Partial Discharge Testing vs Insulation Resistance: When Each Matters

A motor passes its insulation resistance test with a 500 MΩ reading and a PI of 3.2. Six weeks later it fails catastrophically in service. How? The IR test was measuring the bulk resistance of the insulation system, but somewhere inside the winding there was a tiny void in the epoxy-mica tape. That void had been producing partial discharges for months, slowly carbonizing the insulation around it, until the degradation finally bridged enough of the insulation to cause a ground fault.

Insulation resistance testing couldn’t see that void. Partial discharge testing could.

This article is a deep dive into when each method matters, what they measure, and — critically — why a full insulation diagnostics program uses both. It covers the standards (IEC 60270 for PD measurement, IEEE 43-2013 for IR), the physical phenomena each test detects, and the decision framework for which test to apply to which equipment.

All technical specifications are verified directly against IEC 60270:2000+AMD1:2015 (the primary international standard for PD measurements) and IEEE 43-2013.

Two Tests, Two Different Questions

The fundamental difference between insulation resistance testing and partial discharge testing isn’t the equipment. It’s the question each test answers.

Insulation resistance asks: “Does the insulation currently resist the operating voltage?”

Partial discharge asks: “Is anything in the insulation degrading toward eventual failure?”

These are different questions, and they require different measurements. An insulation system can have high resistance (passing the IR test) while simultaneously experiencing internal discharges that will, given time, destroy it. Conversely, an insulation system with slightly low IR might be completely stable with no PD activity — just surface contamination that’s easily cleaned.

Missing the distinction leads to the two most common insulation management errors:

1. Over-confidence from passing IR tests — “the motor passed its megger test, so we’re fine” — missing developing PD-driven failures
2. Over-reaction to low IR — scrapping or rewinding equipment that has a simple surface contamination problem but no actual insulation degradation

Understanding what each test can and cannot see is the foundation of sound insulation diagnostics.

What Insulation Resistance Actually Measures

Insulation resistance testing (covered in depth in our IR Testing Complete Guide) applies a DC test voltage and measures the leakage current through the insulation. The ratio gives resistance in megohms or gigohms.

What IR detects well

• Surface contamination — moisture, salt, dust, oil films that create low-resistance paths across insulation surfaces
• Bulk moisture absorption — water absorbed into the insulation material itself
• Gross insulation failure — conductor-to-ground faults, severely degraded insulation, water ingress to the interior of the insulation
• Overall “condition” of the insulation system as a single aggregate number

What IR cannot detect

• Void discharges — partial discharges inside tiny air-filled voids in solid insulation
• Corona on energized conductors — discharges in air at the edges of high-voltage components
• Surface tracking in early stages — before it carbonizes enough to significantly lower surface resistance
• Developing failures that haven’t yet created measurable leakage paths
• Localized degradation in equipment with large insulation volumes (the problem is too small relative to the total insulation to affect the aggregate reading)

The key limitation: IR testing measures the average behavior of the entire insulation system. A small defect — even one actively degrading toward failure — may not move the needle on a 1 GΩ reading.

A numerical example

Consider a 13.8 kV form-wound motor stator with ~1 GΩ IR at 40°C. Now imagine a 1 mm³ void in the insulation that’s experiencing partial discharge activity. The void’s effect on the bulk IR is essentially undetectable — the surrounding mica-epoxy maintains its high resistance. But the void is actively carbonizing its boundaries, and the carbonized region will eventually extend, grow, and bridge the insulation. The motor will fail — and the IR test will only show the problem after the failure is imminent, not while there’s time to schedule intervention.

This is not a theoretical concern. Most insulation failures in medium-voltage rotating machinery follow this pattern: long PD activity in a void, progressive degradation, and then rapid final failure that IR testing catches hours before it happens rather than months.

What Partial Discharge Actually Measures

Partial discharge testing detects the brief current pulses that occur each time a localized breakdown happens inside the insulation — without bridging the entire gap between conductors.

What PD detects well

• Void discharges in solid insulation
• Corona from sharp points on high-voltage components
• Surface tracking along insulation surfaces
• Delamination at interfaces (coil-to-groundwall, conductor-to-coil insulation)
• Contamination-induced discharges from conductive particles or moisture
• Insulation aging patterns through long-term PD trending

What PD is less useful for

• Bulk moisture absorption — doesn’t necessarily produce PD activity
• Simple surface contamination — may or may not produce detectable PD
• Installation verification — showing an installation is safe to energize (IR/hipot tests do this better)
• Low-voltage equipment (below ~3 kV) — PD inception voltages are often above operating voltage in LV equipment, so PD doesn’t occur

The “precursor” nature of PD

What makes PD powerful is that it’s a precursor to failure, not a measurement of failure itself. Insulation systems in medium-voltage and high-voltage equipment typically experience PD activity for months or years before the insulation finally fails. By detecting PD activity and trending its magnitude and pattern over time, you can identify developing problems long before they become IR-detectable — often with enough warning to schedule intervention during planned outages.

The Physics of Partial Discharge

Understanding why PD occurs makes the rest of this article much clearer.

Why voids matter

Solid insulation — epoxy, mica, polyethylene, oil-impregnated paper — is designed to withstand the operating voltage gradient (typically kV per mm). When the insulation is uniform, the electric field is evenly distributed, and no discharge occurs.

But real insulation has imperfections. The most common is a gas-filled void — a small pocket of air or gas trapped during manufacturing. The dielectric strength of air is dramatically lower than the surrounding solid insulation (roughly 30× lower for the same distance). When the solid insulation carries, say, 2 kV/mm without issue, the gas in the void might only withstand 0.1 kV/mm.

What happens inside a void

1. Voltage is applied across the insulation
2. The electric field concentrates in the void (lower dielectric constant than surrounding material)
3. When the field exceeds the breakdown strength of the gas in the void, an electrical discharge occurs — a tiny arc
4. The discharge is extinguished quickly because the resulting space charge reduces the field
5. The void recharges on the next voltage half-cycle, and the process repeats

Each discharge is small — picocoulombs to nanocoulombs of charge — but the cumulative effect is destructive. Each discharge:

• Chemically attacks the void boundary (UV radiation, ozone, acids)
• Physically erodes the insulation surface
• Carbonizes the surrounding material over time
• Creates conductive tracks that eventually extend the discharge path

From PD to failure

The progression from initial PD to catastrophic failure follows a typical sequence:

1. Void formation during manufacturing, thermal cycling, or mechanical stress
2. Initial PD activity when voltage is applied — small, possibly intermittent discharges
3. Progressive degradation — chemical attack extends the void, carbonization begins
4. Void growth — what was a 0.1 mm void becomes 1 mm, then larger
5. Coalescence — multiple voids connect into a conductive track
6. Final breakdown — when the track extends far enough, a full arc forms, triggering the protection

This process can take months to decades depending on operating voltage, temperature, and insulation type. PD monitoring provides visibility at every stage.

IEC 60270: The PD Measurement Standard

IEC 60270:2000+AMD1:2015 (“High-voltage test techniques – Partial discharge measurements”) is the primary international standard for PD measurement. Every PD test procedure and every commercial PD instrument refers back to this standard.

What IEC 60270 covers

The standard applies to PD measurements on electrical apparatus tested with AC voltages up to 400 Hz or with DC voltage. It:

• Defines the terms used in PD measurement
• Defines the quantities to be measured
• Describes test and measuring circuits
• Defines analog and digital measurement methods
• Specifies calibration methods and instrument requirements
• Gives guidance on test procedures
• Provides assistance on discriminating PD from external interference

The standard does NOT specify test levels for specific equipment. Individual product standards (IEC 60034-27-2 for rotating machines, IEC 60076-3 for transformers, etc.) specify the actual PD levels and interpretation criteria for each equipment type.

The formal definition of PD

Per IEC 60270 Clause 3.1, a partial discharge is: “localized electrical discharge that only partially bridges the insulation between conductors and which can or cannot occur adjacent to a conductor.”

Two critical notes from the standard:

• Partial discharges typically appear as pulses with duration much less than 1 μs
• “Corona” is a specific form of PD occurring in gaseous media around conductors

The “localized” and “partially bridges” parts are key. A full dielectric breakdown is NOT a partial discharge — it’s a complete failure. PD is the smaller, localized activity that precedes full breakdown.

The Quantities Measured in PD Testing

IEC 60270 defines several quantities used in PD measurement. Understanding them is essential for interpreting test reports.

Apparent charge (q) — the primary PD quantity

The apparent charge q is the measured charge that, when injected instantaneously between the terminals of the test object, would produce the same momentary voltage change as the partial discharge itself.

Important: apparent charge is NOT the actual charge released in the void. The actual charge is usually larger, but it’s trapped inside the insulation and not directly measurable from the terminals. Apparent charge is what the external measurement system sees.

Apparent charge is expressed in picocoulombs (pC). Typical values:

• 1–10 pC: Very low — healthy new equipment, background noise level
• 10–100 pC: Moderate — common in aged but functional MV equipment
• 100–1,000 pC: High — active PD, investigate
• 1,000–10,000 pC (1–10 nC): Very high — significant deterioration
• >10,000 pC: Severe — equipment approaching failure

Interpretation depends heavily on equipment type. A 50 pC reading on a 13.8 kV motor might be normal; on a 400 kV power transformer, it could be cause for concern. Always interpret against equipment-specific criteria.

Pulse repetition rate (n)

The number of PD pulses per second. Along with apparent charge, this characterizes the activity level. Low magnitude + high repetition rate suggests different degradation than high magnitude + low repetition rate.

Average discharge current (I)

The sum of absolute values of apparent charge magnitudes over a chosen reference time interval, divided by that interval. Expressed in coulombs per second (C/s) or amperes (A). Useful for trending total PD activity.

Discharge power (P)

The average power fed into the test object due to PD activity. Calculated from the sum of q_i × u_i (charge magnitude times instantaneous voltage at time of occurrence). This quantity represents the energy being dissipated by PD activity and is useful for predicting remaining insulation life.

Phase angle (φ)

The phase angle of the AC voltage at which each PD pulse occurs. This is enormously diagnostic — different types of PD (void discharges, corona, surface tracking, floating-component discharges) produce characteristically different phase-resolved patterns. Modern PD analyzers display these as phase-resolved partial discharge (PRPD) patterns.

PD inception voltage (PDIV) and extinction voltage (PDEV)

• PDIV: The lowest voltage at which PD activity is first detected as voltage is increased from zero
• PDEV: The voltage at which PD activity stops as voltage is decreased from above PDIV

For healthy equipment, PDIV should be above operating voltage. PD activity at or below operating voltage is a clear indicator of a problem.

Offline vs Online PD Testing

PD testing divides into two fundamentally different approaches.

Offline PD testing

The equipment is de-energized, isolated, and connected to a dedicated test voltage source. PD measurements are made under controlled laboratory-like conditions.

Advantages:

• Low background noise (controlled environment)
• Precise measurement of apparent charge in pC
• Can vary test voltage to determine PDIV and PDEV
• Traceable calibration per IEC 60270
• Best sensitivity for small defects

Disadvantages:

• Requires equipment outage
• Requires specialized test voltage source (for HV testing, kV-rated AC sources)
• Expensive dedicated equipment
• Only shows PD behavior at the test voltage, not at actual operating conditions

Online PD testing

PD activity is monitored while the equipment is energized and in normal operation. Sensors (high-frequency current transformers, capacitive couplers, UHF antennas, or acoustic sensors) detect PD signals from the live equipment.

Advantages:

• No outage required
• Shows actual operating conditions (temperature, mechanical stress, voltage)
• Continuous monitoring possible
• Catches PD that only occurs under specific load or thermal conditions

Disadvantages:

• Higher background noise from the electrical environment
• Quantitative calibration in pC is difficult or impossible
• More susceptible to external interference
• Requires sophisticated signal processing to discriminate PD from noise

The right choice depends on the purpose

Commissioning and acceptance testing → offline, per IEC 60270. Quantitative, traceable, comparable to type test data.

Maintenance monitoring of critical in-service equipment → online. No outage, catches real-world behavior.

Troubleshooting an equipment issue → both may be needed. Online testing identifies that there’s a problem; offline testing in a controlled environment characterizes and quantifies it.

For the detailed frequency specifications per IEC 60270 Clauses 4.3.4–4.3.6:

• Wide-band instruments: f₁ between 30 kHz and 100 kHz, f₂ up to 1 MHz, bandwidth 100 kHz to 900 kHz
• Narrow-band instruments: bandwidth 9 kHz to 30 kHz, midband frequency 50 kHz to 1 MHz
• UHF/VHF instruments (not covered by the 2000 edition): typically 100 MHz to 1 GHz, used for GIS and advanced online monitoring

Where IR Excels and Where PD Excels

Each test has a distinct sweet spot.

IR testing excels at

Commissioning and acceptance testing. Quick, inexpensive, unambiguous go/no-go for new installations.

Low-voltage equipment. Below ~3 kV, PD activity is often absent even in aged equipment (voltage gradient is too low to exceed void breakdown). IR is the right tool.

Moisture detection. Water absorbed into insulation dramatically lowers IR but may not produce detectable PD. IR sees moisture; PD often doesn’t.

Contamination detection. Surface contamination creates leakage paths measurable by IR. Contamination may or may not produce PD depending on geometry.

Fleet screening. Fast IR tests on many motors identify which ones need deeper investigation. PD testing each motor in a 500-motor fleet is impractical.

Post-repair verification. After rewind, the IR test confirms the repair work was adequate before energization.

PD testing excels at

Medium- and high-voltage equipment. Above ~6 kV, PD becomes a primary failure mechanism. Equipment operating at 13.8 kV, 22 kV, 66 kV, 132 kV, 400 kV gets progressively more benefit from PD monitoring.

Detecting developing defects. Voids, delamination, sharp points, surface tracking — all detectable by PD long before IR shows anything.

Online monitoring of critical equipment. Continuous PD monitoring on a main transformer or generator provides real-time health visibility no IR test can match.

Differentiating failure modes. PRPD patterns show whether you have a void discharge, a corona source, or surface tracking — each has different remediation.

Life assessment of aged equipment. Quantifying current PD activity and comparing to historical patterns gives remaining-useful-life estimates that IR trending alone cannot provide.

Factory quality control. Large equipment manufacturers (motor, transformer, GIS) use PD testing to detect manufacturing defects before shipment.

The Complementary Relationship

The most important point in this entire article: IR and PD are complementary, not competing.

A complete insulation diagnostic program uses both. Each answers questions the other cannot.

An example from substation testing

A 66/22 kV power transformer shows the following test results during a scheduled outage:

Test	Result	IR Interpretation	PD Interpretation
IR at 5 kV DC	3,200 MΩ corrected to 20°C	Healthy, above 100 MΩ minimum	—
PI (10 min / 1 min)	2.8	Above 2.0 — healthy	—
Offline PD at 22 kV (rated)	150 pC	—	Slightly elevated — investigate
Offline PD at 33 kV (1.5× rated)	800 pC	—	Active PD — trending required
Tan delta	Within spec	Healthy dielectric losses	—

The IR and PI tests say the transformer is fine. The PD test says there’s developing insulation degradation. Both results are correct — they’re measuring different things. The transformer currently has good bulk insulation resistance AND active PD activity that will degrade that resistance over time.

The right response isn’t to panic or to scrap the transformer. It’s to:

1. Document the PD baseline
2. Schedule follow-up PD measurement in 12 months
3. If the PD increases, plan intervention before it becomes critical
4. Meanwhile, operate the transformer normally with confidence based on the IR/PI results

Without PD testing, this transformer would have looked healthy — and the developing problem would have been invisible until it was much further along.

Another example from motor testing

A 4.16 kV motor shows these results:

Test	Result	Interpretation
IR at 1 kV DC	180 MΩ corrected to 40°C	Above 100 MΩ minimum, but lower than new
PI	1.5	Below 2.0 — concerning
Offline PD at 4.16 kV	25 pC	Very low — no significant PD activity

The IR/PI results suggest developing insulation problems. But the PD is essentially zero. What’s happening?

The low PI with low PD typically indicates moisture or contamination, not insulation degradation. The moisture is lowering the bulk resistance but isn’t producing PD activity (moisture typically doesn’t). The right response is:

1. Dry the motor
2. Clean accessible surfaces
3. Retest

In this case, the motor is probably fine after drying — even though the IR/PI results initially looked alarming. PD testing confirmed there’s no structural degradation, just surface/absorbed moisture.

This is the power of complementary testing.

Cost and Complexity Comparison

A significant practical difference between the two methods.

Insulation resistance testing

• Equipment cost: $500–$8,000 depending on voltage rating
• Time per test: 1–15 minutes depending on equipment and test type
• Operator skill: Basic electrical training; 1 week to become competent
• Interpretation: Straightforward go/no-go against standards
• Setup requirements: Portable, field-deployable, minimal support equipment

Offline PD testing (per IEC 60270)

• Equipment cost: $30,000–$200,000+ for calibrated PD measurement systems; additional $50,000–$500,000 for AC test voltage source for HV testing
• Time per test: 1–4 hours for setup, calibration, and measurement
• Operator skill: Specialized training; months of experience for reliable interpretation
• Interpretation: Requires understanding of PRPD patterns, equipment-specific criteria, and noise discrimination
• Setup requirements: Shielded environment ideal; grounded test setup; often requires partial dismantling

Online PD monitoring

• Sensor cost: $500–$10,000 per measurement point
• Monitoring system: $20,000–$100,000+ for multi-channel continuous monitoring
• Ongoing: Requires data management, analysis infrastructure, ongoing expertise
• Time per result: Continuous after installation
• Operator skill: High — pattern recognition and statistical analysis required

The economic reality

For most industrial facilities, PD testing is specialized equipment that isn’t cost-justified for routine maintenance. It’s used for:

• Critical equipment where failure cost is extreme (main transformers, generator step-up transformers)
• New equipment commissioning where the OEM performs the PD test
• Troubleshooting when IR results are ambiguous
• Research and development for equipment manufacturers

IR testing is used for everything else — because it’s cheap, fast, and covers 80–90% of insulation problems across a typical equipment fleet.

When to Add PD Testing to Your Program

A practical decision framework.

Add PD testing when

• Equipment rated above 10 kV with high criticality (main bus transformers, utility generators)
• Catastrophic failure would cause extended outages or safety events
• IR testing results are ambiguous — low but not clearly failed, declining without obvious cause
• Equipment age is significant (>20 years for MV equipment) and you need remaining-life assessment
• A similar unit has already failed — understanding the failure mode helps prevent recurrence
• Manufacturer recommends it in the service manual (increasingly common for large machines)
• Regulatory or insurance requirements mandate it (growing trend in utility and nuclear industries)

Skip PD testing when

• Equipment rated below 3 kV — PD inception voltages usually above operating voltage
• Non-critical equipment where failure consequences are modest
• Budget limits are tight — IR testing gives much better coverage per dollar
• Equipment is new enough that PD is unlikely (under 10 years for most equipment)
• Equipment will be replaced soon anyway — no reason to spend on diagnostics for equipment already slated for retirement

The build-it-gradually approach

Most maintenance organizations implementing PD testing take a phased approach:

Year 1: Identify the 10 most critical pieces of medium-voltage equipment. Use contract PD testing service (not in-house equipment) to establish baselines.

Year 2-3: Repeat on those 10 pieces. Build trending database. Identify which units show developing issues.

Year 3-5: Consider in-house capability only if fleet size and testing frequency justify $100k+ investment.

Year 5+: Online PD monitoring on the top 1–3 most critical units if the trending data shows monitoring would catch problems early.

This phased approach captures most of the benefit at a fraction of the cost of full in-house capability.

FAQ

Key Takeaways

• IR and PD testing answer different questions. IR measures bulk insulation resistance at a moment in time. PD detects localized discharge activity that indicates developing degradation.
• IR excels at: commissioning, low-voltage equipment, moisture and contamination detection, fleet screening, post-repair verification.
• PD excels at: MV and HV equipment, detecting developing defects, online monitoring of critical equipment, failure-mode discrimination, life assessment.
• IEC 60270 is the primary standard for PD measurement. It defines terms, quantities (apparent charge in pC), test circuits, and calibration methods. It does NOT specify equipment-specific test levels — those come from product standards.
• Apparent charge (q) is the primary PD quantity, expressed in picocoulombs. Typical ranges: <10 pC (healthy), 10–100 pC (aged but functional), >1000 pC (active deterioration). Interpretation is equipment-specific.
• Offline PD testing is quantitative and calibrated; online PD monitoring is qualitative but continuous. Both have their place.
• PD appears mostly in equipment above ~3 kV. Below that, PD inception voltage usually exceeds operating voltage and PD activity is rare.
• PD testing is expensive — $30k+ equipment, specialized training. Justified primarily for critical MV/HV equipment.
• Most maintenance programs use IR as primary, PD as secondary — PD added only for the 10–20 most critical pieces of equipment in the fleet.
• The two tests are complementary, not competing. A complete diagnostic program uses both for critical equipment.

Standards Referenced in This Article

Standard	Key Content
IEC 60270:2000+AMD1:2015	High-voltage test techniques — Partial discharge measurements. Primary international PD measurement standard.
IEC 60034-27-2	Rotating electrical machines — On-line PD measurements on stator windings
IEC 60076-3	Power transformers — Insulation levels, dielectric tests, and external clearances (includes PD test requirements)
IEC 62478	Measurement of PD by electromagnetic and acoustic methods (UHF, acoustic)
IEEE 43-2013	IEEE Recommended Practice for Testing Insulation Resistance of Electric Machinery
IEEE 1434	IEEE Guide for the Measurement of Partial Discharges in AC Electric Machinery
IEEE C57.127	IEEE Guide for the Detection, Location and Interpretation of Sources of Acoustic Emissions from Electrical Discharges in Power Transformers