VA-TEK Engineering Every VLF test starts with one number: the test voltage. There is no instrument that tells you the right value — you choose it from the cable’s rating, its age and the reason it is being tested. Choose it too high and you risk the insulation you set out to prove; choose it too low and the test proves nothing.
This article works through VLF test voltage selection for medium-voltage extruded cable: what U0 means, how the 2U0 to 3U0 band is used, how new, aged and repaired cables pull the number in different directions, and why the waveform you pick changes the answer.
Key takeaways
- Test voltage is expressed as a multiple of U0, the rated phase-to-ground voltage — not the line-to-line system voltage.
- Acceptance testing of new cable sits at the top of the band (about 2.5–3 U0); aged in-service cable is tested lower (about 1.5–2 U0).
- The same set of kV works on any cable length; it is duration and frequency, not voltage, that change with length.
- Sinusoidal 0.1 Hz is the reference waveform. Cosine-rectangular has a different peak-to-rms relationship and is treated separately.
- The older the cable, the lower the voltage — a full acceptance level on aged insulation can precipitate the failure you meant to prevent.
1. Start from U0, not from the system voltage
A cable is rated by two voltages: U, the line-to-line system voltage, and U0, the voltage between each conductor and its shield. For a three-phase cable the two are related by the square root of three:
U0 = U / √3
A cable marked 12/20 kV therefore has U = 20 kV and U0 = 12 kV. VLF test voltages are always written as multiples of U0, because that is the stress the insulation actually sees between conductor and earth. A “3 U0” test on a 12/20 kV cable is 36 kV, applied between each phase conductor and the shield while the other phases are grounded.
Confusing U with U0 is the most common arithmetic error in VLF records. It inflates the intended test voltage by 73 percent — enough to turn a routine proof test into an over-stress event on a cable that was already marginal.
2. Three test classes, three voltage bands
The test voltage is not a fixed property of the cable; it is a property of the question you are asking. Field practice recognises three classes, and the standard that governs the test usually names the one that applies.
| Test class | When it is run | Typical 0.1 Hz sine voltage | What it is for |
|---|---|---|---|
| Acceptance | New cable, after installation and before energisation | about 2.5 U0 (band 2.0–3.0 U0) | Prove the installation, the joints and the terminations |
| Maintenance | In service, during a planned outage | about 1.5–2.0 U0 | Confirm the cable still holds; avoid over-stressing aged insulation |
| After repair | After a splice or fault repair | match maintenance, or slightly lower | Prove the repair section and the reinstated cable |
IEEE 400.2 tabulates these as ranges rather than a single number, because the right point depends on the cable type and its history. What matters more than the exact figure is the direction of travel: the older the cable, the lower the voltage. A cable that has carried load for thirty years and may contain water trees does not benefit from the full acceptance level. At that point the test stops being a proof and becomes an additional stress event.
3. A worked table: U0, 2U0 and 3U0 by rating
The arithmetic is simple once U0 is known. The table below gives the conductor-to-shield voltage at common multipliers for standard MV ratings.
| Cable rating (U/U0) | U0 (kV) | 2 U0 (kV) | 3 U0 (kV) |
|---|---|---|---|
| 6/10 kV | 6.0 | 12.0 | 18.0 |
| 8.7/15 kV | 8.7 | 17.4 | 26.1 |
| 12/20 kV | 12.0 | 24.0 | 36.0 |
| 18/30 kV | 18.0 | 36.0 | 54.0 |
| 26/35 kV | 26.0 | 52.0 | 78.0 |
Two points fall out of this table. First, a 26/35 kV acceptance test at 3 U0 needs 78 kV rms — and the peak of that sine wave is 1.41 times higher again, so the set must be insulated and rated for over 110 kV peak. That is a trailer-mounted class, not a handheld. Second, the same “3 U0” is 18 kV on a 10 kV cable and 78 kV on a 35 kV cable: the multiplier travels across ratings, the absolute voltage does not.
4. Duration trades against voltage
Voltage and hold time are not independent decisions. The standard practice is to spend the stress budget over a longer time at a lower voltage rather than a shorter time at a higher one, because insulation degradation is driven by both magnitude and duration, and a slow ramp gives the operator time to watch leakage current and tan delta move.
In practice:
- Higher voltage, shorter hold — acceptance work at 2.5–3 U0.
- Lower voltage, longer hold — maintenance work at 1.5–2 U0, commonly 30 to 60 minutes at 0.1 Hz.
Frequency enters here too. A 0.1 Hz waveform applies one twentieth of the cycles a 50 Hz test would apply in the same wall-clock time, so some standards scale the required hold time with the test frequency to accumulate a comparable number of stress cycles. If you drop to 0.05 or 0.02 Hz to reach a longer cable, check whether the standard requires you to extend the hold. That relationship is set out in VLF test duration and frequency.
5. Waveform: why sinusoidal is the defensive default
VLF sets are built around two waveforms: sinusoidal, and cosine-rectangular (a square wave whose pulse areas are shaped so the voltage reverses through a cosine-like ramp). Both are alternating, and both avoid the space-charge problem that makes DC harmful to extruded insulation. They are not equivalent, however.
A cosine-rectangular wave carries harmonics of the base frequency. Those harmonics change the ratio between the rms value a meter reads and the peak voltage the insulation actually experiences, and they change how partial discharge behaves at the voltage reversals. IEEE 400.2 addresses the waveform explicitly and treats the sinusoidal wave as the reference; where a standard quotes a voltage and a duration without qualification, it means the sinusoidal 0.1 Hz case.
The practical rule: use sinusoidal 0.1 Hz unless the applicable standard or the cable owner explicitly calls for cosine-rectangular. If the rectangular wave is used, record the waveform alongside the voltage, because a “2.5 U0” figure is not transferable between the two without adjustment.
6. Where this gets you
Voltage selection is a decision, not a lookup. Anchor on U0, place the cable in the right test class, use the top of the band only for new work, and move down as the cable ages. When in doubt, the lower voltage held longer is the safer engineering choice: a withstand test that passes at a defensible level tells you more about the cable than one that risks it to reach a round number.
With the voltage set, the remaining decisions are the waveform and whether a withstand alone is enough or a diagnostic is required. Those are covered in the practical VLF guide and in VLF withstand vs diagnostic testing. The test set must be able to reach the voltage you select — see VLF hipot test sets for the ranges each class covers.
