Every medium voltage cable that reaches site under IEC 60502-2 has been through a chain of tests. Some ran on that exact drum. Some ran on a sample cut from the same batch. Some ran once, years ago, on a different length of the same design. And the last ones run after you’ve laid it and jointed it.
People mix these up all the time. A factory test report gets treated like a type test certificate. A type test certificate gets treated like proof that this drum is good. A site test gets set at factory levels and damages healthy insulation.
This article walks the cable through its four gates in order. For each gate: who does it, how often, what voltage, and what “pass” means.
Table of Contents
What IEC 60502-2 Covers
IEC 60502-2 covers power cables with extruded solid insulation from 6 kV up to 30 kV, for fixed installations like distribution networks and industrial plants. Part 1 covers 1 kV and 3 kV cables. Accessories are in Part 4.
It does not cover overhead, mining, nuclear containment, submarine or shipboard cables. Above 30 kV, the test regime moves to IEC 60840, covered in our cable partial discharge testing to IEC 60840 guide. For general field methods on any cable, see cable insulation testing.
The current text is Edition 3.1 (2024), which is the 2014 edition plus Amendment 1. The amendment mainly adds interlock tape armour, halogen-free sheaths and fire test requirements. The insulation tests below come from the 2014 base text.
Rated Voltages
Cables are named U0/U (Um):
- U0 is the voltage between conductor and earth or metal screen.
- U is the voltage between conductors.
- Um is the highest system voltage the cable can be used on.
The standard ratings are 3.6/6 (7.2), 6/10 (12), 8.7/15 (17.5), 12/20 (24) and 18/30 (36) kV.
Almost every test voltage in the standard is a multiple of U0. Keep that in mind. It’s the key to reading the whole document.
Picking U0: The Earth Fault Category
The same system voltage can need a different cable, depending on how long an earth fault is allowed to stay on.
- Category A: a faulted phase is cleared within 1 minute.
- Category B: the system runs for a short time with one phase earthed. For these cables, up to 8 hours on any occasion is tolerated, with no more than 125 hours of earth faults per year.
- Category C: everything else.
For Category C, you step up one U0 level. On a 12 kV system, Categories A and B use U0 = 6 kV. Category C uses U0 = 8.7 kV. On a 24 kV system it’s 12 kV versus 18 kV. On a 36 kV system, the standard gives no Category C option at all.
If your network often runs with a standing earth fault, the standard suggests treating it as Category C.
The U0 Ladder
Here’s every key insulation stress in the standard, in order. This one table explains most of the standard.
| Stress | Where it’s used |
|---|---|
| 1.73 U0 | Partial discharge acceptance level (factory and type test) |
| 2 U0 for 10 s | Short pre-stress before the PD reading |
| 3 U0 at 0.1 Hz, 15 min | After-installation VLF option |
| 3.5 U0, 5 min | Routine voltage test on every length |
| 4 U0, 4 h | Sample test and type test |
| 4 U0 DC, 15 min | After-installation DC option (not preferred) |
Notice the pattern. The factory routine test is short and high. The 4 hour test is longer and slightly higher. The site tests are lower, because they hit joints and terminations made in the field and a cable that’s already been pulled.
Gate 1: Routine Tests (Every Length)
Routine tests run on each manufactured length. The number of lengths can be cut, or a different method used, if a quality control procedure is agreed.
There are four:
- Conductor resistance
- Partial discharge (screened cables only)
- AC voltage test
- Electrical test on the oversheath, if agreed
Conductor Resistance
The cable sits in the test room for at least 12 hours at a fairly steady temperature, or 24 hours if there’s doubt the conductor has settled. You can also test a conductor sample after at least 1 hour in a temperature-controlled liquid bath.
The reading gets corrected to 20 °C and 1 km, then checked against the IEC 60228 maximum. This isn’t an insulation test, but it’s on every routine report, so it’s worth knowing.
Partial Discharge
This applies to cables with both a conductor screen and an insulation screen. Screens are mandatory on every cable except 3.6/6 kV PVC (always unscreened) and 3.6/6 kV EPR or HEPR with the thicker unscreened insulation.
The procedure:
- Raise the voltage to 2 U0 and hold for 10 s.
- Lower it slowly to 1.73 U0.
- At 1.73 U0, there must be no detectable discharge above the declared sensitivity.
Routine test sensitivity must be 10 pC or better. On three-core cables, every core gets tested, conductor to screen.
The 2 U0 step is there to give any weak spot a chance to start discharging. The reading is then taken at a lower level, so a defect has to keep discharging at a stress near operating conditions to fail.
The standard adds a blunt note: any partial discharge from the test object may be harmful. In extruded insulation, PD erodes the material. It doesn’t settle down.
AC Voltage Test
The voltage test is done at ambient temperature with power-frequency AC. The voltage is 3.5 U0, held for 5 minutes and raised gradually.
| U0 (kV) | 3.6 | 6 | 8.7 | 12 | 18 |
|---|---|---|---|---|---|
| Routine test voltage (kV) | 12.5 | 21 | 30.5 | 42 | 63 |
- Single-core: conductor to metal screen.
- Three-core with individually screened cores: each conductor to the metal layer.
- Three-core without individual screens: each conductor in turn to all the other conductors and the collective metal layers.
You can test a three-core cable in one go with a three-phase transformer. In that case the phase-to-phase voltage is 1.73 times the table value.
Pass means no breakdown.
Oversheath Test
This is an optional routine test, done only if customer and supplier agree. It follows IEC 60229. Cables with an extruded semi-conductive layer on the oversheath use the DC test from IEC 60229 instead.
Gate 2: Sample Tests (Per Batch)
Sample tests run on pieces cut from finished cable at a set frequency. They catch drift between type tests.
The list:
- Conductor examination
- Dimensions: insulation, sheath, armour, outer diameter
- 4 hour voltage test (above 3.6/6 kV)
- Hot set test (EPR, HEPR, XLPE and elastomeric sheaths)
How Often
Conductor and dimension checks: one length from each manufacturing series of the same type and size, capped at 10% of the lengths in the contract.
Electrical and physical tests: by agreed QC procedure. With no agreement, on contracts over 2 km of three-core or 4 km of single-core cable, the count is:
| Three-core cable length | Single-core cable length | Samples |
|---|---|---|
| Over 2 km up to 10 km | Over 4 km up to 20 km | 1 |
| Over 10 km up to 20 km | Over 20 km up to 40 km | 2 |
| Over 20 km up to 30 km | Over 40 km up to 60 km | 3 |
And so on.
If a Sample Fails
Take two more samples from the same batch and repeat the failed test. If both pass, the batch passes. If either fails, the batch fails.
This “two more” rule is a common point of argument on site, so it’s worth having it in front of you.
Insulation Thickness
This is the check that matters most for insulation. Two rules apply to every piece of core:
- Minimum: t_min ≥ 0.9 t_n − 0.1 mm
- Eccentricity: (t_max − t_min) / t_max ≤ 0.15
Here t_n is nominal thickness, and t_max and t_min are measured on the same cross-section.
The second rule is the one people forget. A cable can meet the minimum and still have its conductor sitting off-centre. That pushes up the stress on the thin side.
Example: 12/20 kV XLPE, nominal 5.5 mm.
- Minimum allowed: 0.9 × 5.5 − 0.1 = 4.85 mm
- If t_max is 6.0 mm, t_min must be at least 6.0 × 0.85 = 5.1 mm to meet the eccentricity rule.
Nominal XLPE insulation thickness, for the common conductor sizes:
| Rated voltage | Nominal XLPE thickness |
|---|---|
| 3.6/6 (7.2) kV | 2.5 mm (rising to 3.2 mm on large conductors) |
| 6/10 (12) kV | 3.4 mm |
| 8.7/15 (17.5) kV | 4.5 mm |
| 12/20 (24) kV | 5.5 mm |
| 18/30 (36) kV | 8.0 mm |
Semi-conducting screens don’t count toward insulation thickness.
4 Hour Voltage Test
This runs on a piece at least 5 m long between terminations. Power-frequency voltage at 4 U0 is applied between each conductor and the metal layers for 4 hours, at ambient temperature.
| U0 (kV) | 6 | 8.7 | 12 | 18 |
|---|---|---|---|---|
| 4 h test voltage (kV) | 24 | 35 | 48 | 72 |
It applies only above 3.6/6 kV. Pass means no breakdown.
Hot Set Test
This checks that XLPE, EPR and HEPR are properly crosslinked. A test piece is loaded at 20 N/cm² in hot air for 15 minutes:
- XLPE at 200 °C
- EPR and HEPR at 250 °C
The elongation under load must not exceed 175%. After cooling, the permanent elongation must not exceed 15%.
Poorly crosslinked XLPE stretches too far here. It’s a fast way to catch a curing problem that no electrical test at ambient temperature would show.
Gate 3: Type Tests (Once per Design)
Type tests prove a cable design before it’s sold commercially. They’re not repeated for every order.
When a Type Test Covers Your Cable
A passed type test also covers other sizes and voltages of the same type, if all three of these are true:
- Same insulation and semi-con materials, and the same manufacturing process.
- Conductor size no larger than the tested cable. Exception: if the tested cable was 95 mm² to 630 mm², all sizes up to 630 mm² are covered.
- Rated voltage no higher than the tested cable.
Conductor material doesn’t matter. Copper approval covers aluminium, and the other way round.
So when a supplier hands you a type test report, check these three points against your cable. A report for a 240 mm² 12/20 kV cable doesn’t cover an 800 mm² cable, or an 18/30 kV one.
The Electrical Sequence (Screened Cables)
A 10 m to 15 m sample goes through these steps in order, on the same piece:
- Bending test, then PD
- Tan δ measurement (can be on a separate sample)
- Heating cycles, then PD
- Impulse test, then AC voltage test
- 4 hour voltage test
The order is the point. The cable is bent, cooked, and hit with impulses before the final voltage tests. A design has to survive the damage first, then still hold voltage.
On three-core cables, every core is tested.
Bending
The sample is wound at least one full turn around a test cylinder at ambient temperature, unwound, then bent the other way without twisting. This cycle is done three times.
The maximum cylinder diameter depends on the construction (d is the conductor diameter, D the cable’s outer diameter):
| Construction | Single-core | Three-core |
|---|---|---|
| Lead sheath or longitudinal overlapped metal foil | 25 (d + D) + 5% | 20 (d + D) + 5% |
| Other cables | 20 (d + D) + 5% | 15 (d + D) + 5% |
Then comes a PD test, same steps as routine but at 5 pC sensitivity or better. Bending cracks semi-con layers and opens gaps at the screen interfaces, which is exactly what PD finds.
Amendment 1 adds a visual check of the oversheath after bending: no cracks, and no deformation that takes it below minimum thickness.
Tan δ
This is required only from 6/10 (12) kV upward.
The sample is heated until the conductor is 5 K to 10 K above its maximum normal operating temperature. Heat it in a liquid tank, an oven, or by passing current. Measure at 2 kV AC or more.
Limits:
| Insulation | Maximum tan δ |
|---|---|
| XLPE | 40 × 10⁻⁴ |
| EPR / HEPR | 400 × 10⁻⁴ |
That’s a ten-to-one gap. EPR is a lossier material by nature, so don’t compare EPR and XLPE tan δ on the same scale. If you’re doing site tan δ on a mixed network, keep that in mind. Our transformer power factor and tan delta guide covers the measurement side.
Heating Cycles
The cable is laid on the floor and heated by conductor current, to 5 K to 10 K above maximum normal operating temperature. On three-core cables, current goes through all conductors.
Each cycle:
- At least 8 hours long
- At least 2 hours at temperature
- At least 3 hours of natural cooling, to within 10 K of ambient
That’s done 20 times, then PD again at 5 pC.
This is where a design proves its screens stay bonded and its insulation doesn’t shrink back from the conductor under repeated expansion.
Impulse Then AC
The impulse test runs hot, again at 5 K to 10 K above maximum normal operating temperature. Each core takes 10 positive and 10 negative impulses, with no failure.
| U0/U (Um) | 3.6/6 (7.2) | 6/10 (12) | 8.7/15 (17.5) | 12/20 (24) | 18/30 (36) |
|---|---|---|---|---|---|
| Impulse peak (kV) | 60 | 75 | 95 | 125 | 170 |
The impulse wave has a front time of 1 µs to 5 µs and a time to half-value of 40 µs to 60 µs.
After the impulses, each core gets a 15 minute AC test at ambient temperature, at the routine voltage of 3.5 U0. No breakdown.
4 Hours at 4 U0
This is the last step: 4 hours at 4 U0, at ambient temperature, conductor to screen. No breakdown.
A new sample can be used for this step, but only if it has already been through bending and heating cycles.
Semi-Con Resistivity
This is done on a separate sample, before and after the ageing treatment, at maximum normal operating temperature ±2 K.
- Conductor screen: 1000 Ω·m maximum
- Insulation screen: 500 Ω·m maximum
Unscreened 3.6/6 kV Cables
At 3.6/6 kV, PVC cables are always unscreened, and EPR or HEPR cables can be. These cables get a different type test set, with a 10 m to 15 m sample per core.
Insulation resistance in water, at ambient temperature. Strip the outer coverings and soak the cores in water for at least 1 hour. Apply 80 V to 500 V DC, conductor to water, for 1 to 5 minutes until the reading is steady.
Insulation resistance in water, hot. The same test, with the water held within ±2 K of maximum conductor temperature for at least 1 hour first.
From the reading, you calculate volume resistivity ρ or the insulation resistance constant Ki, which normalises the reading to the cable’s length and insulation geometry:
ρ = 2π · l · R / ln(D/d)
Ki = 0.367 × 10⁻¹¹ × ρ (in MΩ·km)
Here R is the measured resistance in Ω, l is the length in cm, and D and d are the outer and inner diameters of the insulation.
| Compound | ρ at 20 °C | ρ hot | Ki at 20 °C | Ki hot |
|---|---|---|---|---|
| PVC/B | 10¹⁴ Ω·cm | 10¹¹ Ω·cm | 367 MΩ·km | 0.37 MΩ·km |
| EPR / HEPR | – | 10¹² Ω·cm | – | 3.67 MΩ·km |
Look at the PVC numbers. Resistivity drops by a factor of 1000 between 20 °C and 70 °C. That’s why an insulation resistance reading is meaningless without the temperature, and why temperature correction matters so much on cable.
Then come 4 hours at 4 U0 in water, and an impulse test on a separate sample: 60 kV peak, 10 of each polarity, run hot.
Gate 4: Tests After Installation
This is the gate you’ll run yourself. The standard keeps it short, and lets the purchaser and contractor agree on most of it.
Oversheath DC Test (Recommended)
A DC test between each metal screen or sheath and ground, at the level and duration in IEC 60229. It finds sheath damage from pulling, backfill and crushing.
For it to work, the ground must contact the whole outer surface of the oversheath. A conductive layer on the oversheath helps. If parts of the run sit in ducts or dry backfill, the earth contact there can be poor, so the test may not see damage on those sections.
Insulation Test (If Required)
The insulation test is optional. Where only the oversheath test is done, quality assurance during accessory installation can replace the insulation test, by agreement.
If you do run one, the AC options are:
| Option | Voltage | Frequency | Duration |
|---|---|---|---|
| a) | U (phase-to-phase) | 20 Hz to 300 Hz | 15 min |
| b) | U0 (system voltage) | Power frequency | 24 h |
| c) | 3 U0 RMS | 0.1 Hz | 15 min |
Tan δ and PD can be monitored during an AC test. The 0.1 Hz option came in with the 2014 edition.
Worked example for a 12/20 (24) kV cable:
- Option a): 20 kV for 15 minutes, 20 Hz to 300 Hz (resonant test set)
- Option b): energise at 12 kV to earth for 24 hours (soak test)
- Option c): 36 kV RMS at 0.1 Hz for 15 minutes (VLF)
For the equipment side of these, see AC withstand testing of cables and partial discharge testing of cables.
DC: Allowed, Not Preferred
As an alternative, the standard allows 4 U0 DC for 15 minutes. On a 12/20 kV cable that’s 48 kV DC.
The standard itself warns that a DC test may endanger the insulation under test, and says to use AC where possible. On extruded insulation, DC leaves space charge in the material. When the cable goes back into AC service, that trapped charge adds to the AC stress at any defect. That’s why a cable can pass DC and still fail after it goes back into service.
If someone hands you an old spec calling for DC hipot on new XLPE, push back and point to this clause.
Old Cables
For installations already in service, lower voltages or shorter durations may be used. The values should be agreed based on the cable’s age, environment, breakdown history and the reason for the test.
In plain terms: don’t hit a 20 year old cable with commissioning levels and then blame the test when it fails.
Test Conditions That Apply Everywhere
- Ambient temperature: 20 °C ± 15 °C unless a test says otherwise
- AC test frequency: 49 Hz to 61 Hz, substantially sinusoidal, RMS values (factory tests)
- Impulse wave: front 1 µs to 5 µs, tail 40 µs to 60 µs
- Conductor temperature: should be confirmed by one of the methods in Annex G
Maximum Conductor Temperatures
The type tests run “5 K to 10 K above maximum normal operating temperature,” so you need to know these:
| Insulation | Normal operation | Short-circuit (5 s max) |
|---|---|---|
| PVC/B, up to 300 mm² | 70 °C | 160 °C |
| PVC/B, over 300 mm² | 70 °C | 140 °C |
| XLPE | 90 °C | 250 °C |
| EPR / HEPR | 90 °C | 250 °C |
So XLPE type tests run hot at 95 °C to 100 °C conductor temperature.
The standard flags one field trap. A buried cable run continuously at 90 °C can dry out the soil around it. The soil’s thermal resistivity rises, and the conductor runs well above 90 °C. That’s a design problem, but it ends up as an insulation failure.
Reading the Paperwork
When a cable arrives, check what you actually have:
- Routine test report. Covers this drum: conductor resistance, PD at 1.73 U0, 3.5 U0 for 5 minutes. Check the drum number matches.
- Sample test report. Covers the batch: dimensions, eccentricity, hot set, 4 U0 for 4 hours. Check the sample count against your contract length.
- Type test report. Covers the design. Check materials, maximum conductor size and maximum voltage against your cable. Check it’s a full sequence, not just one test.
- Your site test. Agree the option (a, b, c or DC) and the levels in the contract before the cable is pulled.
Quick Reference
| Test | Gate | Level | Duration | Pass |
|---|---|---|---|---|
| Partial discharge | Routine | 2 U0 then 1.73 U0 | 10 s pre-stress | None above 10 pC |
| AC voltage | Routine | 3.5 U0 | 5 min | No breakdown |
| Insulation thickness | Sample | t_min ≥ 0.9 t_n − 0.1 | – | Plus eccentricity ≤ 0.15 |
| AC voltage | Sample | 4 U0 | 4 h | No breakdown |
| Hot set (XLPE) | Sample | 200 °C, 20 N/cm² | 15 min | ≤ 175% / ≤ 15% |
| Bending + PD | Type | 3 cycles, then 1.73 U0 | – | None above 5 pC |
| Tan δ (XLPE) | Type | ≥ 2 kV, hot | – | ≤ 40 × 10⁻⁴ |
| Heating cycles + PD | Type | +5 to 10 K over max | 20 × 8 h | None above 5 pC |
| Impulse | Type | 60 to 170 kV peak | 10 + 10 | No failure |
| AC after impulse | Type | 3.5 U0 | 15 min | No breakdown |
| AC voltage | Type | 4 U0 | 4 h | No breakdown |
| Oversheath DC | Site | Per IEC 60229 | – | Recommended |
| VLF | Site | 3 U0 at 0.1 Hz | 15 min | No breakdown |
| Resonant AC | Site | U | 15 min | No breakdown |
| Soak | Site | U0 | 24 h | No breakdown |
| DC | Site | 4 U0 | 15 min | Allowed, not preferred |
FAQ
What’s the routine test voltage for a 12/20 kV cable? 3.5 U0, which is 42 kV AC for 5 minutes, conductor to screen. This is done on every length in the factory.
What’s the VLF test voltage for a 12/20 kV cable after installation? 3 U0 RMS at 0.1 Hz for 15 minutes, which is 36 kV. That’s the option given in IEC 60502-2. Other documents may set different levels or durations, so agree the value in the contract.
Is a site insulation test mandatory under IEC 60502-2? No. The oversheath DC test is recommended. The insulation test is “if required.” By agreement, quality assurance during jointing can replace it when only the oversheath is tested.
Why is DC not recommended for XLPE cables? The standard says a DC test may endanger the insulation under test and recommends AC where possible. DC can leave trapped space charge in extruded insulation, which adds to the stress at a defect once AC service resumes.
Does a type test on one cable cover my cable? Only if your cable uses the same materials and process, isn’t larger in conductor size (with the 95 mm² to 630 mm² exception), and isn’t higher in voltage.
Why is the EPR tan δ limit ten times higher than XLPE? EPR is a lossier material by nature. The limit is 400 × 10⁻⁴ for EPR and HEPR versus 40 × 10⁻⁴ for XLPE. Don’t judge an EPR cable against an XLPE number.
Is partial discharge allowed in an MV cable? Not above the test sensitivity. At 1.73 U0 there must be no detectable discharge above 10 pC (routine) or 5 pC (type test). The standard notes any PD from the cable may be harmful.
What does the hot set test show? Whether XLPE or EPR is properly crosslinked. Under-cured insulation stretches too much at high temperature. Electrical tests at room temperature won’t catch it.
