Third Harmonic Method for Surge Arresters: B1 vs B2 and the Voltage Harmonic Error

By | August 31, 2026

If your utility owns one arrester diagnostic instrument, it is probably a third harmonic meter. Clamp the earth lead, read a number, log it. No connection to a VT, no outage, no voltage reference of any kind.

That convenience is the reason the method spread. It is also the reason so many arrester monitoring programmes produce numbers nobody trusts.

For third harmonic analysis, IEC 60099-5 describes two approaches: B1 and B2. The difference between them is one field probe, and it decides whether your reading means anything.

Where the harmonics come from

Feed a metal-oxide arrester with a clean sine wave and the current coming out is not a sine wave. The V-I characteristic is non-linear, so the resistive current is distorted, and that distortion shows up as harmonics.

In the upturn region of the characteristic, the third harmonic is the largest harmonic component of the resistive current. It typically runs 10 % to 40 % of the resistive component.

That is the whole basis of the method. You cannot easily measure resistive current without a voltage reference. But you can measure the third harmonic of the total leakage current with nothing but a clamp, and the third harmonic tracks the resistive current. Measure the proxy instead of the thing.

The proxy is not fixed, though. The harmonic content depends on the magnitude of the resistive current and on how non-linear the characteristic is at that moment — which itself moves with voltage and temperature. Same normalisation problem as any other leakage current method.

What B1 assumes

Method B1 is the bare version. Read the third harmonic in the leakage current, treat it as an indicator of the resistive current, done.

It rests on an assumption: that all the harmonics in the leakage current come from the non-linear resistive behaviour of the arrester.

That assumption is not correct, for two separate reasons.

The smaller one: the capacitive current itself is observed to behave non-linearly, so it makes its own contribution to the harmonics.

The bigger one: your system voltage is not a clean sine wave either. Voltage harmonics drive capacitive harmonic currents through the arrester. These are not small compared with what you are looking for. The standard puts it plainly — the capacitive harmonic currents produced by voltage harmonics can be of the same order of magnitude as the harmonic currents generated by the arrester’s non-linear resistance.

So a B1 instrument reads a sum of two things and reports it as one.

The error is worse than “some inaccuracy”

IEC 60099-5 plots this. The evaluation error in the third harmonic of the leakage current is shown against the third harmonic content in the system voltage, with a band covering different phase angles of the voltage harmonic, different arrester capacitances, and different V-I characteristics.

Two features of that plot matter for field work.

The magnitude. The band opens as soon as any third harmonic is present in the voltage, and it widens steeply from there. You do not need a badly distorted network before the error stops being a tolerance and starts being a multiple. The error becomes significant well before the system voltage looks severely distorted.

The sign. The band runs negative as well as positive. This is the part that gets skipped. Voltage distortion does not only inflate your reading and cause a false alarm. Depending on the phase angle of the voltage harmonic, it can also cancel part of the arrester’s own harmonic current and pull the reading down. A degrading arrester can read healthier than it is.

An uncompensated third harmonic number, taken on a distorted network, is not a conservative measurement. It is just a measurement in an unknown direction.

What B2 adds

Method B2 works on the same principle and addresses the problem directly: measure the interference, then subtract it.

A field probe sits at the base of the arrester. The electric field there induces a harmonic current in the probe that is proportional to the capacitive harmonic current flowing in the arrester. Scale that signal properly, subtract it from the total harmonic current, and what remains is the third harmonic current associated with the arrester’s non-linear resistive current.

No voltage reference is needed. The method stays suitable for in-service measurement. The sensitivity to voltage harmonics drops from high to low.

The trade is handling complexity. The probe has to be positioned correctly and the scaling has to be right, so the measurement takes longer and depends more on the operator. In return you get a measurement far less sensitive to voltage distortion — greatly reduced, not eliminated. Both B1 and B2 have extensive service experience behind them, so this is not a choice between a proven method and an experimental one. It is a choice about whether the answer is trustworthy on your network.

The step both methods share

Neither method measures resistive leakage current directly. B1 gives you the third harmonic of the leakage current and treats it as a proxy. B2 first compensates the capacitive third harmonic, then uses what remains as the diagnostic quantity. Either way, the number in front of you is a harmonic current.

If you want to convert that into a resistive current level, you need conversion information — from the arrester manufacturer, or from laboratory measurements on the same design. And the standard warns that this conversion can introduce large errors of its own.

This is worth sitting with, because it changes how you should use the number.

Comparing a third harmonic reading against an absolute resistive current limit means stacking two error sources: whatever the measurement itself carries, plus the conversion. Comparing a third harmonic reading against the same arrester’s own history, or against sister units of the same design and vintage on the same busbar, skips the conversion entirely.

Trending in the raw measured quantity is more defensible than converting to an impressive-sounding resistive current figure. Normalise for voltage and temperature. Don’t convert unless you have solid manufacturer data.

Picking a method for your site

A short decision path:

  • Low third harmonic content in the system voltage, comparative trending, several identical arresters to compare. B1 can be appropriate, provided the voltage harmonic influence has actually been assessed rather than assumed. Cheap, fast, extensive service experience.
  • Any significant harmonic content in the system voltage. B2. The compensation is not a refinement here, it is the difference between data and noise.
  • You have a VT secondary available and would rather not add the B2 field probe. Method B3 uses only the fundamental components of voltage and current, which practically eliminates the influence of voltage harmonics. It still needs the voltage signal, and accuracy depends on phase shifts in the voltage and current.
  • You want the most robust in-service number and can accept a voltage signal. Method C determines the arrester’s average power loss directly, by integrating the instantaneous product of voltage and leakage current over one cycle. That multiplication and integration greatly reduces the influence of voltage harmonics.

There is an obvious question sitting underneath this: how much third harmonic is actually in your system voltage? Most utilities have power quality data and never think to look at it before an arrester campaign. Pull the THD and harmonic spectrum for the busbar first. That single check tells you whether B1 results from that site are worth keeping.

FAQ

Can I use a third harmonic instrument on any arrester?

The measurement needs a clean current path through the arrester’s earth connection. An insulated earth terminal is what makes that possible on most designs, and several of the diagnostic methods in Annex D depend on having one. Parallel current paths or an unsuitable mounting arrangement can invalidate the reading.

My readings went up 30 % since last year. Is the arrester degrading?

Not established yet. Check three things before concluding anything: the ambient temperature at both measurements, the service voltage at both measurements, and the harmonic content of the system voltage at both measurements. If the instrument is uncompensated, that third item alone can produce a change of this size with nothing wrong at the arrester.

Why the third harmonic and not the fifth or seventh?

In the upturn region of the V-I characteristic, the third is the largest harmonic component of the resistive current, so it gives the best signal to work with.

Is B2 immune to voltage harmonics?

Greatly reduced, not immune. The compensation depends on the probe signal being correctly scaled to the actual capacitive harmonic current. Placement and calibration still matter.

Do I need a different instrument for B1 and B2?

B2 requires an instrument designed to take a field probe input and perform the compensation. A basic clamp meter reading third harmonic content cannot be upgraded into a B2 measurement by technique alone.

Is a low third harmonic reading proof the arrester is fine?

No. It is one indicator among several, and on a distorted network an uncompensated reading can be pulled downward by the interference. Combine it with visual inspection, thermography, and the operating history of the unit.

Author: Zakaria El Intissar

Zakaria El Intissar is an automation and industrial computing engineer with 12+ years of experience in power system automation, electrical protection, and SCADA systems. He founded InsulationTesting.com to share practical, field-tested guides on insulation resistance testing, equipment, and industry standards. His writing is used by electricians and maintenance engineers worldwide. Based in Morocco.

Leave a Reply

Your email address will not be published. Required fields are marked *