There are few arguments on a job site more reliably circular than the one about voltage drop. One man says three percent is code. Another says it is not code, it is a note. A third says his inspector red-tagged a run for it last year. All three are describing something true, which is the problem.
Here is the actual situation. The three percent branch circuit figure and the five percent combined feeder-and-branch figure live in informational notes, and informational notes are explanatory material. Article 90.5(C) says so plainly. They are not enforceable requirements, and the code even tells you they are not. You can confirm that for yourself through NFPA's free access portal, which needs a free account before it will show you the text.
So the man who says it is not code is technically right. He is also the man most likely to hand you a job where the lights dim when the compressor starts.
Where the numbers came from
The percentages are not arbitrary and they are not about the wire. They are about what the equipment at the far end expects. Utilization equipment in the United States is built and listed on the assumption that it sees something reasonably close to nameplate voltage. A 120 volt appliance is generally designed to tolerate a supply band, and once you have burned five percent between the service and the load, you have spent most of the margin the manufacturer allowed for everything else that varies: utility voltage swing, load on the transformer, other loads on the same feeder.
Split three and two, or three and five combined, and the branch circuit gets the larger share because that is where the long thin conductors usually are. It is a budgeting exercise, not a physics constant. That is exactly why it is written as advice.
Where voltage drop actually is mandatory
Being a recommendation in Article 210 does not mean it is a recommendation everywhere. There are places in the code where a drop limit is a hard requirement, and they are the places where the consequence is not inconvenience:
- Fire pumps. Article 695 sets a starting-condition voltage limit at the controller, because a pump that will not start is the failure mode that kills the building.
- Sensitive electronic equipment. Article 647 imposes tighter limits on separately derived 120/240 volt technical power systems.
- Elevators, escalators and similar. Article 620 addresses conductor sizing with the same intent.
Then there is the one that catches contractors sideways: local amendment. A jurisdiction can adopt the NEC with changes that make the informational note figures enforceable, and plenty do. Which edition you are inspected against, and what your state changed, are separate questions with separate answers. IAEI keeps a readable adoption table for all fifty states and DC with effective dates. Read it before you argue with an inspector rather than after.
What actually goes wrong on a long run
The textbook failure is dim lights. That is the least interesting one, and with LED drivers it often does not even happen, which makes people think voltage drop stopped mattering. It did not; the symptom moved.
Motors are where it gets expensive. A motor is close to a constant-power load. Drop the voltage and it draws more current to make the same torque, which heats the windings, which raises resistance, which draws more current. A well drilled pump on a long buried run at the edge of a property, a compressor at the far end of a shop, a condenser fed with undersized copper because the run looked short on the plan: these are the calls that come back. The motor does not fail dramatically. It runs hot for two summers and then does not start one morning.
Starting current is the second problem. Locked-rotor inrush on a motor can be several times running current for a fraction of a second. On a run already sized to the edge, that inrush produces a drop deep enough that the contactor chatters or the control voltage collapses. The circuit tests fine with a meter and misbehaves under load, which is the hardest kind of fault to sell to a customer.
A conductor sized only for ampacity is sized for whether it survives. Sizing for voltage drop is sizing for whether the thing on the end of it works.
Third, and increasingly common: EV charging and heat pumps have put long, continuous, high-current residential runs into houses that never had them. A forty amp continuous load down a hundred and forty feet of 8 AWG to a detached garage is an ordinary request now. It was rare fifteen years ago. The rule of thumb many of us learned — worry about it past a hundred feet — was calibrated for a load profile that no longer describes the work.
How to handle it without overthinking it
Do the calculation. It is one of the few in the trade that is genuinely quick: circular mils, one-way distance, current, and the K value for the metal. Do it at design, not after the wire is on the reel. And do it for the real load, including the continuous load factor, not for the breaker size.
Then make a decision and be able to say why. If you upsized from 10 to 8 because the run is long and the load is continuous, say that. If you stayed at 10 because the load is intermittent and the drop lands at 3.4 percent, say that too. An inspector who asks about voltage drop is usually checking whether you thought about it, not whether you hit a number that is not in his enforceable text.
The people who struggle with this are rarely struggling with the formula. They are struggling with knowing which figures bind and which advise, which is the same skill the licensing exams test hardest. Texas publishes its pass rates, and in FY2025 only 27.52 percent of journeyman candidates passed, with the calculations paper at 20.56 percent.
If that is the wall you are hitting, the Journeyman Electrician Exam Prep Study Guide 2026 is organized around worked calculations rather than around the order of the code book, which is usually the difference between recognizing a question and being able to answer it in ninety seconds.