Current transformer accuracy class and burden in indirect kWh metering — why 0.2S/0.5S and the right VA loading matter

The [guide on power and power factor measurement with a three-phase clamp meter](/guides/practical/vermogensmeting-cosfi-klemmenmeter-driefasig) covers direct current measurement with a clamp meter. For large consumer connections, however, energy metering is often carried out indirectly, via permanently installed current transformers (CTs) between the installation and the meter. This guide covers a detail that is often overlooked when selecting and connecting those current transformers: the accuracy class and the burden, and why a mismatched combination of the two gives a measurement error even though the CT itself is not faulty.

Accuracy class: 0.2S and 0.5S per IEC 61869-2

IEC 61869-2 (current transformers) defines, among other things, the accuracy classes 0.2 / 0.5 / 1 for metering purposes, as well as the special "S" classes 0.2S and 0.5S:

Class 0.2S: maximum ratio error of ±0.2% and phase displacement of ±0.2° at rated current and rated burden, guaranteed down to 1% of rated current (instead of 5% for the regular class 0.2).
Class 0.5S: maximum ratio error of ±0.5%, also guaranteed down to 1% of rated current.

The difference with the regular (non-S) classes therefore lies not only in the error limit, but especially in the range over which that error limit is guaranteed: an S-class CT stays within its stated accuracy even for a lightly loaded installation (for example at night, or for a large consumer with strongly varying demand), whereas a regular-class CT can give a larger error below 5% of rated current. For billing metering at large consumer connections, this matters because part of the consumption regularly falls well below the rated CT current.

Burden: why too little load is also a problem

The burden is the impedance (expressed in VA at rated secondary current) connected to the secondary side of the CT — in practice, the meter itself plus the wiring between them. The stated accuracy class of a CT explicitly applies at the rated burden: a CT with, say, a rated burden of 5 VA is characterized and tested at that load.

A common misconception is that a lower-than-rated burden is always favourable, on the reasoning that the CT is then "less loaded". In reality, the accuracy class applies over a burden range between a minimum (typically 25% of the rated burden) and the rated value; a burden well below that — for example a modern digital meter with a very low current input, connected to a CT designed for a much higher burden — can fall outside the guaranteed accuracy range of the class designation. Conversely, a burden that exceeds the rated value (for instance due to secondary wiring that is too long or too thin) gives a larger error than the class designation promises.

Why this causes a measurement error without the CT being faulty

A CT that was manufactured to specification and shows no defect whatsoever can therefore still measure outside its stated accuracy class once the actually connected burden does not match what the class designation applies to. This is not a fault of the CT itself, but a consequence of a mismatched combination of CT and secondary load — a distinction that is often missed during an unexplained metering discrepancy, because attention then first goes to the CT or the meter itself rather than the burden match between them.

Note: the exact minimum and maximum burden, the permitted power factor range of the burden (typically cos φ = 0.8 lagging for burden ≥ 5 VA, cos φ = 1 for burden < 5 VA), and the precise error limits are laid down in IEC 61869-2; for a specific project, the nameplate data of the CT in question and the meter manufacturer's specification apply.

Practical relevance

When replacing a meter on an existing CT-instrumented connection — for example when switching to a digital meter with a different (typically much lower) current input than the old electromechanical meter — it is worthwhile checking whether the new burden still falls within the range guaranteed for the CT class. For new large-consumer connections, the choice between 0.2S and 0.5S depends partly on the expected load profile: for strongly varying or predominantly light loads, 0.2S is better suited to remain accurate in the lower current range as well.

Common mistakes

Assuming a lower-than-rated burden is always better — below the minimum burden (typically 25% of the rated value), the CT falls outside its guaranteed accuracy class.
Combining an old CT with a new digital meter without checking the burden match — a burden that strongly deviates from the CT's rated value gives a measurement error that is not caused by the meter itself.
Using class 0.5 (without S) for a strongly varying load profile with many lightly loaded hours — without the S guarantee down to 1% of rated current, the error at low consumption can be larger than expected.
Extending the secondary wiring between CT and meter without recalculating the added burden — extra cable length adds impedance to the burden and can push it beyond the CT's rated value.