Controls, metering and communications
Meter accuracy verification compares what an installed meter, transducer or metering input displays with a calibrated reference source or instrument. It shows whether the voltage, current, power and energy values that operators, billing systems and protection studies rely on are trustworthy, and it is run at commissioning and at calibration intervals.
Meter accuracy verification checks that an installed metering device reports the electrical quantity it is actually seeing. The device may be a panel meter, a microprocessor-based multifunction meter, a transducer feeding a control system, the metering function inside a protective relay or breaker trip unit, or a revenue-class meter. A known signal is applied, or a calibrated reference instrument is connected beside the device, and the displayed or transmitted value is compared with the reference. The quantities of interest are voltage, current, frequency, real and reactive power, power factor, demand and energy, along with the analog or digital output sent to remote systems. The test covers the whole measurement chain where that is practical, including the instrument transformers and the wiring to the meter, because an accurate meter behind a miswired or mismatched transformer still reports the wrong answer. It is performed on switchgear, switchboards, generators, transformers, motor control centers and any other equipment whose metering is used for operations, billing, load studies or alarms.
Meters are trusted without being questioned, so an error can persist for years. An energy meter that reads low costs money every month, and one that reads high creates disputes. Load readings that drift lead to circuits being overloaded unnoticed or capacity being added that is not needed. Power quality and demand data feed engineering studies, so wrong inputs produce wrong conclusions. Operators who see a trustworthy voltage and current make better decisions during an event, and metering values that are used for alarms or automatic control can trip or fail to trip when they should not. Typical causes of error include instrument transformer ratio and polarity mistakes, wrong settings in the meter, shorted or open current circuits, loose or wrongly connected potential leads, burden that is too heavy for the transformer, component aging and drift in the meter itself. A verification at acceptance catches installation errors before the system is relied upon, and a periodic check catches drift that no one would notice from the display.
The principle is comparison against a reference of better accuracy. The reference may be a laboratory-grade standard or a portable field instrument that is itself calibrated against a traceable standard, and the general rule is that the reference should be clearly more accurate than the device under test so that its own uncertainty does not dominate the result. A stable source applies known values of voltage and current across the operating range, and at several points, so that linearity can be judged and not only a single spot. Phase angle matters for power and energy: the meter must multiply voltage and current with the right angle between them, so checks are made at unity and at lagging and leading power factors where the instrument allows. Accuracy is expressed as the difference between the meter value and the reference value, relative to the reading or to the full scale, depending on how the device is specified. Because the instrument transformers add their own ratio and phase error, the verification either includes them in the loop or treats them separately, with their burden and ratio taken from test or nameplate data.
The equipment is prepared under the owner's safety procedures. Current transformer secondaries are never opened while the primary is energized, so a shorting block or test switch is used, and wiring is checked before any signal is applied. Where the loop is verified by injection, a calibrated source is connected to the meter's inputs, or to the test switch provided for the purpose, with the transformers isolated, and known values are applied at several points through the working range while the meter's display and any output are recorded. Where the loop is verified live, a calibrated reference instrument is connected in parallel and both readings are taken at the same moment under steady load. The meter's configuration is also checked against the design: instrument transformer ratios, connection type, scaling of analog outputs, communication settings, and demand or integration intervals. Outputs sent to other systems are compared at the receiving end, which makes the test a check of the signal path as well as the meter. If the device is adjustable, adjustments are made only as the manufacturer allows, and the as-found readings are recorded before any change so that drift can be tracked. Conditions that affect the reading, such as temperature, waveform distortion and frequency, are noted with the results.
The record holds the identity of the meter and its function, the quantity checked, the reference value applied or read, the value displayed or transmitted, and the difference between them, together with the as-found state before any adjustment and the as-left state after it. It names the reference instrument or source and its calibration status, notes the instrument transformer ratios and the configuration found, and records ambient conditions and the load or signal level at which each check was made. Any wiring, setting or scaling corrections are listed so that a later test can see what changed and why. The aim is a record that lets the next verification be compared with this one.
Results are judged by the difference between the meter and the reference, expressed in the way the device is specified, against the accuracy class or tolerance published by the manufacturer and, for revenue metering, the applicable utility or standards requirements. This page publishes no numerical limits. The shape of the judgment has several parts. The first is whether the error at each test point stays inside the stated accuracy of the device, which varies by quantity and by point in the range. The second is consistency across the range and across phases, since a single outlier points to a wiring or channel fault. The third is the as-found versus as-left comparison, and the trend against earlier verifications, where steady drift in one direction is a warning even when the device is still in tolerance. The fourth is agreement of configuration with the drawings. The functional meter and metering checks are described in ANSI/NETA ATS-2025 for acceptance and ANSI/NETA MTS-2023 for maintenance; refer to your own licensed copies for the requirements that apply to each equipment class.
At acceptance the aim is to prove the installation as a whole, including ratios, polarity, wiring and configuration, and to establish the as-left record that later checks will be compared with. In maintenance the focus is on drift and on devices that have been changed, replaced or reprogrammed. Revenue and billing meters are usually held to a stricter standard and may be verified by the utility or by a laboratory, often with sealing. Microprocessor-based meters and protective relays add configuration and communication checks that electromechanical meters did not need, and an error in a setting can matter more than an error in accuracy. Transducers and meters feeding control systems are verified together with the receiving analog input. Battery systems add cell voltage and charger metering, and generator and transformer panels add temperature and power metering. In every case the procedure and tolerance follow the manufacturer and the governing standard for that equipment.
The equipment classes and standard sections that call for this test, and the UltraDb forms that carry it.
There is no single interval. It depends on the meter's role, the manufacturer's recommendation, the owner's maintenance program and, for revenue metering, the utility or regulator. Verification at commissioning is standard, and later checks are scheduled by criticality and by how the as-found drift has been trending.
If the reference has an uncertainty close to the tolerance of the meter, a good meter can appear to fail or a poor one to pass. A reference that is clearly better, and traceable to a recognized standard, keeps its own error small enough that the difference seen belongs to the device under test.
Only if the transformers are included in the loop. Secondary injection at the meter proves the meter alone, while checks with real load current or with primary injection include the transformer ratio, polarity and wiring. Many programs do both, because a wrong ratio is a common source of large errors.
Record the as-found readings first, then investigate the cause: wiring, ratio, settings or drift. Adjust or replace the device as the manufacturer allows, repeat the check and record the as-left values. Where billing or studies relied on the data, the owner may need to evaluate the effect of the earlier error.
UltraDb references ASTM, ANSI, IEEE, NETA, NFPA, NEMA, FAA, TIA, and IEC standards descriptively, to identify the published methods its forms are built from. UltraDb and Verlecta are not affiliated with, endorsed by, certified by, or licensed by NETA, ASTM, ANSI, IEEE, NFPA, NEMA, FAA, TIA, IEC, or Megger. Standard names and section numbers are the property of their respective organizations. Always test to the edition your contract specifies. Where a standard is sold through the ANSI webstore, Verlecta may earn a commission on a purchase made through our link, at no extra cost to you. A commission never decides which standards we cite or how a form is built.