On 1 October 2026, ESCO Technologies completed the acquisition of Megger Group for roughly US$2.35 billion — about 14 times projected EBITDA, with the seller taking a fifth of the buyer in shares. Megger becomes part of ESCO’s Utility Solutions Group alongside Doble. Read that as a signal about where the money in cable testing now sits: not in the ability to apply a voltage, but in the ability to say something about the insulation while you do it.
That is the same direction the standards took. IEEE 400.2-2024 replaced leakage current with tan delta and partial discharge as the criteria for a monitored withstand, and NFPA 70B-2023 turned electrical maintenance from a recommended practice into an enforceable standard whose insulation-test list names VLF, tan delta and partial discharge by name. The three techniques are no longer competing answers to one question. They answer three different questions, and a 2026 test programme is expected to say which one it ran and why.
This article separates them: what each proves, what each misses, and how to sequence them without paying twice.
The short answer
- VLF withstand proves the cable holds voltage today. It is a proof test. It trips on gross defects and is blind to distributed ageing.
- Tan delta measures how much energy the insulation loses. It is a bulk measurement — good for ranking a fleet and for trending one circuit over years, blind to a single bad joint.
- Partial discharge finds localised defects — a void, a delamination, a badly made termination. It is the only one of the three that locates anything.
- Run in the order withstand → tan delta → PD on one energisation and you get a proof test plus a condition record for the same site visit. IEEE 400.2-2024 calls this a monitored withstand.
1. Why a cable can pass a withstand test and still fail next year
The confusion behind most “VLF vs tan delta” searches is a category error: VLF is a voltage source, tan delta is a measurement. They are not alternatives. You apply VLF and you measure tan delta on it. The real question is which measurement you attach to the source.
That question only matters because a withstand test answers one thing. Apply the voltage, hold it, watch for breakdown. A circuit that survives is a circuit with no defect severe enough to fail under that stress right now. What it does not tell you is the state of the insulation that survived.
The failure mode it misses is distributed ageing. Water trees grow from voids and contamination in XLPE over years; they raise local conductivity without creating a single gap that breaks down. A cable full of water trees can pass a perfectly executed hipot and fail eighteen months later. The reverse is also true and less discussed: a single well-made void in a joint can sit in a cable whose bulk loss is unremarkable, and the hipot will not find it either, because the defect is too small to trigger breakdown at the test level.
Two blind spots, in opposite directions. One needs a bulk measurement, the other needs a localised one. That is the entire reason tan delta and partial discharge exist as separate tools.
2. VLF withstand: a proof test, and only a proof test
VLF is alternating voltage below 1 Hz, applied at 0.1 Hz in most field work. The reason it exists is arithmetic: the charging current a cable draws is I = 2πfCV, and dropping the frequency from 50 Hz to 0.1 Hz cuts that current by a factor of about 500 at the same voltage. A test set that would need a truck becomes one two people can carry.
What it buys you is a genuine AC stress with polarity reversal, without the space charge that makes DC hipot unacceptable on extruded insulation. What it delivers is a pass or a fail, at the voltage the standard tabulates for the cable class and purpose. The test voltage article works through reading those values; the short version is that IEEE 400.2-2024 Table 3 gives them in kV rms, and the 2024 edition unified every waveform on that basis, so a bid quoting peak against a table quoting rms is not comparing the same quantity.
The honest limitation: a pass is not a clean bill of health. It is a statement that nothing failed today.
3. Tan delta: the measurement that ranks a fleet
Tan delta — dissipation factor — is the ratio of resistive to capacitive current through the insulation. Clean, dry insulation loses almost nothing; the loss angle is small. Moisture, contamination and water trees add conductive paths, and the loss angle grows.
At low frequency that loss is more visible than at mains frequency, which is why tan delta and VLF are usually discussed together: the same 0.1 Hz energisation that stresses the cable produces the reading. A field instrument such as the VLF-34TD or VLF-45TD carries the tan delta measurement in the same enclosure as the withstand source, so there is no second setup and no second outage.
Two things about interpretation matter more than the absolute number.
It is a bulk average. Tan delta sums the whole cable length. A large population of water trees is needed before it moves; a single defective joint barely registers. That makes it excellent for “which of these forty circuits should I replace first” and poor for “where is the fault”.
The judgement is relative. IEEE 400.2-2024 adds tan delta assessment tables that differ by insulation type — PE, filled EPR and PILC do not share a pass line. Beyond the table, engineers read the stability: how loss changes with test voltage (tip-up), and how it changes against the same circuit’s previous readings. A single reading in isolation is weak evidence; the third reading in a series is strong evidence. The tan delta article goes through the interpretation in detail. For the method across asset types, including transformer and bushing dissipation factor, see tan delta testing.
Searches for an “acceptable tan delta value” usually want one number. The honest answer is that the number depends on insulation type, temperature, cable construction and history — which is exactly why the standard tabulates rather than fixes.
4. Partial discharge: the only one that locates anything
Partial discharge is a small, localised breakdown that does not bridge the insulation. It occurs in voids, at delamination interfaces, at the semicon edge, and in badly installed joints and terminations. It degrades the insulation progressively if it continues.
PD is the only one of the three tests that points at a place. Tan delta tells you the cable is lossy; PD tells you which joint.
Three practical points decide whether a PD measurement is worth anything:
- Test high enough. CIGRE TB 728 recommends PD testing at or above 1.7 U₀, because most discharge inception voltages sit above 1.5 U₀. Test below inception and you will measure nothing on a cable that has a defect.
- Read the trend, not the number. PD readings under VLF vary with residual charge and with test order; inception values repeat poorly, extinction values and trends repeat better. A single absolute pC figure across different cables is not comparable.
- PD needs sine. Cosine-rectangular VLF is a withstand waveform; it cannot support tan delta or PD diagnosis. If the deliverable includes either measurement, the waveform must be sine.
PD is added to a VLF set through external coupling — HFCT or acoustic — rather than built into the source. The PD article covers coupling and the limits of diagnosis at VLF. For PD measurement as a whole — IEC 60270, on-line monitoring, UHF and acoustic — see partial discharge testing.
5. Monitored withstand: running all three for one outage
The 2024 revision of IEEE 400.2 formalised what field engineers had been doing anyway. A monitored withstand runs the hold while measuring tan delta and partial discharge, so a single energisation yields both the proof and the condition record.
The sequencing rule that follows is straightforward:
| Step | Test | What you get | What it cannot tell you |
|---|---|---|---|
| 1 | VLF withstand, sine, at the Table 3 value | Pass or fail today | Whether the insulation is ageing |
| 2 | Tan delta, at each test frequency or voltage step | Bulk condition; trend against previous tests | Where a localised defect is |
| 3 | Partial discharge, ≥ 1.7 U₀ | Location and severity of a localised defect | Bulk ageing across the whole run |
When the job needs only a commissioning proof — new cable, contractor handover, insurance against workmanship — step 1 alone is defensible and is what the standard requires. Steps 2 and 3 earn their cost on service-aged cable, where the question has changed from “did we install it correctly” to “how much life is left and where will it fail”. The withstand-versus-diagnostic article draws that line in more detail.
6. What changed in 2026, and why this is now a purchasing question
Three forces moved at once, and together they changed what a test record is expected to contain.
The maintenance standard became enforceable. NFPA 70B moved from a recommended practice to a standard in its 2023 edition, rewriting “should” as “shall”. It does not mandate an annual VLF test on every asset — the annual interval applies to infrared inspection — but for equipment rated above 1,000 V it requires the maintenance programme to select insulation-condition test methods from a list that explicitly includes VLF, tan delta and partial discharge. Choosing a method is now a documented decision rather than a habit, and a facility that cannot show why it chose one has a gap in its programme.
The cable test standard raised the bar on the record. IEEE 400.2-2024 unified test voltages on an rms basis, tightened the monitored-withstand criteria to tan delta and partial discharge, raised the minimum HV withstand time from 30 minutes to 60, added a mandatory slow controlled rundown, and added tan delta assessment tables by insulation type. The direction is unambiguous: voltage alone is no longer a complete result.
The fleet is old enough that the question has changed. About 30% of the EU27-plus-Norway electricity grid is more than 40 years old, and around 30% of European distribution systems predate 1980 (Eurelectric, 2024); in the US, roughly 70% of transmission lines and large transformers are 25 years or older. Global grid spending, flat at around US$300 billion a year from 2015, rose to roughly US$400 billion in 2024, with the EU’s Grid Action Plan alone committing €584 billion to 2030. Money is arriving for replacement, and replacement decisions are made on condition data, not on pass/fail counts.
The market has priced this in. Cable testing and diagnostics equipment was measured at about US$890 million in 2024, forecast to US$1,296 million by 2031 at roughly 5.6% CAGR, while the narrower VLF hipot test-set segment was put at about US$540 million in 2024 in a separate estimate. Consolidation followed: the Megger acquisition at the top of this article put the two largest diagnostic brands under one owner, which changes the procurement risk for anyone buying in this category and makes the multi-test capability of a single instrument more valuable, not less.
Market figures are taken from published summary data and are definition-sensitive: narrow “power cable testing and diagnostics equipment” is around US$0.9 billion, while broad definitions that include test-and-certification services exceed US$15 billion. The figures above are not comparable across definitions.
7. Choosing an instrument for the three-test programme
Once the programme needs tan delta, the instrument choice narrows quickly, and the answer turns on one question: does the set need DC and tan delta, or is AC withstand enough? The VLF hipot family splits on exactly that line.
| Requirement | Instrument | Output |
|---|---|---|
| Commissioning and maintenance withstand only | VLF-30-90 family — six models, 30 to 90 kV peak | AC sine, four frequencies, 30 mA, split controller and HV tank |
| Withstand + tan delta + DC + sheath testing | VLF-34TD (5–15 kV cable) or VLF-45TD (5–25 kV cable) | AC sine and square, ±DC, integrated tan delta, sheath fault location, single 22–23 kg unit |
| Transformer and bushing dissipation factor | TDC-3100 | Tan delta and capacitance for apparatus, not cable |
The practical rule: the voltage class decides the set before the diagnostics do. For a 15 kV cable the acceptance level is 21 kV rms ≈ 30 kV peak, which the smallest rating covers. For 25 kV cable the acceptance value is 32 kV rms, which sets the requirement at the 45 kV peak class. Add tan delta and PD only when the cable is in service and ageing is the question — not because the data sheet offers it.
Two tools take the arithmetic out of it: the VLF test set selector works from cable capacitance, test current and frequency to show how long a run a given set can reach, and the VLF test voltage calculator returns the IEEE 400.2-2024 value for a cable class. How to choose a VLF test set takes the same four questions further, and the VLF cable testing guide is the reference article for the whole sequence.
8. Questions that come up in every specification meeting
Is partial discharge testing destructive? No. PD measurement is a passive observation of discharges that already occur at the test voltage; it does not inject energy beyond the withstand stress the cable is already taking. The test voltage is the same order as the withstand level, so the risk profile is that of the withstand itself, not an additional one.
Can I do tan delta with a VLF set that has no tan delta function? Not meaningfully. Tan delta requires the set to measure the phase relationship between voltage and current, which is a different measurement chain from a voltage display. Sets without the function can apply the voltage; they cannot produce the number. That is the reason the TD variants exist as separate models rather than as a firmware option.
What is an acceptable tan delta value for MV cable? There is no single figure. The judgement depends on insulation type — PE, filled EPR and PILC have different assessment tables in IEEE 400.2-2024 — and on the trend against previous tests on the same circuit. Treat any answer that quotes one number without asking about insulation type and history as unreliable.
Do I need all three tests every time? No. Newly installed cable needs the withstand, because the question is workmanship. Service-aged cable in a replacement programme needs the diagnostics, because the question is remaining life and location. The mistake in both directions is expensive: skipping diagnostics on old cable, or paying for a PD survey on a cable commissioned last month.
Which test does NFPA 70B require? It requires the maintenance programme to select a method from a list that includes VLF, tan delta and partial discharge, and to document the choice. It does not mandate a specific test for every asset; it makes the selection a decision you have to be able to defend.
Where this leaves a test programme
The three techniques are not rivals, and the “VLF versus tan delta” framing is a category error that costs people money. VLF is the source. Tan delta and partial discharge are two measurements with opposite blind spots — one sums the whole cable, the other finds a point in it. The 2024 standard made the pairing explicit, the maintenance standard made the selection documented, and the market has consolidated around vendors who can supply both.
A defensible 2026 programme therefore looks like this: VLF withstand at the tabulated voltage, on sine, for the tabulated hold time, with tan delta recorded across the hold and partial discharge measured above 1.7 U₀ where the cable is old enough for the answer to matter. One energisation, one record, three answers.
VA-TEK builds VLF withstand sets and integrated VLF plus tan delta systems for 5–46 kV shielded cable. For a test-voltage recommendation against a specific cable schedule, a load-rating check for a given run length, or the current datasheets, request a quote or contact the engineering team. For the surrounding method — voltage selection, hold time, waveform, DC comparison and the oscillating-wave alternative — start with the VLF cable testing guide and VLF versus DC.
Sources: IEEE Std 400.2-2024; NFPA 70B-2023; CIGRE TB 728; IEC 60502-2; GB/T 3048.8-2025 (published 2025-10-05, effective 2026-05-01); IEA grid investment data; Eurelectric Grids for Speed (2024); European Commission Grid Action Plan; DOE grid-ageing figures; QYResearch cable testing and diagnostics equipment summary, 2025; QYResearch VLF hipot tester summary, 2025; ESCO Technologies announcement of the Megger acquisition (15 April 2026) and completion (1 October 2026). Market figures are from public summaries of paid reports and are definition-sensitive — see the note in section 6. Always confirm test voltages against the current official standard before a live job.
