Current IntelPower LibraryShop main store

Comparison

Insulation resistance testing from switchgear vs motor terminals

Motors and motor starters Published 7 min read

Short answer

Test from the switchgear with the cable connected for routine trending against a commissioning baseline taken in the same configuration; the cable is a parallel path, so the connected reading sits below the motor alone and moves slowly as the winding degrades. When the trend falls or the decision is about the motor itself, break the circuit and test the motor and cable separately.

On this page
  1. What the switchgear-end reading includes
  2. The standards behind the numbers
  3. Worked case: a 4160 V motor at 1000 MΩ behind a 100 MΩ cable
  4. Choosing the test point
  5. Building the baseline at commissioning
  6. Splitting the test safely
  7. Questions asked

Do both, on a rule. For routine checks, test from the switchgear with the cable connected and trend that number against a commissioning baseline taken in exactly the same configuration; the cable is a parallel path, so the connected reading always sits below the motor alone, but as a like-for-like trend it is stable and comparable. When that trend falls, or when the decision is about the motor itself, break the circuit at the terminal box and read the motor and the cable separately. IEEE Std 43-2000 gives you the minimums; the trend gives you the warning. The minimum alone will not.

What the switchgear-end reading includes #

A DC insulation-resistance tester (a megohmmeter) applied at the switchgear with the motor connected sees every insulation path from conductor to ground in that circuit: the cable's and the winding's. Those paths are in parallel, so their conductances add:

1 / R_combined = 1 / R_motor + 1 / R_cable

The combined value is always lower than either component alone, and lower than the cable alone. That bound decides arguments further down.

When the cable reads an order of magnitude below the motor, the cable sets the number: a healthy winding cannot lift the reading above the cable's contribution, and a degrading winding barely moves it until the winding falls to the cable's order of magnitude. The connected test is a circuit-level screen. It tells you whether a low-resistance path exists somewhere in the cable-plus-motor circuit; it cannot tell you which component changed, and it is structurally late at catching winding decline. That is the whole disagreement in one equation: the maintenance department is testing the circuit, and you want to test the machine.

The standards behind the numbers #

IEEE Std 43-2000, the recommended practice for insulation-resistance testing of rotating machine windings, is the basis for the figures in this argument. The classic rule of thumb it superseded is 1 MΩ per kV of winding rating plus 1 MΩ: a 460 V motor computes to 0.46 + 1 = 1.46 MΩ, and at 4160 V the same rule gives 4.16 + 1 = 5.16 MΩ. The 2000 edition raised the floor to 100 MΩ for new stator windings rated 2300 V and above — the temperature-corrected figure your program is quoting.

ANSI/NETA ATS-2025 governs acceptance testing of electrical power equipment for initial energization, which is where the commissioning measurements your trends compare against come from. Both bodies treat the IR reading as a trending quantity: the value earns its meaning logged over time under a consistent method, not from a single pass. Confirm the requirements your program applies with the authority having jurisdiction and the machine and cable documentation.

Online protection cannot watch insulation: current-only schemes see what the motor draws, not what the winding is becoming. That blind spot is why motors burn up with no overload trips on 480 V. What current-only motor protection can and cannot detect sets out the same limit from the relay side, and the logged IR test is what covers it offline.

Worked case: a 4160 V motor at 1000 MΩ behind a 100 MΩ cable #

Inputs, all illustrative but typical of the machines in this argument: a 4160 V form-wound motor whose winding reads 1000 MΩ at commissioning; a power cable reading 100 MΩ under the same test conditions; an acceptance figure of 100 MΩ temperature-corrected, per the IEEE Std 43-2000 minimum for new windings rated 2300 V and above.

Rule in words: the connected circuit presents the two insulation paths in parallel, so conductances add, and the combined resistance is the reciprocal of that sum.

Arithmetic, healthy machine:

1/R = 1/1000 MΩ + 1/100 MΩ = 0.001 + 0.010 = 0.011 per MΩ R = 1/0.011 = 90.9 MΩ

Arithmetic, motor degraded to 250 MΩ (a 75% fall from its own history):

1/R = 1/250 MΩ + 1/100 MΩ = 0.004 + 0.010 = 0.014 per MΩ R = 1/0.014 = 71.4 MΩ

Result: the winding lost 75% of its insulation resistance; the switchgear-end reading fell from 90.9 to 71.4 MΩ, about 21%. Both connected readings sit below the 100 MΩ figure even with the winding at 1000 MΩ, because the connected reading can never reach the cable's 100 MΩ.

Decision: on any circuit whose cable-only baseline is at or below the acceptance figure you apply to the motor, the connected test cannot administer that figure — stop using it for pass/fail on that circuit and read the motor at its terminals every cycle. For routine trends on circuits with healthy cable margins, set the action trigger at a repeatable fall of 20% from the same-configuration baseline: in this case a 75% winding decline arrives looking like 21%, so a looser trigger lets a badly degraded motor read as fine and still trending.

Choosing the test point #

Configuration What the reading includes Commit it to What it cannot do
Switchgear end, cable and motor connected Cable and winding in parallel Routine trending against a combined commissioning baseline, same test voltage and connections Read at or above the cable's own value; localize a change to motor or cable
Motor terminals, cable disconnected Winding insulation only Motor condition decisions: rewind, continued service, test interval changes Say anything about the cable
Cable alone, both ends isolated Cable insulation only Localizing a low combined reading; cable condition assessment Say anything about the motor

Run the connected test as the periodic screen. It needs no work at the motor, and against a like-for-like baseline it will announce a real change in the circuit — in the worked case, as a 21% fall.

Split the test when any of these hold: the combined reading has fallen 20% or more from its same-configuration baseline; the combined reading is below the applicable minimum; you need a motor-specific number for a rewind-or-continue decision; or the cable-only baseline is at or below the motor acceptance figure, which disqualifies the connected test outright.

Never trend across configurations. A connected reading compared against a motor-only baseline shows a permanent degradation that is only parallel arithmetic. And the cost argument points the same way as the physics once you count consequences: the motor is the component you actually replace, and its early warning lives only in the motor-only number.

Building the baseline at commissioning #

Take three readings at commissioning and write the configuration beside every number: the motor alone with the cable disconnected; each power cable run alone, both ends isolated; and the cable-plus-motor assembly connected, tested from the switchgear. That is the established commissioning practice for this equipment, and it is what makes the periodic connected test meaningful — your periodic screen compares to the third value only.

Temperature-correct every reading by the standard's method before logging it, and hold the test voltage and connection scheme fixed between readings. If the commissioning file holds only a combined value, it is still a valid baseline for future combined readings, but it is not a motor history; a large departure from it is a trigger to split the test, not a diagnosis of which component moved.

Splitting the test safely #

The motor-terminal test is an open-panel job. Plan it under your site's permit process, and open the terminal box only after the circuit is de-energized, locked out, and tested for absence of voltage — the safe work practices NFPA 70E, 2024 edition covers. A winding just tested with DC holds charge: discharge and ground it per the tester's procedure before touching a lug.

The sequence:

  1. Isolate the circuit and prove it de-energized at the point of work.
  2. Break the terminations at the motor terminal box.
  3. Test the motor at its terminals and the cable from the switchgear with the far end isolated, each against its own baseline and criteria.
  4. Remake the joints and restore the insulation system by the site-approved method.

Where a machine needs repeated access — the ones you are babysitting — remake its joints once with voltage-class motor pigtail splice kits built to be opened and remade, instead of rebuilding heat-shrink medium-voltage terminations at every test; on a 4160 V circuit that means 5 kV-class hardware.

Questions asked

Our switchgear-end reading meets 100 MΩ temp-corrected. Is the motor healthy?

It proves no low-resistance path anywhere in the cable-and-motor circuit, not the winding's condition. The connected value sits below the motor alone and below the cable alone. Judge the motor against its own history: a machine that read 1000 MΩ at commissioning and reads 250 MΩ today is still above 100 MΩ and still in decline, and that fall is your action trigger.

Can we trend switchgear readings against the motor-only commissioning value?

No. The combined reading is always below either component, so that comparison shows a permanent offset that is arithmetic, not insulation. Compare same configuration to same configuration. If commissioning captured only a combined number, use it as the combined baseline and accept that no motor history exists until you take a motor-only reading.

What minimum applies to a 460 V motor?

The classic rule is 1 MΩ per kV of winding rating plus 1 MΩ: 0.46 + 1 = 1.46 MΩ. The 100 MΩ figure from IEEE Std 43-2000 applies to new stator windings rated 2300 V and above, not to a 460 V machine. Apply the machine documentation's figure, temperature-correct every reading, and trend regardless, because the minimum is a floor, not a condition assessment.

Does DC testing damage medium-voltage cable?

A routine insulation-resistance test at the voltage and duration the machine standard specifies is not an overpotential test. The field concern behind the question is extended DC overpotential testing of service-aged extruded-dielectric cable. If your program includes overpotential tests, confirm the method against the cable and accessory documentation before applying anything above the IR test voltage.

How do we avoid redoing high-voltage terminations at every test?

Reserve the motor-terminal test for trigger events, such as a 20% trend fall, a below-minimum reading, or a rewind-or-continue decision, rather than every cycle. On machines that need watching repeatedly, remake the joints once with voltage-class motor pigtail splice kits built to be opened and remade, so the next split test is an hour of work instead of a termination rebuild.

Keep reading

  • Troubleshooting

    Motor running over nameplate amps after soft starter bypass

    The motor averages 192.7 A against 164.0 A FLA with only about 0.90% voltage-magnitude unbalance, so the supply is not the proven cause. Treat excess shaft load or excess motor loss first, and confirm negative-sequence voltage with a three-phase power-quality analyzer reading V2/V1, total kW and speed before replacing the starter or the motor.

    Motors and motor starters

  • Troubleshooting

    Why motors burn up with no overload trips on 480 V

    Crystallized insulation several feet back from motor terminals is the signature of high-resistance terminations cooking the leads from I²R heat, while unbalanced voltage adds negative-sequence heating that current-magnitude overload heaters never see. Record all three line-to-line voltages and all three phase currents under load, verify the transformer grounding, and measure insulation resistance on every failed branch before any restart. Welded contacts and ruptured starters are consequences of closing onto failed motors, not a separate mystery.

    Motors and motor starters

  • Troubleshooting

    Setting voltage imbalance limits on a 480 V phase monitor

    Keep voltage unbalance at the motor terminals at 1% or less to run full rated power; above that NEMA MG 1 requires derating, reaching 75% of rated power at 5%, with operation above 5% not recommended. A monitor set to 15% voltage imbalance is far too loose to protect a motor. Measure A-B, B-C and C-A at the starter stopped and running, and at the main panel, and call the utility only when the main panel itself is unbalanced.

    Power quality and harmonics