Revenue Meter Accuracy - Field Test

What it is

The field accuracy test of an ac watthour revenue meter records how closely one meter, self-contained or instrument-transformer rated, registers the energy that flows through it, as found and, when it was adjusted, as left, against the in-service limits of the metering code ANSI C12.1-2026. The meter is compared with a field standard at three test loads: full load and light load at unity power factor and a power factor point at 0.5 lagging (0.707 lagging on a polyphase delta service), each reported as percent registration, with the average percent registration computed by the method the commission or program prescribes. The record also carries the field standard's verification date and certificate error, the four inspection items checked before the meter is placed or returned in service, a register test against a reference reading, and the demand or interval register test where one is fitted. Instrument transformers are not tested here: their ratio agreement is graded on the linked Instrument Transformer, High-Accuracy - Acceptance Test record, and acceptance tests against manufacturer accuracy data belong on the metering devices acceptance forms.

Why it is performed

A revenue meter is the record of how much energy a customer consumed in a billing period. A meter that registers high overcharges the customer; one that registers low understates the utility's revenue, and either error repeats on every bill until the meter is tested. A field test against a verified standard shows whether the meter's registration has drifted outside the in-service band, whether an earlier adjustment is still holding, and what the meter was reading before anything was touched, because the as-found result stays on the record as a finding even after adjustment. Grading the field standard itself - its verification age and its certificate error - keeps the judgment about the meter resting on a reference that is traceable, and recording the program context, the reason for test and the sampling basis lets the owner show a commission why this meter was tested and on what plan.

When it is performed

The reason for test is recorded as one of new device acceptance, periodic, variable interval, statistical sample, customer request, suspected defect or removed from service, with the test program, statistical basis and acceptable quality limit where a sampling plan applies. The interval is the utility program's: the metering code's periodic schedule (8 years for AMI meters and 16 years for other periodic meters) is an informative guide, so time since the last test beyond it raises an issue and never changes the verdict. A meter taken out of service may be retired untested, and a meter under an in-service test plan may go back into service untested; both dispositions are recorded. Acceptance tests of a new meter against the manufacturer's accuracy data are recorded on separate forms; this form records the field accuracy test as found and as left against the in-service limits of ANSI C12.1-2026, and a utility commission or tariff that sets different limits governs through the program document recorded on the form.

How it is typically performed

The field standard is identified with its certificate and traceability, whether it is used as a portable or a reference standard, the date of its last accuracy verification (graded against 365 days before the test date), its error at reference conditions from the certificate and, where stated, its error over the voltage and current range. Before the meter is placed or returned in service, four inspection items are answered: the meter undamaged and fit for service, the socket or enclosure undamaged, the connections correct as installed and the installation inspected for safe and accurate operation; the test switch or bypass position and the seal numbers as found are recorded. For an instrument-transformer rated meter the utility record ratios, the ratios programmed in the meter, the register constant and any transformer-loss compensation are recorded and the instrument transformer result is linked. The as-found test applies full load (test amperes, or the CT secondary rating as the base), light load at 10 percent of the base and the power factor point, recording the applied current, the power factor and the percent registration at each; the form computes the load as a percent of base, the percent error and the average percent registration by Method 1, 2, 3 or 4. If full or light load lies outside the band the meter is adjusted where possible or removed from service, and the as-left test repeats the same loads. A register tested for other than accuracy compares the register energy over the test with the reference energy. An electronic demand or interval register records its program validation by time run, its test current as a percent of full-load test amperes, its watthour registration error and its interval timing error over one week; a mechanical demand register records up to three points between 25 and 100 percent of full scale with their errors, any adjustment with the as-left error, and the time-of-day element error over one week. The record closes with tester, date, seals as left, disposition, ambient temperature, comments and deficiencies.

What gets recorded

Identification: manufacturer, model, serial and meter numbers, designation, meter type, form and service, accuracy class (recorded only), test amperes, rated voltage, elements, firmware, and AMI or ISO metering above 20 MW. Program context: reason for test, program and statistical basis, acceptable quality limit, last test date and the time since. Field standard: identity, use as portable or reference standard, verification date, certificate and traceability, error at reference conditions, error with voltage and current variation and whether a performance record is kept. Inspection: the four inspection items, test switch or bypass used and position, and seal numbers as found. Transformer-rated meters: CT and VT ratios on the utility record, ratios programmed, register constant, loss compensation and the linked instrument transformer result with its verdict. As found and as left: full-load base and CT secondary rating, and for full load, light load and the power factor point the applied current, power factor and percent registration with the computed load percent, error and margins; polyphase delta, the average registration method and Method 3 load point; the adjustment answer and what was adjusted. Register and demand: reference and register energy, demand register type, program validation, electronic demand test current, registration error and interval timing error, mechanical demand points, errors, adjustment and as-left error, and time-of-day error. Record of test: tester, date, seals as left, disposition, ambient temperature, comments and deficiencies.

How results are evaluated

The verdict uses the as-left results when the meter was adjusted and the as-found results otherwise; the as-found test stays on the record as a finding, and a meter marked adjusted without as-left readings waits for them. To pass, the field standard must have been verified within 365 days of the test date and its certificate error at reference conditions must be at most 0.05 percent for a portable standard or 0.02 percent for a reference standard; every inspection item must be satisfactory; each test load must sit inside its band (full load and the power factor point from 97 to 103 percent of the base, light load from 7 to 13 percent of test amperes); the average percent registration must lie from 98.0 to 102.0 percent, ends included, and after an adjustment the as-left full-load and light-load registrations must each lie in that band as well. A meter whose full-load or light-load registration lies outside the band must be adjusted where possible or removed from service; one left in service unadjusted fails, and a missed band with no adjustment answer leaves the record indeterminate. Where fitted and tested, the register error against the reference reading must be within 2 percent for retail metering or 1 percent for ISO metering above 20 MW, an electronic demand register follows the same watthour limit with interval timing within 15 s over a week and its program on a new device validated by a time run or other test, and a mechanical demand register must be within 4 percent of full scale at each point tested between 25 and 100 percent of full scale. Issues that never change the verdict: a field standard variation error outside the code table, a test-point power factor more than 0.05 from the specified value, a mechanical demand error above the 2 percent adjustment trigger (an UltraDb reading), time-of-day drift over 2 minutes a week, time since the last test beyond the informative guide, and a missing instrument transformer link or a linked ratio verdict of Do not agree. The meter accuracy class never changes the grading, and an entered load, register or demand row that cannot be graded leaves the record indeterminate rather than reading as untested.