Meter Installation Mistakes: A Field Checklist

An applied extra to Cobler's Electricity Fundamentals course.

The meter powered up, Modbus came alive on the first poll, and the dashboard filled with numbers before lunch. Three phase voltages, three currents, a running kW figure, a power factor. Every register the contractor mapped is answering. It looks commissioned. Six months later it becomes a ticket that reads "the platform's numbers look wrong," and by then the contractor is off site and the trail cold. Here is the trap every energy meter installation hides: a wrongly wired meter communicates perfectly. The meter looks normal. The data may not be.

This is a checklist to run before you sign practical completion, one anyone on the project can use, not only the person who held the screwdriver. It builds on how to spec a CT; here we read the data the meter produces and ask whether the physics behind it is possible.

Why does a wrong meter installation still look commissioned?

Because "talking" and "correct" are different tests, and only the first is automatic. A live Modbus map proves the meter is powered and the register addresses are right. It says nothing about whether the CTs are on the correct cables, facing the correct way, landed on the matching phase, or programmed with the correct ratio.

The errors are catchable because of arithmetic. An electronic meter computes power one phase at a time as P = V x I x cos(theta), where theta is the angle between that phase's voltage input and its current input. Every wiring mistake below is really an error in which voltage the current is measured against, or in the sign or scale of the current, so they surface as impossible numbers that anyone who knows a normal building can read.

Why is the power factor stuck at 0.30 on an ordinary load?

A steady, implausibly low power factor that does not match the load is the classic signature of a CT-to-phase association error. The CT clamped on the red phase conductor has been landed on the meter's yellow phase current input, so the meter multiplies red's current against yellow's voltage. Those two sit about 120 degrees apart, so a load running near unity is computed as though the current carries an enormous phase lag. The offset masquerades as reactive power that was never there.

A normal mixed commercial load sits around 0.80 to 0.95. A rock-steady 0.30 to 0.50 that does not budge when equipment starts and stops is a wiring accusation, not a capacitor problem, and often the three per-phase figures disagree. The exact decimal depends on which pair got crossed (cos 60 degrees is 0.50, cos 120 degrees is minus 0.50), so trust the pattern rather than a magic number: implausibly low, steady, out of step with the load. A phase crossed a full 120 degrees can even compute negative, since its cosine is minus 0.50, so a bad association sometimes trips the negative-kW flag below as well. The fix is to remap the current inputs so each CT lands on the same phase as its voltage reference, then confirm on the phasor screen.

Why is the meter showing negative kW when the site has no solar?

Negative kW with no generation on site is physically impossible, so it is a pure wiring verdict: a reversed CT. The split-core arrow, or the P1 marking, points toward the load instead of the source, or the S1 and S2 secondary leads are swapped at the meter. Either way the meter sees the current 180 degrees from reality, so the cos term flips the sign of P and it believes power flows backward.

The data tells on it cleanly. kW reads negative on the affected phase, its power factor reads negative, and the energy register counts down instead of up under load. Reverse one CT and one phase goes negative while total kW is deflated; reverse all three and the whole meter reads negative. Schneider's own PowerLogic guidance names the cause plainly: the CT fitted the wrong way round (P1 should face the supply and P2 the load), or the secondary wiring crossed (Schneider FA86375). Flip the CT so the arrow faces the source, or swap the secondary leads, then confirm kW is positive and the register climbs.

Why do the submeters add up to more than the main incomer?

Because a child in the metering tree cannot legitimately exceed its parent. Every submeter draws its current through the incomer, so the children must sum to a few percent under the parent, never over; the legitimate gap is metering tolerance, un-submetered loads and distribution losses. A sum that overshoots is a topology or ratio fault, not measurement noise.

Three causes produce an overshoot: a submeter's CT clamped on a cable already counted by another (the same current billed twice), the wrong CT ratio programmed on one meter so it over-reads (see below), or a CT on the wrong cable entirely, catching a bigger feeder than intended. Walk each clamp against the single-line diagram, confirm no cable is double-counted, and re-run the sum. This is the same reconciliation discipline that explains why a submeter and the TNB meter never match exactly, applied one level up inside your own board.

The scale error that hides: wrong CT ratio programmed

This is the hardest fault to catch because the data looks plausible, just scaled. If the meter is set to 800/5 but the installed CT is 400/5, every amp, kW and kWh on that meter reads exactly double the truth; program a lower ratio than the CT actually fitted and it reads low instead. Power factor stays correct, because a scaling cancels out of a ratio, so nothing on the screen looks wrong in isolation.

The phasor screen is blind to it, because only the magnitude is off. The only way to catch a clean 2x or 0.5x scale error is against a reference: clamp the actual cable and compare amps, or lean on the submeter-versus-parent sum above. Read the CT nameplate, set the meter's CT primary to match, and re-verify under a known load.

The quieter faults worth one pass each

Four more errors produce a missing or misconfigured reading rather than an impossible one.

One phase reads near-zero current while its voltage is normal. The CT shorting block was left closed after wiring, so the secondary circulates through the short instead of into the meter. Voltage present, current absent, total kW down by about a third on a balanced load. Open the shorting block into the run position.

One phase voltage reads zero while its current still flows. The voltage-reference fuse on that phase is blown or the sense lead is off. With no voltage to multiply against, that phase's kW vanishes though current is present, the inverse of the case above. Replace the fuse after isolating the reference.

Total kW and PF do not reconcile though each phase looks fine. The meter's wiring mode (four-wire versus three-wire) does not match the install, so it applies the wrong internal formula. Set the mode to match reality and re-verify against a clamp meter.

One phase reads low or flickers under a steady load. A split-core CT has not fully clicked shut, and the air gap weakens the coupling. Re-seat it, clean the mating faces, and latch until it clicks.

What does healthy meter data look like at commissioning?

Write the baseline down before you hunt faults. Every line below should be true at a healthy handover, and each one is a check you can tick on the spot:

- Voltages: all three phases present and balanced, about 230 V phase to neutral on a Malaysian LV four-wire system, none reading zero.
- Currents: all three present and plausible for the running load, matching a clamp spot check.
- kW: positive on every phase and in total, with the energy register counting up.
- Power factor: in the sensible 0.80 to 0.95 band, not a steady 0.30, not negative, not above one, and the three per-phase figures agreeing.
- Submeters: each sums a few percent under its parent, never over.
- Phasor screen: each current arrow hugging its own voltage, the whole set 120 degrees apart in correct rotation, three tidy fans, none reversed.

The cross-check ritual: angles, magnitudes, totals

No single instrument catches everything, so run three checks in order. The phasor screen first, which most Class 0.5S revenue meters provide: a crossed phase parks a current arrow beside the wrong voltage, a reversed CT flips one 180 degrees, a lost input drops an arrow. Then a clamp-meter magnitude check on the actual cable, because the phasor screen is blind to a ratio-scale error. Then a utility-meter reconciliation under a known load, so the whole tree ties back to TNB's figure. A live register map is none of the three; it only proves the meter is talking.

Safety non-negotiables

Never open an energised CT secondary. With the primary live, an open secondary drives the core into saturation and induces lethal kilovolts across the terminals. Short before you touch: short the secondary at the shorting block before disconnecting any meter or lead, because a shorted CT secondary is its safe idle state. Isolate and verify the voltage-reference fuses dead before working on voltage-sensing wiring. All of this is competent-person work under the Electricity Supply Act and Suruhanjaya Tenaga regulations, and the full open-secondary explanation lives in the CT spec guide.

Sign nothing until the data matches reality

Check the metering data against physical reality before you sign practical completion, because a live Modbus map says nothing about whether the CTs are on the right cables, the right way round, on the right phase, at the right ratio. Run the three checks, confirm the healthy baseline, then sign. Skip it and a 30-second check at handover becomes a wrong-numbers ticket next year, once nobody remembers which cable the yellow CT ended up on.

This is where a commissioning view earns its keep. A CobiNeural deployment surfaces per-phase kW, per-phase power factor and a live submeter-versus-parent ledger, so a negative kW tile, a stuck-at-0.30 power factor or a child exceeding its parent jumps out to anyone on the project in minutes. Cobler's own deployment checklist runs exactly these checks before a site is accepted.

This is an applied extra to Cobler's Electricity Fundamentals course. It builds on how to spec a current transformer and how electricity meters work, where the CT and the per-phase power calculation first appear.

Want your meters commissioned so their numbers survive an argument with TNB's bill? Book a demo and we will show you the checks we run before we accept a site.