In recent years secondary or sub-meters have been developed to control energy consumption and monitor usage. They are specifically designed for their market – energy management, cost accounting and so on – and differ substantially from tariff meters. However, the only way of specifying the performance of such meters was to use the standards written for tariff meters, which was not ideal.

A new Code of Practice covering secondary or sub-meters is about to be released by the British Standards Institute (BSI). 

Background to the new code

Four years ago ESTA (the Energy Systems Trade Association) decided to approach BSI for a new standard covering sub-metering. What has been prepared is an inclusive document that covers “Newly manufactured electronic secondary or sub-meters for the measurement of an electrical energy or energies either alone or together with other electrical parameters associated with the energy.”

It requires meters to be clearly and simply identified with regard to accuracy, functionality, performance and so on, irrespective of whether the unit is a precision multifunction device or a simple kWh meter. It is designed to complement rather than replace the existing metering standards IEC 62052 and IEC 62053, and has drawn extensively on the many years of experience that has gone into drafting these documents. 

It also provides manufacturers and users with a single document that can be used to define a meter or a metering requirement. The present situation, where existing standards cannot be used in total, means that each manufacturer selects those parts that are considered important – and perhaps omits those that may be inconvenient.

The most commonly used section of existing standards is the accuracy definition, but this is interpreted differently by individual users. A specification may require “Class 1 accuracy per IEC 1036”. However, a manufacturer may state that his accuracy complies with the requirements of “IEC 1036 Class 1, 20% to 120% of full load” (whereas the standard specifies accuracy from 2%) or even “Accuracy Class 1 per clause/table xx of IEC/BS zzz” (excluding all influencing effects, including temperature). A cynic may question the reasons for defining accuracy in such a limited way.

Another difference between this new Code of Practice and existing standards is that it takes into account the sort of power waveforms that are experienced in practice, rather than the theoretical sine waves used in a test laboratory.

Figure 1

Figure 1 shows the differences between the type of signal used for testing and real-world conditions. In the real world there is system noise, loads vary, people switch things on and off while electronic control allows fast on/off load switching (often for temperature control). Phase 1 shows the effect of load switching using a burst fire controller; such changes could easily happen 50 times a second.

Then there are non-linear loads. Virtually all types of electronic equipment form non-linear loads. On phase 2, some electronic equipment is turned on part way through a cycle, while on phase 3 a dimmer is in ½ phase mode. Both have significant harmonic content.

These are the sorts of load we need to measure with accuracy, yet some meters can take 10 to 20 seconds to accurately measure a change in load.

The Code of Practice has been written for manufacturers, specifiers and users, to provide one single reference document. It covers six aspects of a meter – safety, performance, accuracy, definitions, marking and testing.


The new Code of Practice does not deal directly with safety but refers to EN 61010-1, the generic safety standard for electrical equipment. This will help to remove the present confusion where the safety specification of a kWh meter is defined by EN 61036, a kVArh meter by EN 61287 or EN 61010-1 (different requirements) and a kVAh meter by EN 61010-1.

The Code does, however, deal with the issue of galvanic isolation of CT operated meters. In recent times some meter manufacturers have offered non-isolated meters in order to minimise costs. Under normal conditions the CT connections of such non-isolated meters would typically float at a potential near to neutral. However, under certain fault conditions dangerous voltages can be found on the unearthed CT connections of such meters. The new Code of Practice assumes that the current circuits of CT operated meters are galvanically isolated from the voltage terminals and can be earthed for safety. Where this is not the case, the manufacturer must explicitly say so.


The Code of Practice considers measurement technique, as it has been found that certain types of meter may not measure accurately under certain load conditions. For example, on loads subject to fast on/off switching control, it is possible for certain types of meter to measure a zero load despite current being drawn through the circuits.

The Code defines two types of operation.

  1. Continuous measurement, where the meter measures each and every cycle of the power signals (voltage and current).
  2. Intermittent measurement, where the meter is prevented from measuring each and every cycle for whatever reason.

It further divides polyphase measurements into two different types.

  1. Parallel measurement, where the measurement technique used ensures that each and every cycle of both current and voltage on all phases is measured, thus ensuring that no single event is missed.
  2. Serial measurement (or sequential measurement) where the measurement technique used is such that the meter will not necessarily measure each and every cycle of all phases. A meter shall be considered to use ‘serial measurement' if the measurement element is switched to measure each phase in sequence in such a way that all individual cycles are not measured.

The ideal system is ‘continuous parallel' as this ensures correct measurement under all conditions. However, where loads are known to be relatively stable, and where electronic load control is not used, even ‘intermittent serial' type meters could provide adequate accuracy.

The Code of Practice also considers all types of output – pulse, analogue and digital (computer). Requirements for accuracy, speed of response, ripple content and compliance voltage (analogue output) etc are defined.


The accuracy of a meter is defined as a ‘class index'. However, if a meter is classified as Class 1, it does not follow that the error under practical conditions of use will be within 1% of the actual value. It means that the error should not exceed 1% of the actual value under closely specified conditions, and only over a limited part of the measurement range. The permissible error of a meter under working conditions is the sum of the permissible intrinsic error and of the permissible variations due to each of the influence quantities (eg temperature, magnetic field). In practice the actual error is likely to be much smaller than this, because not all the influence quantities are likely to be simultaneously at their most unfavourable values, and some of the variations may cancel each other out.

The main requirements of the Code of Practice with reference to accuracy are:

  • Accuracy must be quoted as a percentage of reading.
  • Any accuracy defined otherwise must be clearly identified.
  • If only a single accuracy is quoted, it must apply to all measured parameters.


There are several ways of defining certain parameters and different ways of measuring them. The Code of Practice defines the commonly measured parameters and, where multiple definitions exist, demands that the manufacturer states which is being used. This should help users to understand the reasons why different types of meter give different readings on identical loads.
The Code also defines all parameters and functions necessary for defining a meter.


The Code of Practice demands that the meter be fully and comprehensively marked and that all necessary extra information is provided in user documentation. These must include:

  • Manufacturer's name and meter model reference
  • Connection type 
  • Serial number and year of manufacture
  • Nominal voltage, current range and frequency
  • Accuracy
  • Auxiliary supply requirement and tolerance.


A full test procedure is defined to allow manufacturers to prove compliance with the Code.

It includes:

  • General testing procedures, including reference conditions
  • Mechanical testing, including shock and vibration
  • Climatic testing
  • Safety
  • Electrical testing.

The new Code offers users the ability to specify a meter that will measure precisely and accurately under practical conditions. It also acts as a reference to which meters can be designed and manufactured – for 50/60Hz, for dc, for 400Hz etc. There are very few meters on the market that could not apply this Code of Practice.

The full text of the report on which this article is based is available on our website