The pump worked fine for eleven years. Then the homeowner replaced the old pressure tank, and now the motor hums, the lights dim, and the breaker holds. Nothing is wrong with the motor. Nothing is wrong with the breaker. The wire has been the same the whole time. What changed is that the pump now has to start against a heavier load, and the wire has quietly been taking a bite out of its strength on every start since the day it was installed.
That bite is the part of voltage drop most people never see. We usually talk about voltage drop as lost efficiency, a few percent of your power turning into warm copper. For a light bulb, that picture is close enough. For an induction motor it badly undersells the problem, because a motor does not lose pulling power in proportion to voltage. It loses it in proportion to the square of voltage.
The square that changes everything
In an induction motor, the stator's rotating magnetic field induces current in the rotor, and the two fields interact to make torque. Double-count the effect of voltage: lower voltage weakens the stator field, and it weakens the current induced in the rotor by that field. Each factor scales with voltage, so torque scales with voltage times voltage.
To a good approximation, at a given speed, torque is proportional to V².
Put numbers on it:
- At 95% of rated voltage, torque is about 0.95² = 90% of normal.
- At 90% of rated voltage, torque is about 0.90² = 81%.
- At 85% of rated voltage, torque is about 0.85² = 72%.
A 10% sag costs you 19% of the motor's pull. That's why a motor on a marginal circuit can run a light load without complaint and then refuse a heavy one. The torque it needs to start has not changed. The torque it can make has.
Why starting is the moment of truth
A stopped motor behaves almost like a transformer with a shorted secondary. It draws locked-rotor current, commonly around six times its full-load current, until the rotor spins up. (If you've read about locked-rotor current, this is the other half of that story: the surge isn't just hard on the breaker, it is hard on the voltage.)
And voltage drop in the supply conductors is current times resistance. Six times the current means six times the drop, at the exact moment the motor needs the most torque it will ever need.
Here's a worked example, using rough resistive-only numbers:
- A 10 HP, 230-volt three-phase motor. NEC Table 430.250 lists its full-load current as 28 amps.
- It's fed by 8 AWG copper, 150 feet from the panel. Chapter 9, Table 8 gives uncoated 8 AWG copper about 0.778 ohms per 1,000 feet, so 150 feet is roughly 0.117 ohms of conductor.
- Three-phase drop is √3 × I × R.
Running: 1.732 × 28 A × 0.117 Ω ≈ 5.7 volts, about 2.5% of 230. Perfectly respectable.
Starting: call locked-rotor current 168 amps (six times 28). Then 1.732 × 168 × 0.117 ≈ 34 volts, about 15%. The motor sees roughly 196 volts. That is 85% of rated, so available torque is about 0.85² ≈ 72% of what the nameplate promises.
This is a back-of-the-envelope estimate. It ignores conductor reactance, the impedance of the utility transformer, and the fact that the inrush current itself falls as the voltage sags. The real numbers will shift. But the shape of the result is the point: a circuit that looks fine at running current can be starving the motor at startup.
Low voltage makes a running motor hotter too
Starting is the dramatic failure. There's also a quieter one. A motor driving a fixed mechanical load needs a roughly fixed amount of power. If the voltage is low, the motor compensates by drawing more current, and it runs with higher slip. Higher current means more I²R heating in the windings, so a motor on a chronically sagging circuit runs hotter and ages faster. Insulation life is famously sensitive to temperature. A motor that survives for years on a good circuit may cook itself in a few on a bad one, and nobody connects the failure back to a wire that was sized by the book.
Motors are built with some tolerance for this. NEMA MG 1 expects a motor to operate within ±10% of nameplate voltage, but "operate" is carrying a lot of weight in that sentence: performance numbers, including torque, current and temperature rise, are only guaranteed at rated voltage. You get tolerance, not equivalence.
What the Code does and doesn't say
The NEC's 3% branch-circuit and 5% total recommendations live in informational notes, not in enforceable requirements, so you can pass inspection with a circuit that strangles a motor. That's not a loophole to be proud of. The Code's job is to keep the installation from being a fire or shock hazard. Whether the equipment works is, in the end, your job.
The practical lesson is to evaluate the drop at starting current for any motor that has to start under load: well pumps, compressors, conveyors, hoists, big fans on long feeders. Evaluating it at running current alone tells you the motor will be efficient. It does not tell you the motor will start.
Your next moves
- Run the starting-drop number, not just the running one. Take the motor's full-load current from NEC Table 430.248 or 430.250 (not the nameplate), multiply by six for a conservative locked-rotor estimate, and compute the drop with the same formula you'd use for running current. If the motor lands below roughly 85–90% of rated voltage during that surge, treat it as a red flag.
- Square the sag. Convert any measured voltage at the motor into percent of nameplate, then square it. 208 volts on a 230-volt motor is 90%, so about 81% torque. That single calculation tells you how much margin you actually have.
- Measure instead of guessing. With a meter that has a min/max or inrush function, read voltage at the motor terminals while it starts, not at the panel. A reading taken at the panel hides the entire wire run you're trying to evaluate.
- Upsize the conductor before you upsize the motor. On a long run, going up one or two wire sizes is often cheaper than replacing a motor that kept stalling and tripping. Recompute the starting drop with the larger wire to see the improvement before you pull anything.
- Check the other end of the wire too. Loose lugs and corroded connections add resistance that never shows up in a calculation. Verify torque on motor-circuit terminations, then re-measure the start-up voltage.
Putting it in your pocket
The arithmetic here is simple but tedious, and it's exactly the kind of thing that gets skipped on a job site because it's raining and the truck is two hundred feet away. Voltly is built for that moment: a voltage drop calculator with conductor resistance tables and the NEC references on your phone, working offline in a trench or a crawlspace. You can plug in the starting current instead of the running current and see right away whether the wire is helping or hurting. If that sounds useful, you can take a look at https://voltly.lumenlabs.works. The principle works with a pencil and a Chapter 9 table, too.