Monitoring Shows Zero Production But Utility Meter Runs Backward

Why this matters

When the monitoring portal reports zero production but the utility's net meter is clearly running backward (exporting), the system is generating power - the customer is getting paid - and the problem is in the MEASUREMENT and COMMUNICATION layer, not the array. This is a reassuring diagnosis to deliver, but it must be confirmed, not assumed, because a real zero-production fault that coincides with a separate metering quirk can hide behind the same symptom. The fault is almost always a production CT (current transformer) miswire, a comm/gateway dropout, or a monitoring-account misconfiguration. This tree separates a measurement artifact from a genuine outage.

Symptom presentation

The monitoring app shows 0 W or a flat zero line for production, often with a "no data" or "offline" indicator on the gateway/combiner, while the utility meter (or the customer's own confirmation) shows net export - the meter dial reverses or the smart meter shows negative kWh. The inverters themselves may show green status lights and the array is in full sun. The customer says "the app says nothing is happening but my bill shows I'm sending power to the grid."

Quick checks

Confirm production at the inverter before touching the monitoring layer. Read the inverter's own display or local interface (not the cloud portal):

• Inverter local display shows real AC output (e.g., 5.2 kW) but the cloud shows zero: the array works; the fault is in monitoring/comm. Go to the comm/CT branch.
• Inverter local display ALSO shows zero, yet the meter runs backward: contradiction - re-verify the meter direction and confirm it is THIS system exporting (not a neighbor's, not a meter read error). A truly producing system shows output at the inverter.

Then check the gateway/communication device status and the production CTs (consumption/production metering on the gateway or a separate revenue-grade meter).

Isolation tree

Branch A - Communication / gateway dropout. The inverters produce and the meter exports, but the gateway (Enphase Envoy/IQ Gateway, SolarEdge inverter Wi-Fi/cellular, or a separate datalogger) lost its uplink, so the cloud shows zero or "offline." Confirm: inverter local output is healthy and the gateway shows no internet/cell connection or a stale last-report time. Causes: dead Wi-Fi, changed router password, failed cellular module, unplugged gateway, or ISP outage. Remedy: restore the gateway's connection and confirm the portal backfills.

Branch B - Production CT miswire or failure. Systems that meter production via clamp-on CTs report zero if a CT is reversed, on the wrong phase, clamped on the wrong conductor, unplugged, or failed. Confirm: production reads zero on the portal while consumption metering and inverter output are normal, or production reads negative/backward. A CT installed backward reads production as zero or negative. Remedy: verify CT orientation (arrow toward the load/grid per the manufacturer), correct phase association, and seating; re-commission the metering.

Branch C - Monitoring account / commissioning error. The system was never fully commissioned in the portal, the inverters/microinverters were not all provisioned, or the array map is incomplete, so the portal shows zero even though hardware reports. Confirm: the device list in the portal is empty, partial, or shows un-provisioned units while hardware is live. Remedy: complete provisioning/commissioning, add missing serial numbers, and finalize the array configuration.

Branch D - Partial fault masked by net export. A subset of the array could be down while the rest still exports enough to spin the meter backward, AND a comm issue zeros the portal. Confirm by reading EACH inverter/microinverter locally: if some report and some do not, there is a real per-device fault layered on top of the comm issue. Remedy: address the down device(s) separately from the comm fix.

Branch E - Meter reading misinterpretation. Rare but worth ruling out: the "backward" meter is actually the consumption register, a demand register, or a multi-meter site read on the wrong meter. Confirm the meter and register actually correspond to this PV system's export. Remedy: confirm the correct meter; if the system truly is not producing at the inverter, pivot to a no-production diagnosis (DC disconnect open, AC breaker off, inverter fault).

Confirming diagnosis

The single confirming read is the inverter's LOCAL output versus the cloud portal. Local inverter shows production + meter exports + portal shows zero = measurement/comm artifact, and the array is fine. Local inverter shows zero = the array is genuinely down and the metering story needs re-checking. From there, gateway connection status separates comm dropouts (Branch A) from CT/commissioning issues (Branch B/C), and a per-device local read catches a partial fault hiding behind net export (Branch D). Never close this as "just a monitoring glitch" without the local inverter read.

Remediation

Restore gateway connectivity (Wi-Fi credentials, cellular module, power, ISP) and confirm the portal resumes and backfills data. Correct any reversed or mis-clamped production CT per the manufacturer's orientation and re-commission the metering. Complete portal provisioning and the array map so every device reports. Repair any genuinely down inverter/microinverter found in the per-device read. Verify the correct utility meter and register. After the fix, confirm the portal production matches the inverter local output and tracks the meter's export over a full sun day.

References

• Enphase IQ Gateway and SolarEdge monitoring documentation (gateway connectivity, production/consumption CT installation and orientation, commissioning).
• NEC Article 705 - Interconnected Electric Power Production Sources (interconnection and metering context).
• ANSI C12 / utility net-metering bidirectional meter references (export registration, register identification).
• Manufacturer commissioning guides (Enphase, SolarEdge) for device provisioning and revenue/production metering setup.