What Did the UPS Really Deliver? Verifying Protected-Load Power

A UPS exists for one purpose: to keep a disturbance on its input from reaching the load on its output. Yet most facilities can’t actually prove that it did. The UPS reports an input event; the team cannot tell whether the protected load saw anything. Or the monitoring shows output power higher than input power, and someone files it as “impossible data”.

This guide explains how to measure both sides of a UPS on one timeline, how to read the result, and the arithmetic traps that make healthy UPS systems look broken.

The question to answer

When the input is disturbed, what actually happens at the protected output?

The measurement plan

1. Label the two channels explicitly as UPS input and UPS output. Do not rely on an assumed A/B convention — a swapped label reverses every conclusion.
2. Match phases, probe ratios, polarity and timestamps on both sides.
3. Record voltage, current, active power and events at input and output.
4. Add UPS telemetry where the UPS exposes it — operating mode, battery state and alarms — through a validated interface and register map for that specific UPS.

Inspector V4’s two isolated channels are designed for this paired measurement, with event capture on both sides and protocol support (Modbus, SNMP and others) for reading UPS telemetry.

Reading the evidence

Compare the protected output during the same event window as the input disturbance, then place that event in the UPS operating mode at the time.

• An input dip with a clean output: the UPS did its job. Keep that record — it is evidence too.
• An input dip that also appears at the output: investigate against the load’s tolerance, the UPS alarms and its transfer history.
• An output disturbance with no input event: look downstream or inside the UPS.

When output looks bigger than input

Battery operation is the usual explanation. During a disturbance, the battery can supply part of the load while AC input falls; when the battery recharges, input power rises without any change in output load. The power balance is:

AC input + battery discharge = load output + internal losses + battery charging

During ordinary operation the battery term is either charging or discharging, not both.

Three checks before reporting an efficiency anomaly

1. Physical points. Do the input and output channels really represent the intended points? Are phase, polarity and ratios correct?
2. Same time, same averaging. Do the readings share a timestamp and an averaging interval? Compare event records with event records — never one transient sample with an unrelated 10-minute average.
3. Mode and load. What did the battery and mode telemetry say? Low-load efficiency carries much more uncertainty than full-load efficiency.

Report results as a defined test condition, not as a universal efficiency value for that UPS. Vendor-specific operating modes and register maps must be verified against that UPS’s documentation.

Acting on the evidence

• Investigate any output disturbance against the load’s tolerance, UPS alarms and transfer history.
• For efficiency checks, use matched steady-state active power and document battery conditions and auxiliary loads.
• Test approved operating scenarios with the UPS owner — never trigger an unscheduled transfer just to collect data.
• Keep a record of input events that were successfully isolated from the load, as well as those that were not.

What to measure to show value

• the number of output disturbances,
• unexplained transfers,
• response time to UPS alarms,
• how repeatable matched input/output measurements are.

Related reading

• Why production lines stop “for no reason” — including generator changeovers
• Harmonics and poor power factor
• Equipment fault or site power?