine has a lower insulation resistance than a small one in the same condition, simply because there is more material in parallel. The ratio cancels all of that: it compares the material with itself, so a 5 MΩ reading that rises to 12 MΩ is healthy at any size, and a 200 MΩ reading that rises to 210 MΩ is not.
That is why PI is the preferred check on equipment that cannot be taken out of service for a long test. It needs ten minutes and no reference value.
Where the index lies
- Very high absolute resistance. Above a few teraohms the meter noise dominates and the ratio becomes meaningless. Read the absolute value first.
- Temperature differences. A ten minute test on a machine that is cooling changes the resistance for thermal reasons, not dielectric ones.
- Surface contamination. A conductive film on a terminal block can hold the ratio near 1 while the bulk insulation is perfect, and cleaning changes the result completely.
- Capacitive charging. Large capacitance keeps the charging current flowing longer, which can depress the one minute reading and inflate the index.
Running it so the number means something
Discharge before the test and after it. Record the temperature, the humidity and the test voltage. Use guarded measurement if surface leakage is suspected, and keep the leads clear of the specimen. Then take the two readings from the instrument display rather than from a hand-held stopwatch, because 60 seconds and 75 seconds give different answers.
A insulation resistance tester with a timed PI function removes the timing error entirely, and the site index lists the testers by output voltage. For equipment with a history, plot the index over years: a slow decline from 3.5 to 2.2 is a better warning than any single test.
Standards referenced: IEC 61010-1; IEC 60884-1.
