How a Motor Draws What It Draws

Why this matters

A clamp meter on a motor lead is the most-used and most-misread instrument in the trades. The number it gives is not a health reading and it is not a load percentage. It is a measurement of how hard the shaft is being worked, at whatever voltage the motor is actually receiving, and it means nothing until you know that second number.

Motors get condemned over this weekly. A tech reads high amps, compares against the nameplate, and quotes a replacement, when the real fault is upstream and the replacement motor will read exactly the same on day one.

What sets the current: torque first, then voltage

An induction motor is a self-regulating machine. Put more load on the shaft and the rotor slows slightly relative to the rotating magnetic field. That increase in slip induces more current in the rotor, which draws more current through the stator. Take load off and the whole process reverses. Nothing in the motor decides how many amps to draw; the load decides, and the motor complies.

That makes a clamp meter a torque instrument you already own. It also means the reading is only interpretable against the voltage that produced it, because the same shaft load pulled at a lower voltage requires proportionally more current: the mechanical work is fixed, so as volts fall, amps rise.

Those two sentences are the whole article. Everything below is what happens when a reading is taken without the second one.

The gate

Per motor, at thermal steady state, under its normal running load. Conclude that the motor or its driven load is the problem only when both are true:

• supply voltage measured at the motor terminals while it runs is inside the nameplate's rated band - NEMA MG 1 establishes successful operation for general-purpose motors within plus or minus 10 percent of rated voltage, and some designs and some plates are narrower, so read the plate rather than assume the 10 percent, and
• measured current exceeds nameplate full-load amps multiplied by the nameplate service factor.

Both, not either. If the voltage condition fails, correct the voltage first and re-measure, because a current reading taken at low voltage compared against a nameplate written at rated voltage is a comparison between two different conditions. When the gate does trip, act in whole causes - a load change, a bearing, a misapplied motor - not in percentage adjustments.

Every reading below is taken on running equipment. That is energized work, permitted under 29 CFR 1910.333(a)(1) only where de-energizing would introduce additional or increased hazards or is infeasible due to equipment design or operational limitations, with diagnostic measurement as the qualifying case, and it requires a clamp and leads rated CAT III or better at the voltage present. Take the reading at an accessible point in the disconnect or starter enclosure rather than inside a machine guard: a rotating shaft, coupling or fan will take a clamp lead and your hand with it, and removing a guard to reach a conductor puts you in a mechanical hazard that would require isolation under 29 CFR 1910.147, which you cannot do and still be measuring a running motor.

Case A: 110 percent of nameplate and nothing to fix

Nameplate: 230 V, full-load amps 16.0 A, service factor 1.15. Service-factor amps are therefore 16.0 x 1.15 = 18.4 A.

Measured: 17.6 A at 231 V at the motor terminals, running, warm.

Voltage is 0.4 percent above nameplate, comfortably inside the band, so the first condition of the gate is satisfied and the reading is interpretable as taken. Current is 17.6 / 16.0 = 110 percent of full-load amps, and 17.6 is below the 18.4 A service-factor limit, so the second condition is not met. The gate does not trip. This is not a motor fault.

What it is, is a motor working above its nameplate rating within the allowance the manufacturer granted, and that allowance is narrower than it looks: NEMA MG 1 defines the service factor as applying at rated voltage and frequency, permits a higher temperature rise in the service-factor region than at rated load, and assumes the motor is in its rated ambient. Run there continuously, in a hot space, and you are spending insulation life to do it. So the correct output of this call is not a part, it is a finding: something is loading this machine above its rating, and the candidates are on the driven side - restriction, pressure, buildup, a damper or valve position, a belt tension - plus a written baseline so the next tech knows 17.6 A at 231 V is where this machine sat in August.

Case B: the same reading corrected, and the verdict reverses

Identical nameplate. Measured: 19.8 A at 208 V at the motor terminals.

The reading that gets motors replaced: 19.8 / 16.0 = 123.8 percent of full-load amps, well past the 18.4 A service-factor limit, so the motor is overloaded and needs replacing. That conclusion compares a current drawn at 208 V against a nameplate figure defined at 230 V. It is two conditions treated as one.

Run the gate properly. 208 V against a 230 V nameplate is (230 - 208) / 230 = 9.6 percent low, which fails the first condition. So correct before comparing. For a load whose torque requirement has not changed, current scales roughly with the inverse of voltage, so a motor at exactly rated load would be expected to draw 16.0 x (230 / 208) = 17.7 A at 208 V. Measured 19.8 A against that voltage-corrected expectation is (19.8 - 17.7) / 17.7 = 11.9 percent above rated load, not 23.8 percent above it.

Correct the supply and check the same reading from the other direction: at 230 V the same shaft load would draw about 19.8 x (208 / 230) = 17.9 A, which is 112 percent of full-load amps and sits under the 18.4 A service-factor limit. That is the same neighbourhood as Case A. The motor is not the fault; it is a normally loaded machine being fed 9.6 percent under its rating, and the extra current is the supply's doing.

The current it is drawing while the fault persists is not harmless. Winding loss goes with current squared, so at 19.8 A instead of the 17.9 A it would draw at corrected voltage, resistive loss in the windings runs about (19.8 / 17.9) squared = 22 percent higher. Sustained temperature above the insulation's rating shortens life on the widely used engineering rule that insulation life roughly halves for each 10 C of sustained excess rise. That is a rule of thumb rather than a manufacturer guarantee, and it is enough to say the fault is worth fixing this week rather than next season.

What flips the recommendation. All of the above assumes 208 V is a sag on a nominally 230 V supply, so the fix is upstream: a connection, a conductor, a tap, a shared feeder under load. If instead the building is a 208 V nominal system and someone installed a motor rated 230 V only, no amount of upstream work will help, and the correct repair is a motor rated for 208 V or dual-rated 208-230 V. Establish which situation you are in by measuring the supply with the motor off, at the panel and at the terminals, before you write either recommendation.

Why "percent of full-load amps" is not "percent loaded"

This is the most common quiet error in motor work. An induction motor draws magnetizing current whether it is loaded or not, and that component barely changes with load. Small and medium motors commonly sit somewhere around a quarter to a half of full-load current with nothing on the shaft, which means the current axis does not start at zero.

So a motor reading half its full-load amps is not half loaded. It is considerably less than half loaded, and any estimate that divides measured amps by nameplate amps overstates the load, always in the same direction.

Estimating actual load with slip

Slip gives a better estimate because it starts at zero. Percent load is approximately the measured slip divided by the full-load slip:

• Synchronous speed for a four-pole motor at 60 Hz is 1800 rpm.
• Nameplate full-load speed: 1750 rpm. So full-load slip is 1800 - 1750 = 50 rpm.
• Measured shaft speed: 1768 rpm. Measured slip is 1800 - 1768 = 32 rpm.
• Estimated load = 32 / 50 = 64 percent.

Compare that against what the current suggested. The same motor was clamped at 12.0 A against 16.0 A full-load, which by the naive ratio reads as 75 percent loaded. Cross-check it with the no-load current: this motor draws about 6.0 A with the coupling off, so interpolating between 6.0 A at no load and 16.0 A at full load gives an expected 6.0 + 0.64 x (16.0 - 6.0) = 12.4 A at 64 percent load, which is within a few tenths of the 12.0 A actually measured. The slip estimate and the current estimate agree once the magnetizing current is accounted for; the naive ratio was high by 11 percentage points on a machine nobody would have questioned.

Two caveats belong with the method rather than after it. Slip estimation needs a genuinely accurate speed reading, so use an optical tachometer rather than a contact type, and apply the reflective target while the machine is isolated and locked out under 29 CFR 1910.147 for the unexpected-startup hazard, not while it turns. And the method loses accuracy below roughly half load, where the slip numbers get small enough that measurement error dominates.

What the amp reading will not tell you

It does not report efficiency, which needs real power. It does not report power factor, which needs voltage and current sampled together. It does not report bearing condition except indirectly and late, once drag has grown enough to move the shaft load. It does not report winding insulation condition at all, because an ohmmeter-level or operating-level current says nothing about a path that only conducts at higher voltage.

And it does not report temperature. Do not judge a motor's frame temperature with your hand: a motor at full load in a warm space can exceed the burn threshold on a surface that looks unremarkable. Use a non-contact thermometer, and compare against the nameplate's rise and ambient rating rather than against how hot motors usually feel.

Verifying a load reading before you act on it

Take voltage and current at the same point, at the same moment, at thermal steady state, on all supply conductors rather than one. Write down all of it, including the voltage, because a current reading recorded without its voltage is unusable to the next person and to you in six months.

Then run one arithmetic check on yourself: multiply the corrected current by the voltage ratio you applied and confirm you land back on the raw reading. If the corrected figure and the raw figure cannot be converted into each other, one of them has the wrong voltage attached to it, and that is the error that reverses a verdict.

References

• NEMA MG 1 for rated voltage tolerance, service factor definition and the associated temperature-rise allowance
• 29 CFR 1910.333(a)(1) for the energized-work justification governing live current measurement, and 29 CFR 1910.147 for isolation against unexpected startup of rotating equipment
• Motor manufacturer documentation for nameplate full-load speed, no-load current and permissible ambient
• See related: What Power Factor Actually Costs a Shop; Why Inrush Current Trips Things That Should Hold; How to Use Voltage Drop to Find a Bad Connection