A NETA-certified test company once failed a 138 kV transformer on polarization index. The IR measurement was healthy. The power factor was fine. The TTR, excitation current, DGA, leakage reactance, and SFRA all came back acceptable. Everything compared cleanly to previous tests. The transformer was returned to service over the contractor’s objection — and continued to run without issue.
The lesson buried in that story is the central point of this article: insulation resistance testing on oil-filled transformers is not the same test as on motors, and the most common interpretation framework — the IEEE 43 motor PI thresholds (1.5 for Class A, 2.0 for Class B and above) — does not directly apply. Megger, the company that effectively invented modern DC insulation testing, states explicitly in its application documentation that “PI testing is not appropriate for oil-filled transformers.” Yet PI is reported on virtually every transformer test sheet, and it’s regularly used as if it were an acceptance criterion.
This is a practical guide to IR testing on transformers — the test as it actually works on oil-paper insulation, the criteria that apply, what PI and DAR genuinely tell you, and where the test fits in the broader diagnostic suite. For the conceptual basis of IR/PI testing and how it works on motor insulation, see the IEEE 43 clause-by-clause article. The motor case is the foundation; the transformer case is the practical reality.
The core point in one sentence: The instrument is the same as for motors; the interpretation framework is not.
Table of Contents
Transformer IR Testing Quick Reference
| Item | Guideline |
|---|---|
| Minimum acceptable IR | (kV + 1) MΩ at 20°C |
| Preferred in-service range | Above 100 MΩ |
| Excellent condition indicator | Above 1,000 MΩ (supported by other diagnostics) |
| PI acceptance threshold | No universal threshold for oil-filled transformers |
| Typical healthy transformer PI | 1.2 to 1.8 |
| Primary field testing standard | IEEE C57.152-2013 |
| Factory test standard | IEEE C57.12.90 / IEC 60076-1 |
| Primary acceptance test for insulation condition | Power Factor / Tan Delta |
| Common test voltages | 500 V to 10 kV DC, matched to winding rating |
| Standard reading times | 60 seconds (IR60) and 600 seconds (for PI calculation) |
These are widely-applied reference values, not universal limits. Actual criteria depend on transformer voltage class, age, insulation system, and the asset owner’s standards.
How IR Testing on Transformers Differs from Motors
The IR test itself — sometimes called the transformer megger test or simply a megger test on transformer windings — is mechanically identical between motors and transformers: apply a DC voltage between conductor and ground, measure the resulting leakage current, calculate the insulation resistance of the transformer winding by Ohm’s law. The instrument is the same — a megohmmeter. The procedure is largely the same.
What differs is the insulation system being tested, and that difference is significant enough to invalidate some of the interpretive framework that works on motors.
Motor insulation is predominantly solid: mica, polyester, epoxy, varnish, paper. The dielectric structure is rigid. When DC voltage is applied, four currents flow simultaneously: capacitive (decays in seconds), absorption (decays slowly over minutes — this is the current that makes PI meaningful), surface leakage (stays roughly constant), and conduction (stays constant). The absorption current on healthy solid insulation takes long enough to decay that the 10-minute reading is substantially higher than the 1-minute reading, producing PI values in the 2-4 range for healthy insulation.
Transformer insulation is a composite: solid components (paper, pressboard, varnish) impregnated with liquid (mineral oil, ester, or silicone). The dielectric behavior is fundamentally different. The absorption current decays much faster on oil-paper systems because the oil itself contributes a different current behavior than solid dielectrics. The 10-minute reading is often only modestly higher than the 1-minute reading. Even on perfectly healthy transformers, PI values commonly fall in the 1.2 to 1.8 range — below the IEEE 43 motor threshold of 2.0.
IEEE C57.152-2013, Clause 7.2.13.4, addresses this directly: transformer PI values are typically lower than motor PI values because transformer oil has a polarization index close to 1.0. The standard doesn’t set a hard PI acceptance threshold for transformers the way IEEE 43 does for motors, precisely because such a threshold would mismatch the actual behavior of oil-paper insulation.
There’s a second physical issue. On very new transformers with excellent insulation, the IR can be so high that both the 1-minute and 10-minute measurements approach the instrument’s range limit. The ratio becomes “infinity over infinity,” which mathematically resolves toward 1.0 — making PI appear poor on what is actually the cleanest possible insulation. Modern digital test sets handle this better than older analog meggers, but the principle stands: PI loses meaning when the absolute resistance values are extremely high.
The practical consequence: PI on transformers is not a primary acceptance criterion the way it is on motors. It’s supplementary data — useful for trending and for catching specific degradation patterns, but never the basis for failing a transformer in isolation.
Why Motor PI Rules Fail on Transformers
| Insulation System | Typical Healthy PI |
|---|---|
| Large motors | 2.0–4.0 |
| Dry-type transformers | Often above 2.0 |
| Oil-filled transformers | Often 1.2–1.8 |
| New oil-filled transformers | Sometimes near 1.0 |
The reason: oil contributes a polarization index close to 1.0, which pulls the overall PI ratio down even when the cellulose component is in excellent condition. The IEEE 43 motor framework was built around solid insulation systems and assumes absorption-current behavior that oil-paper composites don’t exhibit in the same way.
What About Dry-Type Transformers?
Dry-type transformers — cast resin, vacuum-pressure-impregnated (VPI), or open-ventilated designs — behave more like motors than oil-filled transformers because their insulation system is predominantly solid. The implications:
- PI values are often higher, frequently above 2.0 on healthy units
- IEEE 43-style PI interpretation is more directly relevant
- Moisture affects results differently — surface and barrier effects dominate rather than absorption into cellulose
- Manufacturer-specific guidance becomes particularly important, because dry-type insulation systems vary widely
This article focuses on liquid-filled transformers governed by IEEE C57.152-2013. For dry-type units, the motor framework (IEEE 43) is a closer starting point, with manufacturer recommendations layered on top.
The Standards That Apply
Three documents govern transformer IR testing:
IEEE C57.152-2013 — Guide for Diagnostic Field Testing of Fluid-Filled Power Transformers, Regulators, and Reactors. The primary field testing standard for fluid-filled transformers in North American practice. Covers IR, PI, tan delta, winding resistance, TTR, excitation current, and the broader diagnostic suite. This is the standard to cite for in-service IR testing on transformers, not IEEE 43.
IEEE C57.12.90 — Standard Test Code for Liquid-Immersed Distribution, Power, and Regulating Transformers. Covers factory test procedures including IR. The 2010 revision is notable for the same reason it matters for power factor testing: generic temperature correction tables were removed because they don’t reliably fit all transformers.
IEC 60076-1 — Power Transformers — Part 1: General. The IEC equivalent for routine factory tests, including IR.
The IEEE C57.152-2013 framework is what most field testing programs follow, with manufacturer-specific guidance overlaid for individual units. Treat the values in this article as widely-applied reference points, not universal acceptance limits — actual criteria depend on the transformer’s voltage class, age, insulation system, and the asset owner’s specific standards.
The Test Itself
The setup is straightforward:
Test voltage selection. Matched to the winding rating. IEEE and IEC guidance converge on roughly:
| Winding rated voltage | Recommended DC test voltage |
|---|---|
| Below 1 kV | 500 V |
| 1 to 2.5 kV | 1,000 V |
| 2.5 to 5 kV | 2,500 V |
| 5 to 15 kV | 5,000 V |
| Above 15 kV | 5,000 to 10,000 V |
Lower voltages on lower-rated windings to avoid over-stressing the insulation; higher voltages on higher-rated windings to get meaningful signal across the much larger insulation system. Some test sets offer 10 kV for power transformers; this is the upper end and is appropriate for HV windings of large units.
Winding connections. Every accessible terminal of the winding under test is connected together — bushed terminals shorted across each other at the test set side. The winding behaves as a single conductive node, eliminating any voltage division between turns during the test.
Windings not under test are short-circuited to themselves and grounded. This sets a definite voltage reference and prevents capacitive coupling from producing misleading readings on the measured winding.
Tank and core grounded. Always. The tank is one electrode of every insulation path being measured.
Test combinations. A complete IR test on a two-winding transformer typically includes three measurements:
- HV-to-ground (LV winding grounded along with tank/core)
- LV-to-ground (HV winding grounded along with tank/core)
- HV-to-LV (HV winding energized, LV winding connected to measurement input return, tank/core grounded)
The HV-to-LV measurement isolates the inter-winding insulation specifically. Without it, the three-test set tests two paths and leaves the third unexplored.
Reading durations. Standard practice is to read at 60 seconds (the “IR60” value used for absolute acceptance), and at 600 seconds (10 minutes) if PI is being calculated. Some protocols also record 15-second readings (for DAR calculation) and intermediate readings at 30 seconds, 1 minute, 2, 3, 5, and 10 minutes for trending.
Temperature. Record the winding temperature with every reading. As with power factor testing, temperature has a strong and non-uniform effect on the result. Allow thermal equilibrium between top and bottom oil before testing — typically 3-8 hours of de-energized rest on larger units.
Discharge after the test. The transformer’s winding stores significant capacitive energy after the test. Discharge through a suitable resistor before disconnecting test leads. Minimum discharge time should be at least 4 times the test voltage application time — so a 10-minute PI test needs at least 40 minutes of grounded discharge. Skipping this is a real safety issue and a real risk to subsequent measurements.
Acceptance Criteria
The most commonly cited rule of thumb for transformer IR minimums is:
IR ≥ (rated kV + 1) MΩ at 20°C
For an 11 kV winding, that’s 12 MΩ minimum. For a 138 kV winding, 139 MΩ. The formula is simple, widely cited, and treated as a floor — anything below it warrants investigation, anything above it is at least clearing the basic threshold.
More refined practice distinguishes between several quality bands at 20°C:
- Below the (kV + 1) floor: poor condition. Investigate before energization.
- (kV + 1) to ~100 MΩ: acceptable but not impressive. Combine with other diagnostics.
- 100 MΩ to 1000 MΩ: good. Healthy in-service condition.
- Above 1000 MΩ: generally considered excellent for most oil-filled transformers, particularly when supported by good power factor and oil test results. IR alone doesn’t prove excellent insulation condition — extreme IR values can also reflect “infinity over infinity” measurement effects on very new transformers.
These are guidance bands, not pass/fail limits. Healthy modern power transformers commonly read in the gigaohm range, well above any threshold.
What Should the IR Actually Read in Different Conditions?
The bands above apply broadly, but engineers often need a quicker reference for the question “what should I expect for this transformer right now?” The typical pattern for liquid-filled transformers:
| Condition | Typical IR Range |
|---|---|
| New transformer at commissioning | Often several GΩ to tens of GΩ |
| Healthy in-service transformer | Hundreds of MΩ to several GΩ |
| Investigate / borderline | Near the (kV + 1) MΩ threshold, or rapidly declining trend |
| Unacceptable | Below the (kV + 1) MΩ threshold, or continuing rapid decline |
A new transformer reading hundreds of megohms when several gigaohms would be expected is a real concern even though the value passes the absolute threshold. An older transformer reading the same hundreds of megohms with a stable history may be perfectly fine for continued service. The expected reading depends as much on the transformer’s life stage as on the threshold itself.
The (kV + 1) formula traces back to early industry practice and predates modern dry transformers and processing methods. It’s a conservative floor that catches gross problems; it’s not a tight benchmark for healthy units.
A few important qualifications:
These are reference values at 20°C. Temperature correction is required for trending across different test conditions. The standard temperature correction approximation — IR doubles for each 10°C decrease, halves for each 10°C increase — works adequately for moderate temperature differences on oil-paper systems but has the same limitations as power factor correction: the actual temperature behavior depends on insulation condition, and the correction is least reliable on the transformers that need it most. IEEE C57.12.90-2010 removed the generic correction tables for the same reason as for power factor. Test at consistent temperatures when possible; correct cautiously when not.
Acceptance criteria depend on the transformer type and age. A new transformer reading 50 MΩ on a 138 kV winding is alarming. The same value on a 40-year-old unit may be acceptable if it’s stable over years of testing. Trending against the transformer’s own history is more useful than threshold comparison on in-service units.
Never act on a single IR measurement alone. This applies to transformer testing as much as it does to power factor and bushing testing. Surface contamination, weather, instrument noise, or test setup problems can produce false low readings. Confirm any concerning result through a repeat measurement under controlled conditions, supported by power factor / tan delta, DGA, oil quality, and review of the transformer’s history.
PI and DAR: What They Actually Tell You on Transformers
Both PI (10-minute / 1-minute IR ratio) and DAR (1-minute / 30-second or 60-second / 30-second ratio depending on convention) are derivative measurements from the IR test. The transformer PI test, when reported separately, is simply the same measurement carried through the 10-minute reading. Both PI and DAR are intended to characterize the rate at which IR rises during the test — a proxy for the dryness and quality of the insulation.
The IEEE 43 motor thresholds for PI are widely cited:
- PI < 1.0: unacceptable
- 1.0 to 2.0: marginal
- ≥ 2.0: good
These thresholds do not directly apply to oil-filled transformers. The motor case relies on the absorption current behavior of solid insulation, which dominates the IR curve over the 1-minute to 10-minute range. On oil-paper systems, absorption current decays faster and the PI ratio is naturally lower.
What PI actually means on a transformer:
PI well below 1.0 (e.g., 0.7). Indicates that IR is decreasing over the test duration. This is a strong warning sign — the transformer is conducting more as voltage is sustained, which happens with severe moisture, contamination, or active discharge. Investigate regardless of the absolute IR value.
PI close to 1.0 (0.9 to 1.1). Common on new, well-dried, well-impregnated transformers where the IR is very high. The “infinity divided by infinity” mathematical artifact dominates. May also indicate severe contamination or saturation effects. Distinguishing requires looking at the absolute IR — a 1.0 PI with IR in the gigaohm range is fine; a 1.0 PI with IR at the (kV+1) threshold is a problem.
PI in the 1.2 to 1.8 range. Normal for many in-service oil-filled transformers. Healthy insulation in the typical range.
PI ≥ 2.0. Generally indicates dry, healthy insulation, but it is not required for acceptable transformer performance and is less common on oil-filled transformers than on motors.
Falling PI on successive tests over years. Often more diagnostic than any single value. Indicates a developing trend — moisture ingress, contamination accumulation, oil degradation.
DAR (often calculated as IR at 60 seconds divided by IR at 30 seconds, or 30 seconds divided by 15 seconds depending on convention) is useful when the absorption current stabilizes quickly. For transformers where the 10-minute PI approaches 1.0 because the absorption is nearly complete by minute 1, the shorter-time DAR captures the early absorption behavior more meaningfully. DAR values for healthy oil-paper insulation typically fall around 1.25 to 1.6.
The practical position on PI and DAR for transformers: useful as supplementary data for trending, useful for catching severe contamination or moisture issues (where PI drops below 1.0), not appropriate as a standalone acceptance criterion. Power factor / tan delta is the more reliable acceptance test for transformer insulation condition.
Environmental Factors That Distort the Test
IR testing on transformers is sensitive to several conditions that, if not controlled, produce misleading results:
Surface moisture and humidity. Avoid testing during rain, fog, condensation, or periods of very high humidity whenever possible. Surface moisture on bushings can significantly reduce the measured insulation resistance — a clean, dry transformer can read an order of magnitude lower IR in humid conditions than the same unit tested in dry weather. Many utilities still test above 75% RH when necessary, but the results should be interpreted with this limitation in mind and confirmed under better conditions when the values are concerning.
Surface contamination on bushings. Salt, dust, and other deposits on bushing porcelain create leakage paths that mimic internal insulation problems. Clean the bushings before testing if they’re visibly contaminated.
Temperature. Already discussed — but worth re-emphasizing: trending across different temperatures requires correction, the standard correction has known limitations, and the most reliable approach is to test at consistent temperatures.
Test lead arrangement. Long test leads or leads bunched together can produce stray capacitance effects that affect the measurement, particularly on very high resistance readings.
Residual charge from previous tests. If a winding resistance test or other DC test was just done, the winding may have residual polarization. Allow time for discharge before starting the IR test.
Variable frequency capability. Some modern test sets can run IR-equivalent measurements at low frequencies (well below power frequency) using DFR techniques. The frequency response over a low-frequency range can be more diagnostic than a single DC measurement and is less affected by instrument calibration issues at very high resistance values.
Diagnostic Patterns
Pulling the interpretation together, what specific patterns reveal:
All three IR measurements (HV-G, LV-G, HV-LV) low and roughly equal. Bulk insulation contamination or moisture — affects all paths similarly. Investigate oil quality, moisture content, and DGA.
HV-G low, LV-G and HV-LV normal. Problem in the HV-to-ground insulation. Could be HV bushing contamination, HV winding-to-tank insulation issue, or oil contamination affecting the HV side specifically.
LV-G low, HV-G and HV-LV normal. Same logic applied to the LV side.
HV-LV low, HV-G and LV-G normal. Inter-winding insulation problem. Pressboard barriers between windings affected by moisture or contamination. This is often the early signature of moisture ingress.
All values acceptable but PI dropping over successive tests. Slow degradation. Moisture accumulation in the cellulose is a common cause. Pair with DFR for direct moisture assessment.
Falling IR with stable PI. Bulk contamination or oil degradation affecting all current components proportionally. Less common pattern; investigate oil quality.
Single low reading inconsistent with history. Confirm before acting. Weather, contamination, or instrument issue are common false-alarm sources.
Very high IR (gigaohm range) with PI close to 1.0. Likely a healthy transformer with the “infinity over infinity” mathematical artifact. Confirm by other diagnostics (PF, oil tests). Do not condemn on PI alone.
Worked Examples: Reading Real Results
The patterns above are easier to apply against concrete numbers. Here are three representative cases from typical field testing on power transformers.
Example 1: Healthy 138 kV Transformer
| Test | Result |
|---|---|
| HV-to-Ground | 5,000 MΩ |
| LV-to-Ground | 4,200 MΩ |
| HV-to-LV | 6,500 MΩ |
| PI (10 min / 1 min) | 1.3 |
| Temperature at test | 22°C |
| Power factor (supporting) | 0.32% |
| Oil moisture (supporting) | 8 ppm |
Interpretation. IR values are excellent, well above the 100 MΩ preferred range. The PI of 1.3 is normal for oil-paper insulation; applying motor-style PI thresholds would wrongly flag this transformer. Supporting power factor and oil moisture results confirm healthy condition. No indication of moisture contamination or insulation degradation. Continue routine monitoring at standard intervals.
Example 2: Moisture-Contaminated 33 kV Distribution Transformer
| Test | Result |
|---|---|
| HV-to-Ground | 85 MΩ |
| LV-to-Ground | 70 MΩ |
| HV-to-LV | 45 MΩ |
| PI (10 min / 1 min) | 0.95 |
| Temperature at test | 18°C |
| Previous IR (3 years prior) | 220 MΩ / 180 MΩ / 165 MΩ |
| Power factor (supporting) | 1.2% |
| Oil moisture (supporting) | 35 ppm |
Interpretation. All IR values have dropped substantially from baseline — roughly 60% reduction across the board. The HV-LV path is the most degraded, pointing at inter-winding moisture in the pressboard barriers. PI below 1.0 means IR is decreasing during the test, which combined with elevated power factor and high oil moisture confirms moisture contamination. The transformer is still above the (kV + 1) = 34 MΩ minimum, but the trend and supporting tests indicate active degradation. Action: oil processing / drying recommended before continued service.
Example 3: New 230 kV Transformer at Commissioning
| Test | Result |
|---|---|
| HV-to-Ground | 50,000 MΩ |
| LV-to-Ground | 38,000 MΩ |
| HV-to-LV | 65,000 MΩ |
| PI (10 min / 1 min) | 1.05 |
| Temperature at test | 25°C |
| Power factor (acceptance) | 0.28% |
| Oil moisture | 6 ppm |
Interpretation. Extremely high IR values typical of a well-processed new transformer. PI close to 1.0 is the expected “infinity over infinity” mathematical artifact — both measurements are at the upper end of the instrument’s useful range. Applying the IEEE 43 motor PI threshold of 2.0 would falsely fail this transformer. Power factor and oil moisture confirm excellent condition. PI is meaningless here; the absolute IR values and supporting diagnostics are what matter. Accept and record as the baseline for future trending.
Common Transformer IR Testing Mistakes
The patterns above show how the test works when done right. These are the mistakes that produce false conclusions when done wrong.
Applying IEEE 43 motor PI limits to transformers. The most common mistake. The 1.5 / 2.0 motor thresholds don’t apply to oil-filled transformers because oil-paper insulation has fundamentally different absorption behavior. A transformer with a PI of 1.4 may be perfectly healthy. A transformer with a PI of 1.05 may be a brand-new excellent unit suffering from the mathematical artifact. Never fail a transformer on PI alone, and never apply motor thresholds without considering whether they fit the insulation system.
Testing dirty bushings. Surface contamination on bushings creates leakage paths that can reduce measured IR by an order of magnitude. The result looks like internal moisture or aging when the problem is just dust and salt deposits on porcelain. Clean bushings before testing if they’re visibly contaminated.
Ignoring temperature. IR varies strongly with temperature — roughly doubles for each 10°C decrease, halves for each 10°C increase, with the exact relationship varying by insulation condition. Comparing winter readings to summer readings without correction produces false trends. Always record temperature and apply correction cautiously when comparing across different test conditions.
Measuring only HV-to-ground. A complete IR test runs all three combinations: HV-G, LV-G, and HV-LV. The HV-LV measurement specifically isolates the inter-winding insulation, which is where moisture contamination often shows up first. Skipping it misses the most diagnostically useful comparison.
Failing a transformer based on PI alone. PI is supplementary trending data on transformers, not a standalone acceptance criterion. Power factor / tan delta is the primary test for insulation condition. Any concerning PI result should be confirmed through power factor, DGA, oil quality testing, and review of the transformer’s history before any action is taken.
Skipping the discharge after the test. The capacitive energy stored in a transformer winding after a 10-minute IR test is substantial. Inadequate discharge is a real safety issue and contaminates subsequent measurements with residual polarization. Minimum discharge time is 4× the test duration — so 40 minutes after a 10-minute PI test.
Treating extreme high IR as definitively excellent. Above 1000 MΩ is generally a good sign, but extreme values (tens of thousands of megohms) often reflect instrument range limitations as much as true insulation condition. The “infinity over infinity” effect on PI is a common consequence. Pair extreme IR readings with other diagnostics rather than treating them as guarantees of insulation health.
Single-measurement decisions. A single low IR reading rarely justifies major action. Standard utility practice before any significant intervention is to confirm through a repeat measurement under controlled conditions, supported by power factor / tan delta, DGA, oil quality testing, and DFR/FDS for moisture assessment. Surface contamination, weather effects, instrument issues, and test setup problems can all produce false low readings.
How IR Fits in the Transformer Diagnostic Suite
IR is the simplest, fastest, and most frequently run diagnostic test on transformers. A megger and ten minutes is enough for a complete IR test. It’s also one of the least sensitive — catching gross problems but missing subtle ones.
Where it fits with the other tests:
- IR catches: gross moisture contamination, severe surface degradation, gross winding-to-ground or inter-winding insulation failures, gross deterioration of oil quality affecting bulk resistance.
- IR misses: developing dielectric problems that show in power factor first, mechanical changes (SFRA’s domain), turns-related problems (TTR), tap changer problems (DRM), early-stage moisture or aging that affects power factor before it drops the IR.
A complete diagnostic program uses IR as the first-line screening test — quick, cheap, runs at every maintenance visit — and follows up on any concerning IR result with more sensitive tests (PF, DFR, DGA, oil quality) to localize and characterize the problem.
For acceptance and trending, the relative priority on modern transformers:
- Power factor / tan delta — primary acceptance test, most sensitive to dielectric degradation
- DGA — primary monitoring tool for active faults
- DFR/FDS — moisture-specific diagnostic when needed
- IR — screening test, baseline tracking, catches gross problems
- PI/DAR — supplementary information from the IR test, useful for trending, not for acceptance
This is a different prioritization from motor testing, where PI is genuinely a primary acceptance criterion. The shift reflects the different insulation systems and what each test reveals.
The Takeaway
IR testing on transformers is conceptually identical to IR testing on motors but interpretively different. The motor framework — particularly the IEEE 43 PI thresholds — does not directly transfer to oil-filled transformers, and applying it can produce false failures on healthy units. IEEE C57.152-2013 is the appropriate field test reference for transformers, and the (kV + 1) MΩ rule of thumb is the widely-used minimum threshold.
The discipline that makes IR testing diagnostic on transformers: run all three test combinations (HV-G, LV-G, HV-LV) rather than just one; record temperature and apply correction cautiously; treat PI as supplementary trending data, not as an acceptance criterion; confirm any concerning result through repeat testing and supporting diagnostics before acting; and place IR in the diagnostic hierarchy as a screening test, not the primary tool for transformer condition assessment.
Done well, IR testing catches gross moisture and contamination problems early and provides a consistent trending baseline for in-service transformers. Done with motor-style PI interpretation applied to oil-paper insulation, it produces failures on healthy transformers and gives false confidence on transformers that have real problems other tests would catch. The physical test is straightforward — the discipline lies in remembering which insulation system you’re measuring and reading the result against the right framework.