How to Work Out Whether a Run Is Too Long

Why this matters

"Too long" is not a property of a run. It is a verdict about one specific load's worst moment, and a run that is comfortably fine for a heater is unusable for the motor that replaces it. The habit of answering the question with a 3 percent rule of thumb produces both errors: runs upsized for no reason on loads that never notice, and runs signed off as compliant that cannot start the machine on the end of them. In the worked case below, a conductor that passed the rule of thumb at 2.4 percent had to go up two trade sizes, and nothing about the rule of thumb could ever have told you that.

Build the answer as a worksheet with six fields. Fill all six before you compute anything, because the two fields most people skip are the two that decide the outcome.

Field 1: nominal voltage and configuration

Write the system nominal voltage and whether the load is single-phase or three-phase, because they use different multipliers and mixing them is a common off-by-two. On a single-phase load current flows out and back, so the drop is across BOTH conductors: 2 times current times the resistance of one conductor over the one-way length. On a balanced three-phase load the return paths share, and the line-to-line drop works out to 1.732 times current times one conductor's resistance over the one-way length. Write which multiplier you used on the worksheet.

Field 2: the load's worst-case current, not its running current

This is the first field people fill in wrong. The number that sizes a run for length is the current at the moment the load is least able to tolerate low voltage, and for most equipment that is not the running current.

• A motor started across the line draws locked-rotor current for the first moments. The nameplate code letter or a stated locked-rotor amperage owns that number. The commonly-carried rule of thumb of about 6 times full-load current is a placeholder for when you do not have the nameplate in front of you, and where the two disagree the nameplate governs and the rule of thumb comes out of the worksheet.
• A motor on a drive or a soft starter does not have that moment. The drive limits input current, which is why changing the starting method is one of the real answers to a run that is too long.
• A resistive load has its worst moment at steady state, so running current is the right number.
• A switch-mode supply or a drive front end draws more current as its input voltage falls, so its worst case is at its own lowest permitted input voltage rather than at nominal (the card on what voltage drop costs the equipment works that through).

Field 3: the minimum voltage that load will accept at that moment

The second commonly-skipped field, and the one that replaces the rule of thumb.

The equipment's own literature owns this number. Motor manufacturers commonly state that a motor is capable of starting its load at 90 percent of rated voltage, and drives publish an undervoltage threshold on their parameter list. Where the equipment states a number, that number is the test.

Note what the code does and does not say here, because it is widely misread. In the NEC as adopted by your authority having jurisdiction, the familiar 3 percent branch-circuit and 5 percent total figures appear in informational notes at Articles 210 and 215, and NEC 90.5(C) makes informational notes explanatory material that is not enforceable as a requirement. They are good design practice and a reasonable default when you have nothing better, and they are not the standard the equipment is held to. A few places in the same code do set enforceable voltage requirements for specific occupancies and systems, fire pump feeders under Article 695 among them, and where one of those applies it is a requirement and it governs.

So: use the equipment's stated minimum where you have it, fall back to the informational-note percentages where you do not, and say on the worksheet which of the two you used.

Field 4: measured one-way length

Walk it or pull it from the as-built, and add the vertical runs, the drops into the enclosure and the length inside the gutter. A length estimated off a floor plan comes in short because nobody counts the risers, and drop is directly proportional to length, so that error transfers straight into the answer.

Field 5: conductor material, size and resistance per 1000 feet

Take resistance from the conductor properties table in the NEC edition your authority having jurisdiction has adopted, at 75 degrees C, and record whether the conductor is copper or aluminium. These values are direct-current resistance and they ignore reactance, which is a fair approximation for the small and medium conductors most field-service runs use. On large conductors, long runs, or runs in steel raceway, reactance is no longer negligible and the alternating-current impedance table with its power-factor columns governs instead. Say which table you used.

Field 6: what the supply itself does at that moment

The calculation gives you the drop across YOUR run, on the assumption that the panel holds nominal voltage while it happens. It does not. The service, the transformer and everything upstream also sag when the load hits, and that sag adds to yours at the equipment terminals.

Measure it: take voltage at the panel with the load off and again during the moment in field 2, and record the difference. This reading is taken on energized conductors, so 29 CFR 1910.333(a)(1) applies - de-energize before working on or near exposed live parts unless the employer can demonstrate that de-energizing introduces additional or increased hazards or is infeasible due to equipment design or operational limitations - and a loaded reading cannot be taken any other way. Work under the employer's energized-work program with boundaries and PPE selected under NFPA 70E-2021 130.5 and 130.7, in the edition your employer's program or your authority having jurisdiction has adopted. Before you start a machine to create the moment you want to measure, clear the driven end: confirm nobody is at the coupling or inside the equipment, guards are in place, and the machine is lined up to run (valves in their running position on a pump or compressor, so the start does not deadhead it).

The worksheet, filled in

A 7.5 hp three-phase motor, started across the line, at the end of a run from a 208 V panel.

• Field 1: 208 V, three-phase. Multiplier 1.732.
• Field 2: nameplate full-load current 24.0 A. For this illustration, locked-rotor is taken at 6 times full-load, so 144 A, flagged as a placeholder because the nameplate code letter governs and would replace it.
• Field 3: manufacturer states starting capability at 90 percent of rated voltage. 90 percent of 208 V is 187.2 V, so the allowable drop at the starting moment is 20.8 V.
• Field 4: 240 feet one way, measured.
• Field 5: 6 AWG uncoated stranded copper, 0.491 ohm per 1000 feet at 75 degrees C.
• Field 6: measured panel sag during a start of the existing equipment, 2 percent.

Running check first, because it is cheap. Drop equals 1.732 times 24.0 times 0.491 times 240/1000. That is 41.57 times 0.491, or 20.41, times 0.240, or 4.90 V. Against 208 V that is 2.36 percent, and the terminal voltage is 203.1 V. By the informational-note default this run passes comfortably, and this is where most sizing exercises stop.

Starting check, which is the real test. Same run, 144 A instead of 24.0 A. Drop equals 1.732 times 144 times 0.491 times 0.240. That is 249.41 times 0.491, or 122.46, times 0.240, or 29.4 V. Terminal voltage at the motor during start is 208 minus 29.4, or 178.6 V, which is 85.9 percent of nominal. The requirement in field 3 was 90 percent. It fails by a wide margin, and it failed while passing the rule of thumb by a wide margin.

Size it against the governing number. The allowable drop at start is 20.8 V at 144 A, so the run's total resistance can be 20.8 divided by (1.732 times 144), which is 20.8 divided by 249.41, or 0.0834 ohms. Over 240 feet that is 0.0834 divided by 0.240, or 0.347 ohm per 1000 feet. The next conductor at or below that in the properties table is 4 AWG at 0.308 ohm per 1000 feet.

Now re-run everything against field 6, which the first pass ignored. With the measured 2 percent panel sag, the voltage available at the head of the run during the start is 203.8 V, not 208. Drop with 4 AWG is 249.41 times 0.308 times 0.240, or 18.44 V, and the terminal voltage is 203.8 minus 18.44, or 185.4 V. Against the 187.2 V requirement from field 3, which is fixed to the motor's 208 V rating and does not move when the panel sags, 4 AWG fails too, by about 1 percent. Field 3 governs, the earlier 4 AWG answer was computed against the wrong supply voltage, and every figure downstream of it has to be redone.

Redo it. With 2 AWG at 0.194 ohm per 1000 feet, drop at start is 249.41 times 0.194 times 0.240, or 11.61 V, and the terminal voltage is 203.8 minus 11.61, or 192.2 V. That is 92.4 percent of 208 V and it clears the requirement. Running drop on 2 AWG is 41.57 times 0.194 times 0.240, or 1.94 V, which is 0.93 percent.

The run needed 2 AWG. Ampacity never came close to being the constraint: 6 AWG copper at the 75 degrees C column carries far more than 24.0 A, and so does everything above it. The conductor was sized by length, not by load, which is the normal outcome on a long run and the reason a size derived from the load table alone is not an answer.

When the answer is not a bigger conductor

Two sizes of copper for one motor is worth pausing over rather than quoting.

• Change the starting method. A soft starter or a drive removes field 2's locked-rotor number entirely, and with it the constraint that forced the upsize. On a long run this is often the smaller job.
• Move the source rather than the conductor. If several loads sit at the far end, a subpanel fed by one properly sized feeder beats several long branch circuits, and the drop calculation gets done once.
• Fix the sag instead of the run. A large field-6 sag is telling you the problem is upstream, and upsizing your 240 feet corrects the smaller half of it.
• Check whether the load is real. A motor oversized at design carries a locked-rotor number the installation is now paying for, which is a conversation with whoever specified it rather than a conductor decision.

Proving it on the finished install

Calculation sizes the conductor; measurement closes the job. With the equipment installed and under the energized-work gate above, take voltage at the equipment terminals at rest and again at the worst moment from field 2, using an instrument that captures a minimum rather than an average, since a start lasts a few cycles and an averaging meter reports a number that never occurred. Compare that captured minimum against field 3's requirement, not against a percentage.

If the measured minimum is materially worse than the calculation, the candidates are a length longer than field 4 recorded, a locked-rotor current higher than the placeholder in field 2, and a connection adding resistance the conductor table does not know about. All three are found by measuring the drop in segments rather than end to end.

References

• NEC Article 210 and Article 215 informational notes on voltage drop, and NEC 90.5(C) on the status of informational notes; NEC Article 695 for enforceable fire pump voltage requirements; conductor properties and alternating-current resistance and reactance tables in NEC Chapter 9 - all in the edition your authority having jurisdiction has adopted
• 29 CFR 1910.333(a)(1), general industry electrical safety-related work practices and the de-energizing gate
• NFPA 70E-2021, 130.5 and 130.7, as adopted by your employer's electrical safety program or your authority having jurisdiction
• Motor and drive manufacturer literature for locked-rotor current, starting voltage capability and undervoltage thresholds
• See related: Voltage Drop Calculations Reference; What Voltage Drop Actually Costs the Equipment