You test a motor at 20°C and read 200 MΩ. Looks good. You test the same motor three months later at 45°C and read 40 MΩ. Has the insulation degraded? Or is it just the temperature?
Without correction, you can’t tell. A 200 MΩ reading at 20°C and a 40 MΩ reading at 45°C can represent exactly the same insulation condition. Temperature changes everything — and if you’re trending insulation resistance over time without correcting for temperature, your trend is meaningless.
This guide covers the correction formulas from IEEE 43-2013, the differences between thermoplastic and thermosetting insulation, worked examples, and the practical shortcuts that save time in the field.
Table of Contents
Why Temperature Affects Insulation Resistance
Insulation resistance decreases as temperature rises. This happens because higher temperature increases the mobility of charge carriers within the insulation material — ions, electrons, and polar molecules move more freely, allowing more leakage current to flow.
The effect is significant. A typical motor winding tested at 20°C might read 200 MΩ. The same winding at 50°C might read only 25 MΩ — even though nothing has changed about the insulation’s actual condition.
This creates a practical problem: if you test a motor after it’s been running (winding temperature maybe 60°C) and compare that reading to a previous test taken when the motor was cold (20°C), the hot reading will always look worse. Without correction, you’ll think the insulation is degrading when it’s actually fine.
The Basic Rule: 2× per 10°C
The most widely used approximation:
Insulation resistance approximately doubles for every 10°C decrease in temperature, and halves for every 10°C increase.
This rule works well enough for quick field estimates. But it’s an approximation — the actual relationship varies by insulation type, and IEEE 43-2013 provides more precise formulas.
The quick formula
R₄₀ = R_measured × 0.5^((40 − T) / 10)
Where:
- R₄₀ = corrected resistance at 40°C (in MΩ)
- R_measured = measured resistance at temperature T (in MΩ)
- T = winding temperature at time of test (in °C)
If T is below 40°C: the correction factor is less than 1.0 — the corrected value is lower than the measured value. This makes sense: the insulation reads higher when cold, and correcting to 40°C (warmer) gives a lower value.
If T is above 40°C: the correction factor is greater than 1.0 — the corrected value is higher than measured. The insulation reads lower when hot, and correcting to 40°C (cooler) gives a higher value.
IEEE 43-2013 Correction Formulas
IEEE 43-2013 provides the definitive temperature correction method for rotating machinery in Clause 6.3 (Clause 8 in the 2000 edition). The standard uses 40°C as the reference temperature.
The general equation (Clause 6.3)
Rc = KT × RT
Where:
- Rc = insulation resistance corrected to 40°C (MΩ)
- KT = temperature coefficient at temperature T
- RT = measured insulation resistance at temperature T (MΩ)
The standard provides two different formulas for KT depending on the insulation type — thermoplastic or thermosetting.
Thermoplastic vs Thermosetting Insulation
IEEE 43-2013 distinguishes between two insulation families because they respond differently to temperature:
Thermoplastic (asphaltic) insulation
Older insulation systems using asphalt, shellac, varnished cambric, or oleoresinous compounds. Found on motors manufactured before the 1970s and some specialty applications. These insulations soften when heated — their resistance drops more steeply with temperature.
Formula (IEEE 43-2013, Clause 6.3.3.1):
KT = 0.5^((40 − T) / 10)
This is the same as the basic 2× per 10°C rule. For thermoplastic insulation, the approximation and the standard formula are identical.
Thermosetting (epoxy/polyester) insulation
Modern insulation systems using epoxy-mica, polyester-mica, or similar materials. Found on most motors manufactured after the 1970s (Class B, F, and H insulation). These insulations don’t soften with heat — their resistance-temperature relationship is slightly different.
Formula (IEEE 43-2013, Clause 6.3.3.2):
KT = 0.5^((40 − T) / C)
Where C is a constant that depends on the specific insulation system. For many thermosetting insulations, C is approximately 7 to 10. When C = 10, the formula is identical to the thermoplastic formula. When C = 7, the resistance changes more rapidly with temperature.
In practice: Most technicians use C = 10 (the 2× per 10°C rule) for both types unless they have specific correction data from the motor manufacturer. The difference between C = 7 and C = 10 is significant at large temperature deviations from 40°C but modest within the normal testing range of 20°C–50°C.
How to know which type you have
Check the motor nameplate for insulation class:
| Insulation Class | Typical System | Formula |
|---|---|---|
| Class A (105°C) | Often thermoplastic on older motors | KT = 0.5^((40−T)/10) |
| Class B (130°C) | Thermosetting (epoxy-mica) | KT = 0.5^((40−T)/10) or C=7 |
| Class F (155°C) | Thermosetting (epoxy-mica) | KT = 0.5^((40−T)/10) or C=7 |
| Class H (180°C) | Thermosetting (silicone/epoxy) | KT = 0.5^((40−T)/10) or C=7 |
If in doubt, use C = 10. It gives a slightly less aggressive correction for thermosetting insulation, which means your corrected value is conservative (slightly higher than actual) — a safe approach.
Worked Examples
Example 1: Cold motor (below 40°C)
Measured: 300 MΩ at 20°C
Correction to 40°C:
- KT = 0.5^((40 − 20) / 10) = 0.5² = 0.25
- Rc = 0.25 × 300 = 75 MΩ
That 300 MΩ reading at 20°C is actually only 75 MΩ when corrected to the 40°C reference. Still well above the 5 MΩ minimum for a random-wound motor — but a lot less impressive than it looked.
Example 2: Hot motor (above 40°C)
Measured: 40 MΩ at 55°C
Correction to 40°C:
- KT = 0.5^((40 − 55) / 10) = 0.5^(−1.5) = 2.83
- Rc = 2.83 × 40 = 113 MΩ
That 40 MΩ at 55°C is actually 113 MΩ at the reference temperature. The insulation is in good shape — the low reading was entirely due to the high temperature.
Example 3: Comparing two readings over time
Test 1 (January): 500 MΩ at 15°C
- KT = 0.5^((40 − 15) / 10) = 0.5^2.5 = 0.177
- Rc = 0.177 × 500 = 88 MΩ
Test 2 (July): 80 MΩ at 35°C
- KT = 0.5^((40 − 35) / 10) = 0.5^0.5 = 0.707
- Rc = 0.707 × 80 = 57 MΩ
Without correction: It looks like the insulation dropped from 500 MΩ to 80 MΩ — an 84% decline. Alarming.
With correction: The actual decline is from 88 MΩ to 57 MΩ — a 35% decline. Still worth monitoring, but not the emergency the raw numbers suggested.
Example 4: When correction reveals a real problem
Test 1 (last year): 150 MΩ at 30°C
- KT = 0.5^((40 − 30) / 10) = 0.5^1 = 0.5
- Rc = 0.5 × 150 = 75 MΩ
Test 2 (this year): 30 MΩ at 25°C
- KT = 0.5^((40 − 25) / 10) = 0.5^1.5 = 0.354
- Rc = 0.354 × 30 = 10.6 MΩ
With correction: The insulation dropped from 75 MΩ to 10.6 MΩ — an 86% decline. This is a real problem, not a temperature artifact. Investigate immediately.
Quick Reference Correction Table
Correction factors (KT) to normalize measured IR to 40°C
| Winding Temp (°C) | KT (multiply measured IR) | What happens |
|---|---|---|
| 10 | 0.125 | Measured IR × 0.125 = corrected to 40°C |
| 15 | 0.177 | Cold windings read high — correction lowers the value |
| 20 | 0.250 | |
| 25 | 0.354 | |
| 30 | 0.500 | |
| 35 | 0.707 | |
| 40 | 1.000 | Reference temperature — no correction needed |
| 45 | 1.414 | Hot windings read low — correction raises the value |
| 50 | 2.000 | |
| 55 | 2.828 | |
| 60 | 4.000 | |
| 65 | 5.657 | |
| 70 | 8.000 | |
| 75 | 11.314 |
Based on KT = 0.5^((40−T)/10), which is the IEEE 43-2013 formula for thermoplastic insulation and the commonly used approximation for all insulation types.
How to use this table
- Measure the winding temperature
- Find the closest temperature in the table
- Multiply your measured IR by the KT value
- The result is the IR corrected to 40°C
- Compare this corrected value to the IEEE 43-2013 minimums
When NOT to Correct
Temperature correction is powerful, but it has limits.
Below the dew point
IEEE 43-2013 states clearly: if the winding temperature is below the dew point, do not attempt temperature correction.
Below the dew point, moisture condenses on the insulation surface. This surface moisture creates leakage paths that are not related to the bulk insulation condition. Correcting a moisture-affected reading to 40°C gives a meaningless result.
If you must test below the dew point, record the reading as-is and note the conditions. Compare only to other readings taken under similar conditions.
PI doesn’t need correction
The polarization index is a ratio of two readings taken at the same temperature within a 10-minute window. Temperature effects cancel out. Do not apply temperature correction to PI values.
However, IEEE 43-2013 notes that if the winding temperature exceeds 40°C, the PI itself may be unreliable. Let the motor cool below 40°C before running a PI test.
Very high IR values (above 5 GΩ)
At extremely high resistance values, the correction formula amplifies measurement noise. If your motor reads 10 GΩ at 20°C, correcting to 40°C gives 2.5 GΩ. The insulation is clearly excellent at either value — the correction adds precision but no practical information.
When temperatures are already near 40°C
If the winding is between 35°C and 45°C, the correction factor is between 0.71 and 1.41. The correction changes the reading by less than 50%, which is within the normal test-to-test variation for most equipment. For routine assessments in this range, you can often compare uncorrected readings directly.
Transformer Temperature Correction (IEEE C57.152)
Transformers use a different reference temperature than motors.
IEEE C57.152-2013 uses 20°C as the reference temperature for transformer insulation resistance measurements.
Correction formula:
R₂₀ = R_measured × 0.5^((20 − T) / 10)
| Oil Temp (°C) | KT (multiply measured IR) |
|---|---|
| 10 | 0.500 |
| 15 | 0.707 |
| 20 | 1.000 |
| 25 | 1.414 |
| 30 | 2.000 |
| 40 | 4.000 |
| 50 | 8.000 |
Critical difference: Motors correct to 40°C. Transformers correct to 20°C. If you apply the motor correction formula to a transformer reading, you’ll get the wrong result.
Common Mistakes
Using the wrong reference temperature. Motors use 40°C (IEEE 43). Transformers use 20°C (IEEE C57.152). Mixing them up gives incorrect corrected values.
Correcting below the dew point. If moisture has condensed on the insulation, the reading is dominated by surface leakage — not bulk insulation resistance. Temperature correction doesn’t account for moisture and gives misleading results.
Applying correction to PI values. PI is already temperature-independent because it’s a ratio. Correcting it adds error, not accuracy.
Forgetting to measure the actual winding temperature. Ambient temperature is not winding temperature. A motor that was just running may have a winding temperature 30–50°C above ambient. Use an RTD, thermocouple, or the motor’s built-in temperature sensor if available. If you can only measure ambient, wait until the motor has cooled to ambient before testing.
Over-relying on the correction at extreme temperatures. The 2× per 10°C rule is an approximation. At large deviations from 40°C (e.g., testing at 0°C or 80°C), the correction factor becomes very large and amplifies any measurement error. If possible, test within 20°C–60°C for the most reliable corrected values.
Not recording the temperature. Every IR reading without a recorded temperature is data you can’t use for trending. Write it down every time.
FAQ
Why do motors use 40°C and transformers use 20°C?
IEEE 43 chose 40°C because it represents a typical winding temperature for a motor that has been idle in a warm industrial environment. IEEE C57.152 chose 20°C because it represents a typical ambient temperature for transformer oil. The choice is largely historical and based on the conditions under which each type of equipment is most commonly tested.
My megger does automatic temperature correction. Can I trust it?
Most modern megohmmeters with automatic correction use the 2× per 10°C rule (C = 10). This is acceptable for most insulation types. Some advanced instruments let you select the insulation type and apply the specific IEEE 43-2013 formula. If your instrument does automatic correction, verify which formula it uses and whether it corrects to 40°C or 20°C.
Do I need to correct if I always test at the same temperature?
If you consistently test at the same winding temperature (within ±5°C), you can compare uncorrected readings directly for trending purposes. But if you need to compare your readings to IEEE 43-2013 minimum values, you must correct to 40°C — the minimums in the standard are specified at 40°C.
How do I measure winding temperature?
In order of accuracy: built-in RTD or thermocouple in the motor (most reliable), infrared thermometer on the motor frame (approximate — frame temperature is close to but not exactly winding temperature), or ambient temperature after the motor has cooled for several hours. For transformers, use the oil temperature gauge.
Does temperature correction apply to cable IR testing?
IEC 60364-6 does not specify a formal temperature correction for cable testing. However, the same physics applies — cable insulation resistance decreases with temperature. If you’re trending cable IR over time, recording the temperature and applying the 2× per 10°C rule gives more meaningful comparisons.
Key Takeaways
- Insulation resistance approximately halves for every 10°C rise in temperature. Without correction, your readings are incomparable across seasons.
- IEEE 43-2013 uses 40°C as the reference for motors. Formula: R₄₀ = R_measured × 0.5^((40−T)/10).
- IEEE C57.152 uses 20°C as the reference for transformers. Don’t mix them up.
- IEEE 43-2013 provides two formulas: one for thermoplastic (asphaltic) and one for thermosetting (epoxy) insulation. For most field work, the 2× per 10°C approximation (C = 10) works for both.
- Do not correct below the dew point — moisture on the surface invalidates the correction.
- PI does not need correction — it’s already temperature-independent.
- Always record the winding temperature alongside every IR reading. Without it, the data is useless for trending.
Standards Referenced in This Article
| Standard | Key Content |
|---|---|
| IEEE 43-2013 | Temperature correction to 40°C (Clause 6.3), thermoplastic formula (Clause 6.3.3.1), thermosetting formula (Clause 6.3.3.2), dew point limitation |
| IEEE C57.152-2013 | Transformer IR correction to 20°C reference temperature |
| IEC 60085 | Thermal class definitions (A, B, F, H) referenced for insulation type identification |