If you’ve been doing motor insulation testing for any length of time, you’ve heard the standard cited a hundred times. “IEEE 43 says the PI minimum is 2.0.” “IEEE 43 requires correction to 40°C.” Most techs and engineers work from the rules without ever opening the document.
That works fine until it doesn’t. Until an auditor asks where the 100 MΩ figure comes from. Until you’re writing an acceptance spec and need to know whether a 1.6 PI is a fail or a marginal. Until someone tells you the 4.0 PI rule from their old manual is still valid (it isn’t, and it never was an IEEE 43 number).
This article walks the actual standard, clause by clause. Not the procedure — there are plenty of how-to guides. This is the reference. What each section says, what it means, and where people get it wrong.
The current version is IEEE 43-2013, approved 11 December 2013, published 6 March 2014. It superseded IEEE 43-2000.
Table of Contents
A Quick History of IEEE 43
The standard didn’t start as IEEE 43. The original was published in 1950 by the AIEE (American Institute of Electrical Engineers — the predecessor body that merged into IEEE in 1963). It was a guide, not a recommended practice, and it covered insulation resistance testing in general terms.
Revisions followed in 1961 and 1974. Both kept the original framework: thermoplastic insulation systems, asphaltic-mica being the dominant material, and PI thresholds built around those materials’ behavior.
The 1970s changed everything. Thermosetting epoxy and polyester systems replaced asphaltic-mica in new machine production. The new insulation behaved differently under DC test. Absorption currents decayed faster. PI values came out higher even on healthy windings. The old thresholds didn’t fit.
That triggered a substantial revision. IEEE 43-2000 was the first version to clearly separate thermoplastic from thermosetting treatment. The 2013 revision built on that, adding new minimum values, clarifying the PI test’s limitations, and bringing in newer findings on stress-grading effects and continuous voltage stress control systems.
The 2013 edition also removed material that no longer belonged. Dry-out procedures, which had been in Annex A of older editions, were dropped — maintenance procedures are outside the standard’s scope now.
Scope and What the Standard Doesn’t Cover
Clause 1.1 sets the boundary. IEEE 43-2013 covers armature and field windings in rotating machines rated 750 W or greater. That includes synchronous machines, induction machines, DC machines, and synchronous condensers. Fractional-horsepower machines are explicitly out.
What’s also out, though the standard doesn’t list these as exclusions:
- Partial discharge testing (covered by IEEE 1434)
- Surge comparison testing (manufacturer practices, no single IEEE standard)
- AC hipot testing (IEEE 95 covers DC hipot; AC hipot is in IEEE 4 and machine-specific standards)
- Dielectric dissipation factor / tan delta (separate test category)
- Capacitance measurements as a standalone test
People sometimes cite IEEE 43 for tests it doesn’t actually address. The standard is narrow: DC insulation resistance and polarization index. That’s it.
Clause 2: Normative References
This is the list of documents IEEE 43 leans on. Worth knowing what they are because they get cited downstream:
- ASTM F855 — Temporary protective grounds. Safety-related.
- IEC 60085 — Thermal classification of insulation. This is where Class 105 (A) and Class 130 (B) come from.
- IEEE Std 1 — Temperature limits for electrical equipment.
- IEEE Std 56 — Insulation maintenance for large AC machinery (10,000 kVA and larger).
- IEEE Std 62.2 — Diagnostic field testing of electric machinery.
- IEEE Std 67 — Operation and maintenance of turbine generators.
- IEEE Std 95 — DC hipot testing (2300 V and above). This is the standard you go to after you’ve passed IR/PI and want to do an overvoltage test.
- IEEE Std 510 — Safety in high voltage testing.
- NEMA MG-1 — Motors and generators. The general construction and rating standard.
If you’re writing a spec that references IEEE 43, these are the standards your spec implicitly leans on too.
Clause 3: Definitions That Matter
The definitions section is short but important. Get these wrong and the rest of the standard falls apart.
Insulation resistance (IRt). The quotient of applied DC voltage of negative polarity divided by current across the insulation, corrected to 40°C, taken at time t from start of voltage application. Note three things buried here: negative polarity is the default, 40°C correction is part of the definition, and time matters. The subscript convention is also formalized: values 1 through 10 are minutes, values 15 and greater are seconds. So IR1 is the 1-minute reading. IR60 is the 60-second reading. Same number, different things.
Polarization index (PI). The ratio of IR at t2 divided by IR at t1. If t1 and t2 aren’t specified, they default to 1 minute and 10 minutes. Same subscript convention applies. PI60/15 means IR at 60 seconds divided by IR at 15 seconds.
Four current components. The standard names four currents that flow when you apply DC to a winding:
- Geometric capacitive current (IC) — short-duration, exponential decay, controlled by winding capacitance and instrument resistance.
- Absorption current (IA) — molecular polarization and electron drift. Decays slowly. This is the current that makes PI work.
- Conduction current (IG) — steady-state current through the bulk insulation.
- Surface leakage current (IL) — steady-state current over insulation surfaces (endwindings, brush rigging areas).
Understanding which of these dominates at which time is the whole game in interpreting IR results.
Electroendosmosis effect. A phenomenon you mostly see on older asphaltic windings with moisture present. Reversing the polarity of the test leads gives different IR readings. Positive polarity (positive lead on winding, negative to ground) typically reads higher than negative polarity on wet old windings. This is why the standard specifies negative polarity for the winding — keeps results consistent.
Insulation resistance profile (IRP). New in the 2013 edition. IR plotted in discrete time increments (like every 5 seconds) over the test period. More on this in Annex D.
Clause 4: Safety
Read this clause carefully. It’s not just liability boilerplate.
Key safety requirements that matter operationally:
Discharge before testing. The winding must be discharged before testing starts. The standard says “almost zero” discharge current and no discernible return voltage (less than approximately 20 V) after ground is removed. The winding should not be left ungrounded between tests.
Discharge after testing. Through a suitable resistor, sized to limit the instantaneous current. Minimum discharge time equals four times the voltage application duration. So a 10-minute PI test needs at least 40 minutes of grounded discharge. This catches people out — they finish the test, see zero on the meter, disconnect, and don’t realize the absorbed charge is still bleeding out.
Lead handling. The lead between the test set and the winding must be insulated and spaced from ground. Otherwise surface leakage along the lead itself contaminates your measurement. Shielding is acceptable.
Surge reflection mitigation. If accessible, the phase neutral and line ends of each winding should be connected together during testing. If the winding fails under test, this minimizes high-voltage surge reflections.
The standard is explicit that the listed measures aren’t exhaustive. Personal protective equipment, hot sticks, insulated ladders, restricted access — all called out, but the responsibility for full hazard assessment is on the test personnel.
Clause 5: The General Theory
This is the longest theoretical section and the hardest to read on first pass. It matters because everything in Clauses 11 and 12 (interpretation and minimum values) only makes sense if you understand what’s actually happening to the current.
5.1 The Four Currents in Detail
When you apply DC to a stator winding, you get the four currents named in Clause 3 flowing simultaneously. The total current IT is their sum:
IT = IC + IA + IG + IL
What dominates depends on time and on the condition of the insulation.
In the first few seconds, IC dominates. The capacitive charging current is large but short-lived. This is why IR readings taken before about 30 seconds are unreliable — the capacitive current is still settling.
From roughly 30 seconds out to 10 minutes and beyond, IA dominates in a clean dry winding. Absorption current is described empirically as:
IA = K × t^(-n)
where K is a function of applied voltage, capacitance, and the insulation system, t is time, and n is an exponent characteristic of the insulation. On a log-log plot, this gives a straight line with negative slope. That’s the curve PI is measuring.
At very long times (much greater than 10 minutes), IA drops low enough that IG and IL start to dominate. These are the steady-state currents. In a healthy winding they’re small. In a contaminated or wet winding, they’re large from the start and never let IA dominate.
The standard notes something important about the equivalent circuit (Figure 1 in the standard): you can’t model a stator bar with a simple RC lumped circuit. The interfacial polarization mechanism within the insulation has a distribution of relaxation times. There are multiple dissimilar interfaces — mica to epoxy, semi-conducting shield to insulation, and so on — each with its own time constant. So real-world IR curves aren’t clean exponentials. They’re sums of many exponentials with different time constants.
5.2 Characteristics of the Measured Current
The practical takeaway: comparing how IT changes over the test duration tells you about cleanliness and dryness.
- Clean and dry winding: IT decreases noticeably with time. Absorption current dominates.
- Contaminated or wet winding: IT stays roughly constant. Surface leakage and/or conduction current is larger than absorption current and swamps the time-dependent behavior.
This is the qualitative basis for the PI test. If IT decreases over the test, PI is high. If IT is flat, PI is close to 1.0.
5.3 Insulation Resistance Readings
Two operational points from this section:
Test voltage must match the winding rating. The standard’s Table 1 gives the guidelines:
| Winding rated voltage (V) | IR test voltage (V) |
|---|---|
| < 1000 | 500 |
| 1000–2500 | 500–1000 |
| 2501–5000 | 1000–2500 |
| 5001–12,000 | 2500–5000 |
| > 12,000 | 5000–10,000 |
Rated voltage means line-to-line for three-phase AC machines, line-to-ground for single-phase, and rated DC voltage for DC machines or field windings.
Negative polarity is the default. This is to handle electroendosmosis on older insulation. On modern epoxy-mica it doesn’t matter much, but the standard sets the convention.
The 1-minute reading is the standard IR value. It’s IR1 in the subscript notation.
5.4 Polarization Index Readings
PI is normally IR10 / IR1 — the 10-minute reading divided by the 1-minute reading. The 2013 standard acknowledges that other ratios exist (15s/30s/1min variants) and addresses them in Annex A.
One important note in 5.4: PI may not apply to small random-wound machines. The absorption current decays to negligible levels in seconds on these windings. Running a 10-minute PI test on a small random-wound motor doesn’t give you useful data.
5.5 Discharge Current
When you remove the applied voltage, two discharge currents flow:
- Capacitive discharge — decays nearly instantaneously, set by the discharge resistance.
- Absorption discharge — decays from a high initial value with the same characteristics as the original charging current but opposite polarity. Can take more than 30 minutes depending on insulation type and size.
This is why the 4x discharge time rule from Clause 4 matters. The absorption discharge is the slow one, and if you don’t let it complete, the residual charge contaminates your next measurement.
Clause 6: Factors Affecting Insulation Resistance
This clause is where the standard gets practical about what changes your readings.
6.1 Surface Condition
Surface leakage current IL depends on what’s on the winding surfaces outside the slot. Dust, salts, oil, carbon — all conductive when wet. Large turbine generator rotors and DC machines have lots of exposed creepage surface, so they’re more affected.
The fix is usually mechanical: clean and dry the surfaces. If your low IR is caused by surface contamination, cleaning restores it. If it’s caused by something internal, cleaning won’t help.
Stress-control coatings on endwindings can also increase IL. More on this in 12.2.3.
6.2 Moisture
Moisture has two distinct effects:
Surface film. If winding temperature is at or below the dew point, a film forms on the insulation surface. Drops IR and PI. Worse if surface is also contaminated or if insulation has cracks.
Absorbed moisture. Some insulation systems are hygroscopic — they absorb water into the bulk. The big offenders are older thermoplastic asphaltic-mica, some thermosetting polyester-mica, and shellac mica-folium. Some insulating strips in large turbine generator rotors are particularly bad. Absorbed moisture increases conduction current IG, drops IR substantially, and pushes PI close to 1.0.
The standard makes a hard recommendation here: a machine with low PI and low IR1 should not be subjected to further high-voltage testing. Hipotting a wet winding is asking for a failure.
Test machines before their winding temperature drops below dew point. In-service machines are usually warm enough. Out-of-service machines without space heaters often aren’t.
6.3 Temperature
This is the section that gets cited most often in spec writing, because temperature correction is non-trivial.
The general rule. IR varies inversely on an exponential basis with winding temperature. Higher temperature = lower IR. The reason is that in insulators (unlike metals), thermal energy frees additional charge carriers and reduces resistivity. Warm insulation conducts more.
This affects all current components except the geometric capacitive current.
Correction to 40°C. The standard’s reference temperature is 40°C. The correction formula is:
RC = KT × RT
where RC is the corrected IR (megohms), KT is the temperature correction factor at temperature T, and RT is the measured IR at temperature T.
KT depends on the insulation system. The standard splits insulation into two families:
Thermoplastic (asphaltic systems, pre-1960s):
KT = (0.5)^((40−T)/10)
This gives a factor of 2 every 10°C — IR halves for every 10°C increase, doubles for every 10°C decrease. Steep curve.
Thermosetting (epoxy and polyester, 1960s onward):
For 40°C < T < 85°C: KT = exp[4230 × (1/(T+273) − 1/313)]
For 10°C < T < 40°C: KT = exp[1245 × (1/(T+273) − 1/313)]
Much flatter curve than thermoplastic.
Table 2 in the standard gives the calculated values:
| T (°C) | KT thermoplastic | KT thermosetting |
|---|---|---|
| 10 | 0.125 | 0.7 |
| 20 | 0.25 | 0.8 |
| 30 | 0.5 | 0.9 |
| 40 | 1.0 | 1.0 |
| 50 | 2 | 1.5 |
| 60 | 4 | 2.3 |
| 70 | 8 | 3.3 |
| 80 | 16 | 4.6 |
A few cautions from the standard:
- The equations were developed from clean dry bars. They may not apply to windings affected by moisture or dust.
- Approximations only. Significant errors are possible outside the 10°C to 60°C range.
- Below the dew point, correction is unreliable because moisture condensation effects can’t be predicted. The standard recommends using machine history under similar conditions instead of temperature correction.
PI correction. The standard’s position is that you don’t need to correct PI for temperature, with one caveat. If the winding temperature changes significantly between the 1-minute and 10-minute readings (which happens when you start hot — say after shutdown — and the winding cools during the test), IR can rise from temperature change alone. That makes PI look better than it really is. The fix is to retest at or below 40°C as a check.
6.4 Test Voltage Magnitude
For insulation in good condition, IR is roughly independent of test voltage. Apply 500 V or 5000 V, you get similar IR.
For insulation with problems, IR drops with increasing test voltage. The drop becomes more pronounced at voltages above rated. This is why hipot testing (covered in IEEE 95) is more sensitive than IR testing — it stresses the insulation enough to reveal problems IR can’t see.
6.5 Existing Charge
Residual charge on the winding gives erroneous IR readings. This is the “memory effect” — covered in more depth in Jonscher’s dielectric relaxation work cited in the bibliography.
The practical rule: measure discharge current at the start of the test to confirm the winding is fully discharged. Grounding time after a previous test should be at least four times the previous charge time.
Clauses 7-10: Procedural Requirements
These are shorter and more operational. Quick coverage.
Clause 7: Conditions. Record ambient temperature, relative humidity, dew point, winding temperature, time out of service, test voltage, and connection arrangement. Convert measurements to 40°C for trending. Generators can be tested while rotating in some cases. Water-cooled windings need water removed and circuit dried (or the manufacturer’s specific method if available, with water conductivity below 0.25 µS/cm if water remains).
Clause 8: Winding connections. Test each phase separately when feasible — lets you compare phases against each other. When testing one phase, ground the other two and the core. Testing all phases simultaneously only tests insulation to ground, not phase-to-phase. Disconnect all external equipment (cables, switches, capacitors, surge arresters, VTs) and ground them. Use a common ground.
Clause 9: Methods. Direct measurement via megohmmeter (hand-cranked, battery, or line-powered rectifier), or via resistance bridge. Calculated measurement using voltmeter and microammeter with an external DC supply — line voltage regulation should be 0.1% or better, otherwise fluctuations contaminate the reading. The microammeter has to be protected (highest range or shorted) during initial charging to survive the inrush.
Clause 10: Precautions. Full test voltage should be applied as rapidly as possible and held constant. Protective resistors in test instruments cause voltage drop that affects readings on low-resistance specimens.
Clause 11: Interpretation
This is where the standard gets careful, because IR data alone doesn’t tell you everything.
11.1 Monitoring insulation condition. The most valuable use of IR/PI testing is trending. A single number doesn’t tell you much. The history of readings over years, taken under similar conditions, reveals what’s changing. Compare similar conditions: same temperature, same voltage, same duration, same humidity. Correct to 40°C if temperature varies between tests.
What trends mean:
- Sharp decline in IR1 or PI = surface contamination, moisture, or insulation damage.
- Steady increase in IR1 with age, particularly on thermoplastic insulation = decomposition of bonding materials. The absorption current decreases over time as the asphalt or shellac breaks down.
11.2 Suitability for operation. Without history, you fall back on the minimum values in Clause 12. If readings are below minimums because of dirt or moisture, cleaning and drying may restore them. If below minimums because of damage, don’t operate and don’t hipot test.
For overvoltage testing or operation, both PI and IR at 40°C should exceed the minimums in Clause 12.
There’s a specific note about old thermoplastic insulation: a very high PI (greater than 8) on varnished cambric, shellac mica-folium, or asphaltic windings may indicate thermal aging. The insulation has dried out and become brittle. Counter-intuitively, very high PI is bad on these systems. Physical inspection (tapping the insulation) can confirm. Don’t clean or hipot test if the insulation is dry and brittle — it can fail at any time.
11.3 Limitations. The standard is honest about what IR can’t tell you:
- IR is not directly related to dielectric strength. You can’t predict failure voltage from IR.
- Machines with large endwinding surface area, large or slow-speed machines, round-rotor field windings, or machines with commutators may legitimately read below the recommended minimums. Trending matters more than absolute values here.
- A single reading can’t tell you whether contamination is concentrated or distributed.
- DC tests don’t detect internal voids from improper impregnation, thermal deterioration, or thermal cycling. AC tests are sensitive to these — see Annex B.
- IR testing while at standstill won’t catch problems that only show up under rotation (loose coils, vibration-related endwinding movement).
Clause 12: Recommended Minimum Values
This is the clause people cite most. Here’s what it actually says.
12.2 Polarization Index Minimums
Table 3 in the standard:
| Thermal class | Minimum PI |
|---|---|
| Class 105 (A) | 1.5 |
| Class 130 (B) and above | 2.0 |
That’s it. Two values. There’s no 4.0 threshold. The 4.0 figure you’ll see in some test equipment manuals and older training material is not in IEEE 43-2013.
12.2.1 Non-insulated field windings. PI doesn’t apply to:
- Squirrel-cage induction rotor windings (not insulated from rotor body)
- DC armatures with exposed copper commutators
- Large turbine generator field windings with strip-on-edge construction (exposed copper, isolated only by insulating strips)
- Salient pole motor/generator field windings with exposed copper
The reason: no insulation encapsulation means no absorption current to speak of. The IR curve is nearly flat between 1 and 10 minutes, so PI is close to 1.0 regardless of insulation condition.
For wound induction rotors and salient pole machines with tape-wrapped windings, PI applies — those have fully encapsulated conductors.
12.2.2 The 5000 MΩ rule. If IR1 exceeds 5000 MΩ, PI may be unreliable. At that resistance level, total current can be in the sub-microamp range. Small fluctuations in supply voltage, humidity, or test connections produce large relative changes in measured current. The standard says PI “may or may not” be meaningful above 5000 MΩ — don’t rely on it as an assessment tool.
This is a clause people miss. They get a PI of 1.3 on a clean dry machine with IR1 of 15,000 MΩ and panic. The clause says: don’t.
12.2.3 Continuous voltage stress control systems. On Roebel-bar windings with very short overhangs, the complete overhang may be treated with stress-control material. If the stress-control material contacts the bare copper at bar ends, surface leakage current can exceed absorption current. Total current stays flat. PI comes out close to 1.0. This doesn’t indicate bad insulation — it indicates that the stress-control treatment has changed what the test measures.
The standard notes that this design is for new windings only. Don’t retrofit it as a repair.
12.3 Insulation Resistance Minimums
Table 4 in the standard:
| IR1 min (MΩ at 40°C) | Test specimen |
|---|---|
| kV + 1 | Most windings made before ~1970, all field windings, and others not described below |
| 100 | Most AC windings built after ~1970 (form-wound coils) |
| 5 | Most machines with random-wound stator coils and form-wound coils rated below 1 kV, and DC armatures |
A few notes that always need attention:
- kV is rated line-to-line RMS for three-phase AC machines, line-to-ground for single-phase, rated DC for DC machines or field windings.
- IR1 min is the minimum after 1 minute, on the entire winding (all phases), at 40°C.
- Single-phase tests with other phases grounded read lower than 3× the all-phases value, due to phase-to-phase contributions.
- The minimums may not apply to windings with very large endwinding surface areas or DC armatures with commutators. Use trending instead.
- The minimums don’t apply to “green” windings before global vacuum impregnation.
- The minimums don’t apply when the complete winding overhang is treated with stress-control material (see 12.2.3).
The kV + 1 rule. This is the old rule, dating back to early editions. For a 13.8 kV machine, IR1 min = 13.8 + 1 = 14.8 MΩ. Very low by modern standards. It was set for asphaltic-mica and other thermoplastic insulation systems where IR values were inherently lower. Still applies to all field windings regardless of age.
The 100 MΩ rule. Modern form-wound stators on epoxy-mica and similar thermosetting insulation should read at least 100 MΩ at 40°C after 1 minute, regardless of voltage rating. This is a flat floor, not voltage-dependent. The kV + 1 rule still applies on these too (since 100 > kV + 1 for typical machines), but 100 is the stricter constraint that controls.
The 5 MΩ rule. For small random-wound motors, low-voltage form-wound coils (under 1 kV), and DC armatures, the minimum is 5 MΩ. This isn’t a general motor minimum — it applies specifically to these classes. People sometimes treat 5 MΩ as the “absolute floor” for any motor, which isn’t what the standard says.
Annex A: PI Variants
Informative annex. The standard’s traditional PI uses 1-minute and 10-minute points. On modern thermosetting insulation, absorption current decays to negligible levels in 2-3 minutes. On random-wound machines, it decays in seconds.
So shorter-time PI variants exist. The annex names two:
- PI = IR1 / IR30s
- PI = IR5 / IR1
These let you finish the test faster and reduce the discharge time. The limitations: there’s no standard for which intervals to use, and there’s no agreed pass/fail criteria as there is for the traditional 1/10 ratio.
The standard encourages users to collect short-time data so future revisions can establish pass/fail criteria. As of 2013, the traditional 1/10 ratio is the only one with established thresholds.
Annex B: DC vs AC Testing
Short comparison. The two test methods see different things.
DC testing is dominated by resistance. The total current reflects ρL/A — resistivity, path length, cross-sectional area. Contaminants like dirt, oil, and water have low resistivity, so surface contamination shows up as low IR.
But the primary insulation in form-wound stators is mica. Mica’s resistivity is essentially infinite. A single layer of mica tape blocks DC current entirely. So internal insulation voids — from improper impregnation, thermal deterioration, thermal cycling — don’t show up in DC tests unless the void extends all the way through to create an electrical track.
AC testing is dominated by capacitance. The current reflects εA/d — dielectric permittivity, area, thickness. Voids and water dramatically affect ε, so AC tests are much more sensitive to internal defects.
The annex’s recommendation: use both. DC for surface contamination and moisture absorption. AC for internal insulation problems. Neither alone gives the full picture.
Annex C: Monitoring Charge and Discharge Currents
This annex covers a more advanced technique than basic IR/PI. After the test voltage is removed, the discharge current can be monitored as a function of time.
The discharge current has two components:
- Capacitive discharge: instantaneous decay.
- Absorption discharge: high initial value, decays with same characteristics as the original charging current but opposite polarity.
What discharge current doesn’t include is surface leakage and conduction current — those are zero once the source is removed. This is useful: comparing charge and discharge currents lets you isolate the leakage component.
The annex introduces the normalized discharge resistance RCdis:
RCdis = Uo × C / Idis
where Uo is the applied charge voltage, C is winding capacitance, and Idis is the 1-minute discharge current.
For modern epoxy-mica systems in good condition, RCdis should be above 2000 seconds at room temperature.
The annex also shows how stress-control systems affect normalized resistance. Two similar epoxy-mica bars — one with iron oxide stress-control paint, one with iron oxide paint plus silicon carbide tape — produced very different results. The SiC-tape bar had 5 kV resistance three times lower than the paint-only bar. PI was strongly affected too: the bar with SiC tape gave higher PI than expected because the SiC system altered the current shape. Guarded measurements (excluding the endwinding current) restored more normal values.
The practical point: on machines with continuous stress-control treatment over the endwinding, standard IR and PI tests don’t isolate the bulk insulation behavior from the stress-control system behavior. Discharge current measurement, particularly guarded, gives cleaner data.
Annex D: Insulation Resistance Profiling
Also new in 2013. IRP is IR plotted in discrete time increments (typically 5 seconds) over the full test duration (typically 10 minutes). What you get is a curve, not just two numbers.
The argument for IRP is that it provides useful information beyond standard IR1 and PI, particularly when IR exceeds 5000 MΩ where PI is unreliable.
Requirements: accurate metering, low-ripple power supply, voltage and current measured at each sample point.
As of the 2013 standard, IRP doesn’t have established pass/fail criteria for specific defect types. The standard explicitly says future IEEE standards will need to address this. As of right now (still under IEEE 43-2013, no superseding version), IRP is data collection without standardized interpretation.
What’s Not in IEEE 43
Worth being explicit about this so you don’t try to cite IEEE 43 for things it doesn’t cover.
- Partial discharge testing. Different standard (IEEE 1434).
- Surge testing / surge comparison. No single IEEE standard; manufacturer practices vary.
- AC hipot. IEEE 4 and machine-specific standards.
- DC hipot. IEEE 95.
- Dissipation factor (tan delta). Not in IEEE 43.
- Capacitance measurement. Not addressed as a standalone test.
- Step voltage / ramp voltage testing. Annex C references ramp DC work, but IEEE 95 is the operational standard.
- Drying procedures. Removed from the standard in 2013. Out of scope.
- Repair acceptance criteria. Not in IEEE 43. Use manufacturer or owner specs.
- PD-free voltage levels. Not in IEEE 43.
If your spec needs any of these, you need additional standards beyond IEEE 43.
Practical Use as a Reference
How to actually use IEEE 43-2013 in real work:
Writing a spec. Cite IEEE 43-2013, name the specific clauses for the requirements you’re imposing. Don’t just say “per IEEE 43” — say “per IEEE 43-2013 Clause 12.3, Table 4” for the IR minimum. That makes the requirement unambiguous.
Acceptance testing. Use the Clause 12 minimums as pass/fail. Remember that minimums are minimums, not targets. A machine that just barely passes is a different machine from one that reads 10× the minimum.
Trending. Clause 11.1 is the controlling section. Same temperature, same voltage, same duration, same humidity conditions, same connections. Correct to 40°C using the Clause 6.3 factors.
Investigating a failure. Clause 11.3 limitations matter. Don’t assume a passing IR means healthy insulation. Don’t assume a failing IR means insulation failure — it could be surface contamination, moisture, or stress-control system effects (Clause 12.2.3).
Audit defense. Know the actual minimums. Know which annexes are informative versus normative. Annexes A through D are all informative — they don’t impose requirements. The normative content is in Clauses 1 through 12.
The standard is roughly 26 pages plus annexes. Worth reading cover to cover if you spec, audit, or interpret motor insulation testing. Worth having on the desk for reference even if you don’t.