Voltage-drop calculators are excellent at one job: estimating the voltage lost in conductors from length, material, size, current, and phase arrangement. They do not know whether the last receptacle was backstabbed, whether a feeder lug was torqued with a calibrated screwdriver, whether aluminum was landed on a copper-only terminal, or whether a wirenut splice was twisted onto conductors with damaged copper. In the field, those small connection details can turn a mathematically acceptable circuit into a hot, weak, intermittent circuit.

This article is for electricians, engineers, inspectors, maintenance technicians, and careful DIYers who use the calculator and then have to explain why the meter still shows too much drop. The topic is not replacing wire math. The topic is separating conductor voltage drop from connection voltage drop so the fix is not automatically “pull larger copper” when the real defect is one loose terminal in a junction box.

A common field scenario came from a 120 V, 20 A workshop branch circuit that calculated at about 4.1 V drop on a long 12 AWG copper run. The far receptacle measured 6.0 V lower than the panel while a compressor was running near 16 A. The missing 1.9 V was not in the cable route. Loaded millivolt checks found roughly 0.7 V across a loose receptacle feed-through and 0.5 V across a wirenut splice in a crowded box. Re-terminating the devices and replacing the damaged splice brought the loaded loss back close to the calculated value without changing the branch-circuit conductor.

TL;DR

A conductor can calculate correctly while one loose splice wastes 0.5 V to 2.0 V under load.
NEC 110.14 and NEC 110.3(B) make terminal compatibility and manufacturer torque values part of the installation check.
Measure connection drop under real load; open-circuit voltage usually hides bad contact resistance.
At 20 A, a 0.5 V bad connection dissipates about 10 W at one hot point.
Use the calculator for wire loss, then add measured terminal or splice loss during troubleshooting.

The design baseline in this article is anchored to National Electrical Code , International Electrotechnical Commission , electrical contact resistance . Those references matter because code language, conductor physics, and equipment behavior usually fail in the same place: a circuit that was technically legal on paper but poorly optimized for the distance, load, or operating temperature in the field.

A long 12 AWG branch circuit might legitimately lose 3 V or 4 V, but one loose receptacle screw can add another 0.5 V by itself. At 20 amps, that is 10 watts concentrated inside a small device box.
— Hommer Zhao, Technical Director

Why Connections Create Their Own Voltage Drop

A termination is the point where a conductor is connected to equipment, a device, a breaker, a lug, a terminal block, or another conductor. A splice is a connection between conductors that continues the circuit through a wirenut, crimp sleeve, split bolt, insulated multi-tap connector, lever connector, terminal strip, or other listed connector. Electrical contact resistance is the resistance at the touching surfaces of a connection, including the effects of pressure, oxide, plating, surface area, contamination, and conductor damage. Those three definitions matter because connection voltage drop is not the same failure mode as ordinary conductor voltage drop.

Conductor voltage drop is distributed along the whole route. A 150 ft branch circuit warms along its length and loses voltage predictably from conductor resistance. Connection drop is concentrated at one physical point. If a poor lug, wirenut, or device feed-through creates only 0.025 ohm of extra resistance, the voltage loss at 20 A is 0.50 V and the heat is I squared R: 20 x 20 x 0.025 = 10 W. Ten watts spread over a long cable is modest. Ten watts inside one device box, terminal compartment, or plastic connector is a defect worth fixing.

The NEC treats connections seriously because they are part of the current path. NEC 110.14 covers terminals, splices, conductor material compatibility, and temperature limitations. NEC 110.3(B) requires listed or labeled equipment to be installed according to the instructions, which includes terminal torque values when the manufacturer provides them. NEC 300.15 and NEC 314 help frame where splices belong and how boxes must remain accessible. IEC projects use different numbering, but IEC 60364 verification practice still expects continuity, suitable terminations, and measured performance that confirms the installation can carry design current safely.

Bad connection drop is load-dependent. A receptacle can read 120.2 V with no load and still collapse under a 12 A heater or 16 A compressor because contact resistance only reveals itself when current flows.
The heat is concentrated. A 0.5 V drop across one splice at 20 A is 10 W at that connector, while the same 0.5 V distributed through cable length is much less severe locally.
Torque matters more than feel. Panel lugs, breakers, neutral bars, mechanical connectors, and equipment terminals often publish inch-pound or newton-meter values; a loose or crushed conductor can both fail.
Material compatibility is part of the voltage-drop story. Aluminum conductors need terminals listed for aluminum, proper preparation where required, and temperature-rating coordination. Oxide and incompatible hardware raise contact resistance.

Comparison Table: Conductor Drop vs Connection Drop

Use the table to decide whether the calculator result, a field measurement, or both should drive the next step. The numbers are practical screening values, not permission to ignore damaged equipment or local code.

Condition	Typical clue	Measured value under load	Code or standard focus	Best next action
Normal conductor voltage drop	Drop matches length, AWG, material, and current	Example: 3.2 V on a 120 V long branch	NEC 210.19(A) note, NEC 215.2(A) note, IEC 60364-5-52	Upsize wire, shorten route, split load, or accept documented result
Loose device screw terminal	Device warm, downstream outlets weak	0.2 V to 1.0 V across one device at 12-20 A	NEC 110.14, NEC 110.3(B)	De-energize, inspect conductor, re-terminate or replace device
Poor wirenut or lever splice	Intermittent load, heat in box, inconsistent readings	0.3 V to 2.0 V across splice path	NEC 110.14(B), NEC 300.15, NEC 314.16	Remake splice with listed connector and correct strip length
Undertorqued feeder lug	Feeder voltage weak on one phase or leg	50 mV to 500 mV across lug at feeder load	NEC 110.14, terminal label torque	Use proper PPE, outage procedure, inspection, and calibrated torque tool
Aluminum termination issue	Darkened conductor, oxide, heat at lug	Rising millivolts as load runs	NEC 110.14, AL/CU listing, temperature limits	Verify connector listing, conductor prep, antioxidant if specified, and torque
Backstab receptacle feed-through	Voltage improves when device is bypassed	0.1 V to 0.8 V across device path at 15 A	Listed device instructions, NEC 110.3(B)	Move to listed pressure plate/screw terminal or pigtail the device

NEC 110.14 is not paperwork language. It is the rule that keeps copper, aluminum, terminals, temperature ratings, and torque from becoming hidden voltage-drop sources after the calculator says the wire size is acceptable.
— Hommer Zhao, Technical Director

Example 1: 20 A Receptacle Circuit With One Bad Feed-Through

A 120 V branch circuit feeds a far workshop receptacle through 135 ft of 12 AWG copper and two upstream device boxes. The connected compressor draws 16 A while running. A calculator check using about 1.588 ohms per 1,000 ft per conductor gives a 270 ft loop resistance of about 0.429 ohm. The expected conductor drop is 16 x 0.429 = 6.9 V, or 5.7% on 120 V. That is already high enough to consider 10 AWG or a shorter route, but the field reading at the far receptacle shows about 9.1 V loss.

The extra 2.2 V should not be blamed on the cable until the connection path is checked. With the compressor running, a meter across the first receptacle feed-through reads 0.65 V, and a second box splice reads 0.48 V. The remaining difference comes from device and connector losses that were never in the wire calculator. Replacing the damaged receptacle, pigtailing the feed-through, and remaking the splice removes more than 1.0 V of avoidable drop before any conductor upsizing decision is made.

Example 2: 240 V EVSE Feeder Lug With 0.18 V Across One Termination

A 48 A EVSE on a 60 A branch circuit has a 150 ft copper run. The conductor voltage-drop calculation may show roughly 2.4% to 3.0% depending on conductor size. During commissioning, the line-to-line voltage at the charger is lower than expected, and one line conductor lug at a disconnect shows 0.18 V across the termination while charging at 48 A. That seems small, but the heat at the lug is 0.18 V x 48 A = 8.6 W.

An EV charger can hold that load for hours. Eight watts at one lug for a five-hour charging session is not a harmless rounding error; it is a thermal defect. The correct response is not to subtract the number from the design target and move on. The work should be de-energized under appropriate procedure, the conductor and lug inspected for damage, the connector listing and conductor material verified, and the terminal torqued to the manufacturer value required by NEC 110.3(B) and 110.14.

Example 3: 277 V Lighting Row With Neutral Splice Heating

A commercial 277 V lighting row draws 14 A continuous and is fed through several accessible junction boxes. The calculated conductor drop is about 5.9 V, or 2.1%, which is reasonable for the branch. After a maintenance call, the far drivers measure low and flicker during startup. A loaded check finds 0.9 V from the line-side to load-side of one neutral splice while the row is on.

That 0.9 V at 14 A is 12.6 W in one box, and because the load is continuous, the connector temperature can keep rising. NEC 110.14(B) requires splices to be made with devices identified for the use, and NEC 314.16 box fill matters because crowded boxes make poor workmanship more likely. The fix is to remake the splice with the correct connector, verify conductor condition, confirm box fill and accessibility, then remeasure the total branch drop.

Example 4: IEC 230 V Final Circuit With Connector Loss

On a 230 V final circuit feeding a 2.3 kW heater, the design current is about 10 A. A 42 m copper run calculates near the project voltage-drop limit, but the load voltage is lower than expected. Measuring across a terminal block under the 10 A heater load shows 0.35 V across one pole. That equals 3.5 W in a small enclosure and adds 0.15% to the total system drop before the conductor is considered.

IEC 60364 verification work is not identical to NEC inspection language, but the practical conclusion is the same. Continuity and functional testing should expose abnormal connection resistance. When the measured circuit loss is higher than the calculated cable loss, the terminal path needs inspection before a designer simply selects a larger cable.

Mistakes That Hide Connection Voltage Drop

Testing only open-circuit voltage.

A bad connection can show full voltage with no load because almost no current is flowing. Use a real load or controlled test load that represents the circuit, such as 10 A, 16 A, 20 A, or equipment nameplate current.

Using the calculator as a substitute for inspection.

The calculator cannot see damaged strands, loose lugs, mismatched conductor materials, overheated insulation, poor strip length, or a device used outside its listing.

Ignoring small volt readings at high current.

A 0.15 V reading may sound minor, but at 80 A it is 12 W at one termination. Always convert connection voltage into watts when judging severity.

Retorquing energized equipment casually.

Troubleshooting may require qualified-person procedures, PPE, lockout, manufacturer instructions, and sometimes replacement rather than retightening a heat-damaged connector.

Letting box fill create bad workmanship.

Crowded boxes make splices harder to form and inspect. NEC 314.16 box-fill compliance supports connection quality as well as physical space.

A Practical Field Workflow

Use this sequence when measured load voltage is worse than the calculated conductor drop. It keeps the wire-sizing question separate from the workmanship and maintenance question.

Calculate the expected conductor drop first. Use the voltage drop calculator with real one-way length, conductor size, material, system voltage, phase, and load current. Record the expected volts and percent.
Measure source and load voltage under the same load. Compare panel, feeder, and load-side readings while the equipment is drawing current. A 16 A or 48 A live load is what reveals resistance.
Walk the circuit by voltage segments. Measure across breakers, lugs, disconnects, receptacle feed-throughs, wirenuts, terminal blocks, and equipment terminals. One connection with 0.5 V at 20 A deserves attention.
Convert suspicious readings to heat. Use watts = volts across the connection x load amps. This turns a vague 0.25 V reading into 5 W at 20 A or 20 W at 80 A.
Repair the defect, then rerun the calculator decision. After re-terminating or replacing damaged parts, measure again. If the remaining loss matches calculated conductor drop, then decide whether wire upsizing is still needed.

FAQ

Can a splice or lug add measurable voltage drop?

Yes. A clean connection should normally add only millivolts at branch-circuit current, but a loose or oxidized splice can add 0.2 V to 2.0 V under load. At 20 A, 0.5 V across one bad point means about 10 W of heat at that connection.

What NEC rule applies to terminal torque and connection quality?

NEC 110.14 covers electrical connections, conductor material compatibility, terminals, and temperature limitations. Listed equipment torque values also fall under NEC 110.3(B), so the installer should use the manufacturer torque value rather than guessing by feel.

How do I test voltage drop across a connection?

Energize the load safely, place meter probes on the line-side and load-side metal of the same termination or splice path, and read millivolts or volts across the connection. A 0.0 V open-circuit reading proves little; the test needs load current, such as 12 A, 20 A, or the equipment nameplate current.

Should connection drop be included in the calculator result?

Use the calculator for conductor length, material, current, and system voltage first. Then treat measured connection drop as an added field loss. For example, 3.1 V calculated conductor drop plus 0.8 V measured across two poor terminations equals 3.9 V total branch loss.

Is back-wiring a receptacle a voltage-drop problem?

It can be. Listed pressure-plate back-wire terminals can be acceptable when used within their rating, but push-in spring backstab connections on 14 AWG or 12 AWG branch circuits have less contact area and are a common troubleshooting point on 15 A and 20 A circuits.

How does IEC practice compare with NEC practice?

IEC 60364 verification emphasizes continuity, insulation, polarity, earth fault loop impedance, and functional testing. It does not use NEC article numbers, but the field principle is the same: a connection must carry design current without abnormal voltage drop or heat.

Separate Wire Loss From Connection Loss Before You Pull More Cable

Run the conductor calculation, document the expected voltage drop, and then measure suspect terminations under load. If a project still looks marginal after the connection path is clean, use the calculator to compare the next conductor size before ordering material.

Contact the Engineering Team Review More Guides

Termination and Splice Voltage Drop: Torque, Heat, NEC 110.14, and Field Checks