The polarization index test is the most reliable way to assess insulation condition on motors, generators, and other rotating machinery. It goes deeper than a standard 60-second megger reading and gives you a result that’s independent of temperature and equipment size.
If you only have time for one insulation test on a critical motor, make it the PI test.
This guide covers exactly how the test works, the correct procedure per IEEE 43-2013, how to interpret the results (including cases where the PI can mislead you), and when to use it versus simpler methods.
Table of Contents
What Is the Polarization Index?
The polarization index (PI) is a ratio of two insulation resistance readings taken during the same test:
PI = IR at 10 minutes ÷ IR at 1 minute
That’s it. No complicated formula. You apply test voltage, record the reading at 1 minute, keep going for another 9 minutes, record the reading at 10 minutes, and divide.
The result tells you how the insulation absorbs and holds electrical charge over time. Good insulation absorbs charge steadily — resistance keeps climbing. Bad insulation is dominated by leakage — resistance stays flat or drops.
A PI of 2.0 means the 10-minute reading is twice the 1-minute reading. The insulation is absorbing charge normally. A PI of 1.0 means the readings are identical — no absorption happening, which points to moisture, contamination, or serious degradation.
Why the PI Test Works
When you apply DC voltage to insulation, three currents flow:
Capacitive current — charges the insulation’s natural capacitance. Dies out in seconds.
Absorption current — flows as the dielectric polarizes under the electric field. Decays slowly over minutes. This is the current that makes good insulation show rising resistance over time.
Leakage current — the steady, constant current that flows through and across the insulation. This is the “bad” current. It doesn’t decay.
At 1 minute, your reading includes a mix of absorption current and leakage current. The capacitive current is already gone.
At 10 minutes, the absorption current has largely decayed. What’s left is mostly leakage current.
In good insulation: Absorption current is significant at 1 minute but has decayed substantially by 10 minutes. So the 10-minute reading is much higher than the 1-minute reading. PI is high (2.0 or above).
In bad insulation: Leakage current dominates from the start. Absorption is weak or masked by the high leakage. Both readings are similar. PI is low (near 1.0 or below).
Why it’s better than a spot reading
A 60-second spot reading gives you one number that’s heavily influenced by temperature and equipment size. A large motor will read lower than a small motor at the same condition, simply because it has more insulation surface area for leakage.
The PI eliminates these variables. Because it’s a ratio of two readings taken minutes apart under the same conditions, temperature effects cancel out. Equipment size effects cancel out. What you’re left with is a pure measure of insulation quality.
That’s why IEEE 43-2013 recommends the PI test as the primary diagnostic for rotating machinery.
What IEEE 43-2013 Says
IEEE 43-2013 (IEEE Recommended Practice for Testing Insulation Resistance of Electric Machinery) is the authoritative standard for this test. Here are the specific requirements, referenced by clause and table.
Scope (Clause 1)
The standard applies to rotating machines rated 750 W or greater — synchronous machines, induction machines, DC machines, and synchronous condensers. It does not apply to fractional-horsepower machines.
Test Voltage (Table 1)
IEEE 43-2013, Table 1 specifies the DC test voltage based on winding rated voltage:
| Winding Rated Voltage | Test Voltage (DC) |
|---|---|
| Below 1,000V | 500V |
| 1,000V – 2,500V | 500V – 1,000V |
| 2,501V – 5,000V | 1,000V – 2,500V |
| 5,001V – 12,000V | 2,500V – 5,000V |
| Above 12,000V | 5,000V – 10,000V |
Note: “Rated voltage” means line-to-line voltage for three-phase AC machines, line-to-ground for single-phase, and rated DC voltage for DC machines.
Minimum PI Values (Table 3 / Clause 12.2)
The minimum acceptable PI depends on the thermal class of the insulation system, per IEC 60085:
| Insulation Thermal Class | Minimum Recommended PI |
|---|---|
| Class A (105°C) | 1.5 |
| Class B (130°C) | 2.0 |
| Class F (155°C) | 2.0 |
| Class H (180°C) | 2.0 |
Most modern industrial motors use Class F or Class H insulation. The minimum PI for these is 2.0.
Class A is the exception — its minimum is 1.5. Class A insulation is rare in new equipment but still found in older motors manufactured before the 1970s.
Minimum IR Values (Clause 12.1)
IEEE 43-2013 also specifies minimum insulation resistance (at the 1-minute mark, corrected to 40°C):
| Winding Type | Minimum IR at 40°C |
|---|---|
| Form-wound coils (manufactured after 1970) | 100 MΩ |
| Random-wound stator coils and form-wound coils rated below 1 kV | 5 MΩ |
| Windings manufactured before 1970 | (kV + 1) MΩ |
| DC armature windings and AC windings not in the above categories | 5 MΩ |
Both the PI and the IR must meet minimums for the machine to be considered suitable for operation or overvoltage testing. Meeting one but not the other is insufficient.
Temperature Correction (Clause 8)
IEEE 43-2013 uses 40°C as the reference temperature for insulation resistance. Readings taken at other temperatures must be corrected using the standard’s correction factors.
The general approximation: insulation resistance halves for every 10°C rise in temperature.
Correction formula: R₄₀ = R_measured × 2^((40 − T) / 10)
Where T is the actual winding temperature in °C.
Important: The PI itself does not require temperature correction. Because both readings are taken at the same temperature within a 10-minute window, the ratio cancels out temperature effects. This is one of the PI test’s key advantages per the standard (Clause 12.2).
However, IEEE 43-2013 notes that if the winding temperature exceeds 40°C, the PI may be unreliable. In that case, let the winding cool before testing.
When PI Is Not Valid (Clause 12.2.2)
IEEE 43-2013 states clearly:
If the 1-minute IR reading (corrected to 40°C) exceeds 5,000 MΩ (5 GΩ), the calculated PI may not be meaningful and can be disregarded.
The reason: at extremely high resistance, the leakage current is so small (nanoamps) that instrument accuracy, humidity, voltage instability, and test lead positioning can all cause measurement noise that distorts the ratio. The insulation is clearly in excellent condition — the absolute value tells you that. The PI adds no useful information.
Step-by-Step PI Test Procedure
1. Prepare the equipment
De-energize, isolate, lockout/tagout, and verify with a multimeter. Disconnect cables at the motor terminal box. Remove star or delta links to test each winding separately.
2. Discharge stored energy
Short all terminals to ground for at least 60 seconds. Longer for large motors or long cables.
3. Record conditions
Measure and record: ambient temperature, winding temperature (if possible), relative humidity. The PI doesn’t need temperature correction, but you still need these for your records and for the IR value.
4. Connect the megger
- LINE lead → winding terminal
- EARTH lead → motor frame (ground)
- GUARD lead → terminal box insulation surface (if available)
5. Select test voltage
Per IEEE 43-2013 Table 1. For a 400V motor: 500V DC.
6. Apply voltage for 10 minutes
Press and hold the test button for the full 10 minutes. Record the reading at these intervals:
- 30 seconds — for DAR calculation
- 1 minute — this is IR₁ (used for both PI and minimum IR assessment)
- 2 minutes, 3 minutes, 5 minutes — useful for plotting the absorption curve
- 10 minutes — this is IR₁₀ (used for PI calculation)
Most modern digital megohmmeters will record these automatically and calculate PI and DAR for you.
7. Calculate PI
PI = IR₁₀ ÷ IR₁
Example: IR at 1 minute = 150 MΩ, IR at 10 minutes = 450 MΩ
PI = 450 ÷ 150 = 3.0 → Good insulation.
8. Discharge the motor
Short all terminals to ground for at least 40 minutes (4 times the 10-minute test duration per standard practice). Large motors tested at high voltage need this time to safely dissipate absorbed charge.
This is not optional. I’ve measured voltages above 100V on motor terminals more than 5 minutes after a 10-minute PI test. The charge builds up during the test and takes time to bleed off.
9. Test remaining phases
Repeat for each winding. Compare PI values across phases — they should be similar. A significantly lower PI on one phase points to localized contamination or damage.
10. Document and compare
Record all readings. Compare PI values to the IEEE 43-2013 minimums and to your historical data. Plot the trend.
How to Interpret PI Results
PI interpretation table
| PI Value | Condition | Action |
|---|---|---|
| Below 1.0 | Dangerous | Do not operate. Severe moisture, contamination, or insulation failure. Investigate immediately. |
| 1.0 – 1.5 | Poor | Do not operate (unless Class A insulation where 1.5 is the minimum). Investigate cause — likely moisture or contamination. |
| 1.5 – 2.0 | Questionable | Acceptable only for Class A insulation. For Class B/F/H, investigate further. Schedule cleaning or drying. |
| 2.0 – 4.0 | Good | Insulation is in healthy condition. Continue regular testing program. |
| 4.0 – 8.0 | Excellent | Clean, dry insulation in very good condition. |
| Above 8.0 | Investigate | Per IEEE 43-2013: For varnished cambric, shellac mica-folium, or asphaltic windings, a PI greater than 8 may indicate thermally aged, dry, brittle insulation at risk of mechanical failure. Physical inspection recommended. |
The PI > 8 warning
This catches many people off guard. An extremely high PI sounds like great insulation. But per IEEE 43-2013, on older insulation types (varnished cambric, shellac-based), a PI above 8 can mean the insulation has dried out and become brittle. It still resists current — but it can crack under vibration, thermal cycling, or mechanical stress.
If you see a PI above 8 on an older motor, tap the insulation gently. If it feels hard and brittle, or if flakes come off, the insulation is at risk of mechanical failure even though the electrical readings look perfect.
This doesn’t apply to modern epoxy-mica insulation systems (Class F and H), where very high PI values simply indicate excellent condition.
When the PI Can Mislead You
The PI is an excellent test, but it has limitations. Knowing them prevents misinterpretation.
Very high IR values (above 5 GΩ)
Per IEEE 43-2013 Clause 12.2.2, if the 1-minute reading exceeds 5,000 MΩ, the PI may not be meaningful. The currents are so tiny that measurement noise dominates the ratio. Don’t worry about the PI in this case — the insulation is clearly in excellent shape.
Winding temperature above 40°C
High temperatures increase leakage current and flatten the absorption curve. A motor tested hot will show a lower PI than the same motor tested cold. IEEE 43-2013 recommends testing below 40°C for reliable PI results.
Low-capacitance equipment
The PI relies on absorption current, which depends on the insulation’s capacitance. Equipment with very low capacitance — short cable runs, small motors, switchgear — may show a flat PI even when the insulation is fine. For these, a spot reading is more appropriate.
Stress control coatings
Some high-voltage stator windings have semiconductive stress-control coatings at the slot exits. These coatings can cause additional current that masks the absorption behavior and gives misleadingly low PI values. IEEE 43-2013 discusses this in the context of current characteristics (Clause 5).
PI vs DAR: Which One to Use
The dielectric absorption ratio (DAR) is a shorter version of the PI:
DAR = IR at 60 seconds ÷ IR at 30 seconds
| Feature | PI Test | DAR Test |
|---|---|---|
| Duration | 10 minutes | 60 seconds |
| Formula | IR₁₀ / IR₁ | IR₆₀ / IR₃₀ |
| Temperature independent | Yes | Mostly (short duration helps) |
| Best for | Large motors, generators, critical equipment | Smaller motors, quick checks |
| Standard reference | IEEE 43-2013 Table 3 | Not in IEEE 43 main tables; widely used in practice |
| Depth of diagnosis | Higher — absorption has more time to develop | Lower — may miss slow-developing absorption issues |
My rule of thumb:
- Motor above 100 kW or in critical service → PI test
- Motor below 100 kW, routine check → DAR or spot reading
- Any motor with concerning spot readings → PI test regardless of size
The DAR is faster but less conclusive. If you have time for 10 minutes, always choose the PI.
When to Run a PI Test
Always recommended:
- Motors above 100 kW
- Generators of any size
- Medium and high-voltage machines (above 1 kV)
- Critical equipment where failure causes significant downtime or safety risk
- Baseline testing on new or rewound motors
- When a spot reading shows a concerning decline from previous values
Not necessary:
- Small motors in non-critical service where spot readings are stable and high
- Equipment with very high IR (above 5 GΩ at 1 minute) — the PI won’t add information
- Very low-capacitance equipment where absorption is minimal
Recommended testing intervals
| Equipment | PI Test Frequency |
|---|---|
| Critical motors and generators | Every 6 months |
| Medium-voltage motors (1 kV+) | Every 6–12 months |
| Large standard motors (100 kW+) | Annually |
| After extended shutdown (30+ days in humid conditions) | Before re-energizing |
| After a VFD ground fault trip | Immediately |
| After rewind or major repair | Before returning to service (new baseline) |
Common Mistakes
Stopping the test early. The PI requires a full 10 minutes. Stopping at 8 or 9 minutes gives an inaccurate ratio because absorption hasn’t fully developed. Be patient.
Testing hot windings. A motor that was just running will be well above 40°C. The PI will be unreliable. Let it cool, or at minimum record the winding temperature.
Not discharging long enough after the test. A 10-minute test stores significant charge in the insulation. Discharge for at least 40 minutes (4× test duration). I’ve seen technicians get shocked because they discharged for “only” 5 minutes.
Comparing PI across different insulation classes. A PI of 1.7 is acceptable for Class A (minimum 1.5) but failing for Class F (minimum 2.0). Always check the motor’s insulation class before interpreting.
Ignoring the IR₁ value. The PI can be 3.0 (good), but if the 1-minute IR is only 2 MΩ, the insulation still has a problem. IEEE 43-2013 requires both PI and IR to meet minimums. A good ratio on a bad absolute value is not a pass.
Trusting PI on very high IR readings. If IR₁ is above 5 GΩ, the PI ratio may be distorted by measurement noise. The standard says to disregard it in this case.
FAQ
What is a good polarization index?
Per IEEE 43-2013, the minimum recommended PI is 1.5 for Class A insulation and 2.0 for Class B, F, and H insulation. A PI between 2.0 and 4.0 indicates good insulation. Above 4.0 is excellent. Above 8.0 on older insulation types should be investigated for brittle insulation.
Does the PI test need temperature correction?
No. Because the PI is a ratio of two readings taken at the same temperature within a 10-minute window, temperature effects cancel out. However, IEEE 43-2013 recommends testing below 40°C for the most reliable results. The 1-minute IR value does need temperature correction if you’re comparing it to the minimum values in the standard.
How long does a PI test take?
The test itself takes 10 minutes. With setup, discharge, and documentation, allow about 20–25 minutes per motor winding. If you’re testing all three phases of a three-phase motor, budget about 1 hour total.
Can I use PI on transformers?
IEEE 43-2013 applies specifically to rotating machinery, not transformers. For transformers, the applicable standard is IEEE C57.152-2013 (Diagnostic Field Testing of Fluid-Filled Power Transformers, Regulators, and Reactors), which has its own PI interpretation criteria. Transformer PI values are typically lower than motor values because transformer oil has a polarization index close to 1.0.
What does a PI of exactly 1.0 mean?
A PI of 1.0 means the 10-minute reading equals the 1-minute reading — no absorption occurred. This strongly indicates moisture saturation, heavy contamination, or severely degraded insulation. Do not operate the machine. Investigate the cause immediately.
Key Takeaways
- The PI is the ratio of the 10-minute IR reading divided by the 1-minute IR reading.
- Per IEEE 43-2013: minimum PI is 1.5 for Class A and 2.0 for Class B, F, and H insulation.
- The PI does not require temperature correction — it’s a ratio taken under the same conditions.
- If IR₁ exceeds 5,000 MΩ (5 GΩ), the PI may not be meaningful per IEEE 43-2013 Clause 12.2.2.
- A PI above 8 on older insulation types (varnished cambric, asphaltic) may indicate brittle, thermally aged insulation.
- Both the PI and the absolute IR value must meet minimums for the machine to pass.
- Minimum IR for form-wound coils (post-1970) is 100 MΩ at 40°C. For random-wound and form-wound below 1 kV, it’s 5 MΩ.
- Always discharge for at least 4 times the test duration after a PI test.
- The PI is the best diagnostic tool for large rotating machinery — but know its limitations.
Standards Referenced in This Article
| Standard | Title | Key Relevance |
|---|---|---|
| IEEE 43-2013 | Recommended Practice for Testing Insulation Resistance of Electric Machinery | Primary reference. Test voltages (Table 1), minimum PI values, minimum IR values, interpretation guidance. |
| IEC 60085 | Electrical insulation — Thermal evaluation and designation | Defines thermal classes (A, B, F, H) referenced by IEEE 43 for PI minimums. |
| IEEE C57.152-2013 | Diagnostic Field Testing of Fluid-Filled Power Transformers | PI interpretation for transformers (separate criteria from IEEE 43). |
| NETA MTS | Standard for Maintenance Testing Specifications | Provides acceptance and maintenance testing criteria consistent with IEEE 43 values. |