A Motor's Insulation Resistance Test Can Pass Every Year and Still Be Telling You It Is About to Fail

Most ETPs and STPs run periodic insulation resistance tests on pump, blower, and mixer motors with a megohmmeter, commonly called meggering after the instrument maker whose name became the verb. The result gets checked against one number: a minimum acceptable resistance, often one megohm per kV of rated voltage plus one megohm as a conventional rule of thumb, or whatever threshold the plant's maintenance standard specifies. If the reading clears that minimum, the test is logged as a pass and the motor goes back into service until next year. That single-threshold check answers a real question, whether the winding insulation is safe to energize today. It does not answer the question that actually predicts whether the motor is about to fail, because a winding's insulation resistance can decline sharply, year after year, while still clearing the same minimum threshold every single time, right up until the year it does not.

A pass against a threshold and a healthy trend are different claims

A spot insulation resistance reading is a point-in-time measurement against a fixed bar. It says nothing about direction, because a single number carries no history. A motor whose insulation resistance reads 1,800 megohms this year, then 1,500, then 700, then 260 over four consecutive annual tests has failed nothing by the standard that only checks a minimum, since even 260 megohms can sit comfortably above a typical pass threshold. The same four readings, looked at as a trend rather than four isolated passes, describe a winding whose insulation is deteriorating at a rate that should have triggered a closer look well before the fourth test, because each prior reading was itself a warning the threshold-only check had no way to register.

Why the pass-fail habit misses what it was built to catch

Grading insulation resistance against a single minimum is not a wrong practice on its own terms, it is answering a genuinely important question: is this motor safe to run right now. The trouble is that the same test, repeated annually and logged only as pass or fail, produces exactly the data needed to answer the harder, more useful question, and that data gets thrown away the moment each reading is reduced to a checkbox. A moisture-ingress pattern, a slow insulation breakdown from heat cycling, or a seal degradation letting process liquid reach the windings, all of these show up as a declining trend long before any single reading crosses the pass-fail line. By the time a declining trend finally does cross that line, the deterioration is usually advanced enough that an unplanned trip or an outright winding failure is closer than a clean maintenance window.

This is the same shape of mistake we described in the piece on arc flash study staleness: a value that passes an individual, correctly specified check at the moment it is checked can still be heading somewhere the check itself was never built to notice, because the check only ever asks whether today's number clears today's bar. There, the number that mattered was fault current drifting against a study last updated years earlier. Here, it is insulation resistance drifting against its own history. In both cases the record shows nothing but passes right up until it doesn't, because passing was always the only thing being recorded.

Picture a plant running a ten-year-old blower motor on a quarterly or annual meggering schedule, the blower being one of several driving the aeration system that keeps the biological process alive, every reading logged, every reading clearing the plant's minimum threshold without exception. Nobody plots the four or five most recent readings against each other, because the maintenance log only has a column for pass or fail, not a column for percent change since last test. The motor's insulation resistance has in fact been falling by roughly half every year for the last three tests, a pattern consistent with the moisture or thermal degradation documented in published predictive-maintenance case studies on motors read this way, and nothing in the plant's own records would surface that pattern unless someone specifically pulled the historical values and looked at them side by side, which the pass-fail logging habit gives nobody a routine reason to do.

An aeration blower failing without warning is a worse event than most motor failures elsewhere on site, because it does not just cost a rewind or replacement, it risks the biological process the whole plant depends on if dissolved oxygen drops before a spare can be brought online. That asymmetry is exactly why the motors most worth trending, rather than just pass-fail testing, are usually the ones already identified as critical to uptime in a plant's own energy and criticality audit, not every motor on the single-line diagram with equal priority.

What catching the trend actually requires

The fix does not replace the threshold check, since a safe-to-energize answer is still a necessary one. It adds a second, cheap step that most plants already have the raw data for: trending each motor's insulation resistance readings over time rather than discarding everything except the pass-fail verdict, and treating a sharp year-over-year drop as a maintenance flag independent of where the absolute number sits relative to the minimum. A complementary test, the polarization index, compares the ten-minute reading against the one-minute reading from the same test and is largely self-correcting for winding temperature, which a raw spot reading is not; a ratio that has drifted down toward 1 across successive tests points at moisture or contamination even when the one-minute reading alone still clears the plant's threshold comfortably. Neither of these requires new instrumentation beyond the megohmmeter the plant already owns. They require keeping the historical numbers somewhere a maintenance planner actually looks, rather than archiving them as a stack of individually passed test certificates that nobody is ever asked to compare against each other.

None of this is a large programme to stand up. A spreadsheet that carries forward each motor's last three or four readings alongside the current one, reviewed at the same time the test itself is logged, catches the pattern a pass-fail column cannot. The cost of building that habit is a few minutes per motor per test cycle. The cost of not having it is discovering the trend only in hindsight, after an unplanned trip, when the maintenance log already contains every number that would have predicted it and nobody had ever looked at them side by side.

If your plant's motor testing programme is still logging insulation resistance as pass or fail without plotting the trend behind each pass, that is a gap worth closing with Spans' electrical and instrumentation team before a motor that passed every test on record fails on the one day nobody was trending it.