Electrical
Electrical device wiring: receptacles and switches field guide
Land the hot, neutral, and ground on the right screws, use the screws and not the back-stabs, torque to spec, and pigtail so the device is not carrying the downstream load.
Direct answer
Device wiring is terminating receptacles and switches onto the branch circuit: hot to the brass screw, neutral to silver, ground to green. Land conductors on the screw terminals, not the back-stab push-ins, torque them to the device spec, and pigtail rather than feed downstream current through the device. The adopted code edition controls the requirements.
Key takeaways
- Land hot on the brass screw, neutral on silver, ground on green; swapping brass and silver creates reversed polarity that energizes the wrong side of plugged-in devices.
- Use the screw terminals or a true back-wire clamp, never the spring-clip back-stab push-in, which loosens, heats, and burns the device.
- On a multi-wire branch circuit the neutral must be spliced and pigtailed at every device, never run through it, or pulling the device opens the neutral and puts 240 V across 120 V loads (NEC 300.13(B)).
- Torque every terminal to the device's specified value with a calibrated tool where required (NEC 110.14); do not eyeball it.
- A 20 A receptacle cannot go on a 15 A circuit; 15 A receptacles are allowed on a 20 A multi-outlet circuit (NEC 210.21(B)).
What device wiring is and why the connection is what fails
Device wiring is the act of terminating a receptacle or a switch onto the branch circuit that feeds it. The wiring method got the conductors to the box, and the device type was chosen for the circuit and the load. This is the last step, where the conductor actually meets the device, and it is the step where most of the heat, the arcing, and the callbacks come from.
Here is the thing rookies underrate: the conductor almost never fails. The connection fails. A length of #12 copper carries its current all day without complaint, but the point where it lands on a screw or a clip is a mechanical joint, and a mechanical joint that is loose, nicked, or under-torqued runs hot. Heat oxidizes the contact, oxidation raises resistance, higher resistance makes more heat, and the joint walks itself toward a glowing connection and a scorched device. You find it by the brown discoloration on the strap or the smell before you find it with a meter.
So three things have to line up at every device. The right termination, made the reliable way. The right device for the circuit and the location. And the right circuit behind it, sized and run correctly. The conductor sizing and the raceway are covered in the wiring methods guide, and picking the device by its NEMA configuration is covered in the receptacle types guide. This guide is the part in between: getting the connection right at the device.
Wiring the receptacle: hot, neutral, and ground
A standard 125 V receptacle takes three conductors, and each has a screw with a job. The hot, the ungrounded conductor, lands on the brass screw. The neutral, the grounded conductor, lands on the silver screw. The equipment grounding conductor lands on the green screw. Brass for hot, silver for neutral, green for ground. Burn that into muscle memory, because the consequence of swapping the brass and the silver is reversed polarity that energizes the wrong part of every plugged-in device.
The two brass screws are tied together internally, and so are the two silver screws, unless you deliberately break the tab between them to split-wire the device. That bridge is what lets a single receptacle feed a conductor onward to the next box, and it is also the trap that tempts people into daisy-chaining the current through the device. More on that below.
On a GFCI or any device that protects what is downstream, the terminals are not interchangeable. One pair is marked LINE, the incoming supply from the panel, and the other pair is marked LOAD, the downstream receptacles the device protects. Land the supply on LINE. Get that backwards and the device protects nothing, and on many newer units it will not reset at all.
- Hot (ungrounded)
- The energized conductor from the breaker; lands on the brass screw
- Neutral (grounded)
- The return conductor; lands on the silver screw, never on the green
- EGC (equipment ground)
- The bare or green-insulated fault path; lands on the green screw
- Line vs load
- On a GFCI, line is the incoming supply, load is the protected downstream wiring
Should you use back-stab or screw terminals?
Use the screw terminals. The back-stab push-in is the single most common failure point in residential device wiring, and the choice between the two is not close.
A screw termination clamps the conductor under a binding-head screw with real surface area and a torque you control. A back-stab, also called a push-in, takes the stripped conductor straight into a hole in the back of the device, where a thin spring clip is the only thing holding it. That clip starts with a small contact patch and a light grip, and under the normal heating and cooling of a loaded circuit it relaxes. The grip loosens, the contact resistance climbs, the joint heats, and you get the burned receptacle that homeowners describe as an outlet that stopped working or started smelling. The conductor was fine. The clip let go.
Back-stabs are also limited by listing to 15 A devices with #14 solid copper, which tells you something about how much the manufacturers trust them. They go fast on a production wiring job, and that speed is the whole appeal. It is not worth it. Strip a little longer, hook the conductor under the screw, and torque it. The minute you spend per device is cheaper than one callback to a scorched outlet.
The back-stab clip, and the back-wire clamp that is not the same thing
Do not confuse the cheap push-in back-stab with the back-wire clamp on a better device. They look similar from the front and they are not the same connection.
The back-stab is the spring-clip hole described above. The conductor goes in straight and a flat spring is the only contact and retention. That is the one that fails.
The back-wire clamp, found on commercial spec-grade and better devices, is a different animal. The conductor goes behind a steel pressure plate that the binding screw drives down onto it, so the screw still does the clamping but the plate spreads the force across the conductor instead of pinching it under the screw head. That is a screw-quality termination with the convenience of a straight strip, and it is what a lot of commercial work standardizes on. If a device offers a true back-wire clamp under the terminal screw, that is fine. If the only option in the back is a spring-clip stab, terminate on the side screw instead. The tell is whether the screw moves the plate. A clamp tightens onto the wire. A stab just holds it.
Why pigtail a receptacle instead of daisy-chaining?
Pigtail the device. Splice the incoming and outgoing circuit conductors together in the box and run a short tail to the device, instead of feeding the circuit through the device terminals so the device carries the downstream current.
When you daisy-chain, you land the feed-in on one screw and the feed-out on the other, and now every load downstream of that receptacle gets its power through that one device's internal tab and terminations. The device has become a splice for the whole rest of the circuit. If that device fails, loosens, or gets removed, everything downstream goes with it, and the failure cascades down the run. People troubleshooting it chase a dead outlet three rooms away when the problem is a loose screw upstream.
Pigtailing breaks that dependency. The splice carries the through current with a wire connector rated for it, and the device only carries what is plugged into that one outlet. Pull the device to change it and the downstream circuit stays live and intact. For a standard string of outlets the pigtail is good practice. On a multi-wire branch circuit, pigtailing the neutral is not optional, and the reason is dangerous enough that it gets its own section below.
Torque the device terminals to spec
A device terminal has a torque value, and the loose connection is where the heat is born, so the screw gets torqued to that value. The NEC requires terminations to be made up to the manufacturer's specified torque, and recent editions, around the 110.14 termination rules, have pushed the use of a calibrated torque tool to actually hit the number rather than tightening by feel and gunslinger wrist.
On receptacles and switches the value is small, commonly in the low single digits of inch-pounds up into the teens depending on the screw and the device, and it is printed on the device or in the packaging. Do not invent it and do not eyeball it. Use the value the device lists. A device terminal under-torqued runs hot the same way a feeder lug does, just at a smaller scale, and there are a lot more of them in a building.
One detail the field gets wrong: you do not loosen a properly torqued termination and re-torque it to the same value expecting the same joint. The conductor has already taken a set under the screw. If you back it off, re-seat the conductor fresh. The terminations-by-torque discipline is covered more fully in the wiring methods material; at the device, the short version is land it on a screw and tighten it to the number the device gives you.
Grounding the device
The equipment grounding conductor lands on the green screw, and the path it makes is the one that has to work when everything else has gone wrong. An open ground means a fault on the device or the thing plugged into it has no low-impedance way back to trip the breaker, so the metal stays energized and waits for a person to complete the circuit.
In a plastic box the green screw on the device is the whole ground connection, so that screw matters. In a metal box the box itself has to be bonded too, and the device's mounting yoke can be part of that path. A self-grounding receptacle has a spring clip or a special screw on the yoke that bonds the device to a grounded metal box through the mounting screw, so you do not need a separate bonding jumper from the box to the device. Without that feature, the code wants a bonding jumper, and the grounding continuity in the box must not depend on the device, the same continuity-not-through-the-device principle that governs the neutral.
On a multi-gang metal box feeding several devices, the grounds get spliced with a pigtail to each device rather than chained through the straps. Test the ground after, do not assume it. A green screw with a conductor on it is not proof the other end is bonded to anything.
Wiring a single-pole switch
A single-pole switch is the simplest device in the box and the one whose one rule people still break. It breaks the hot, never the neutral. The switch interrupts the ungrounded conductor so that when it is off, the load downstream of it is de-energized and safe to work on. Switch the neutral instead and the lamp goes dark but the fixture stays hot, which is how someone gets bitten changing a bulb on a switch they thought was off.
Electrically it has two brass terminals and a green ground screw, and there is no line and load distinction on a plain single-pole because it is just an open or closed contact. One terminal gets the hot feeding in, the other gets the switch leg going out to the load. The two are interchangeable on a standard single-pole, though the device may mark a common.
The wrinkle is the white conductor. In an older switch loop, a cable run to a switch and back used the white as the return leg, which is a switched hot, not a neutral. Recent code wants that re-identified with tape or marking and, increasingly, wants an actual neutral present in the switch box. That neutral requirement is its own section. If you see a white conductor on a switch, confirm whether it is a re-identified hot before you treat it as grounded.
How do you wire a 3-way switch?
A 3-way pair lets two switches control one light, and the whole thing turns on getting the common terminal right. Each 3-way switch has three current-carrying screws: one common, usually the dark or black-colored screw, and two travelers, usually brass or marked. The two travelers run between the two switches, and the commons are the connection points to the rest of the circuit.
Wire it like this. At the first switch, the incoming hot from the panel lands on the common. The two travelers run from that switch to the two traveler terminals on the second switch. At the second switch, the common goes out as the switch leg to the light. The neutral runs straight through to the fixture and never touches either switch. Now either switch flips the connection between its common and one of the two travelers, and any combination of positions either completes or breaks the path, so either switch toggles the light regardless of the other.
The classic confusion, and the number one 3-way callback, is landing a traveler on the common or the hot on a traveler. The lights then work from one switch but not the other, or only in certain combinations. When a 3-way acts possessed, it is almost always a common-and-traveler mix-up. Identify the common screw on each device first, before you land a single conductor, and the rest falls into place.
Adding a 4-way switch in the middle
When you need three or more switches on one light, the extras go between the two 3-ways as 4-way switches. The 3-ways still sit at the two ends of the run with their commons facing the hot and the load. The 4-way drops into the traveler pair in the middle.
A 4-way switch has four terminals and no common. It takes the pair of travelers coming in from one 3-way and the pair going out to the next device, and depending on its position it either passes the travelers straight through or crosses them over. That swap is what lets a middle switch reverse the state and toggle the light. Add as many 4-ways as you need and they all stack in the traveler line between the two 3-ways.
The wiring trap is matching the terminal pairs. The two travelers from the upstream device go on one designated pair, the two heading downstream go on the other pair, and crossing those pairs at the device is the usual reason a 4-way circuit works erratically. Follow the terminal grouping the manufacturer marks on that specific device, because the pairing is not standardized across brands.
The neutral at the switch box
Recent code editions require a grounded conductor, a neutral, at most lighting switch locations, and the reason is the smart switch. An electronic switch, a timer, an occupancy sensor, or a dimmer that talks to a hub needs a small standing current to run its electronics even when the load is off, and without a neutral in the box those devices used to steal that current through the load, which causes flickering and early failures. This requirement is commonly cited around 404.2(C) in editions that carry it.
The point is to have the neutral present for a future device, not necessarily connected to a plain switch today. So you pull a neutral to the switch box and cap it if the current switch does not use it, leaving it for whoever installs the smart switch later.
The requirement has real exceptions, and they vary by edition, so confirm against the adopted code. Common carve-outs cover switch boxes fed by a raceway large enough to add a neutral later, boxes accessible for adding a cable without tearing out finish, and switches controlling a receptacle rather than lighting. The old switch loop without a neutral is not automatically wrong on an existing install, but on new work in a covered location, run the neutral. It is cheap now and a wall-opening retrofit later.
The GFCI receptacle: line, load, and where it goes
A GFCI receptacle compares the current going out on the hot to the current coming back on the neutral, and if they differ by a few milliamps it assumes the missing current is leaking through something, possibly a person, and trips in a fraction of a second. It protects against the shock path the breaker is too slow and too coarse to catch.
The wiring distinction that matters is line versus load. LINE is the supply from the panel. LOAD is the downstream wiring you want the GFCI to also protect, so a single device at the start of a run can protect every ordinary outlet after it. Land the incoming supply on LINE. If you reverse them, the device protects nothing downstream, and most modern GFCIs with self-test will simply refuse to reset, which is the manufacturer telling you the wiring is wrong. The TEST button creates a tiny fault to prove the trip works, and RESET restores it; tell the owner to push TEST monthly, because a GFCI that will not trip is worse than no GFCI, since people trust it.
Where GFCI protection is required is set by code and has expanded over the cycles, commonly cited around 210.8: bathrooms, kitchens, garages, outdoors, basements, laundry areas, and more in newer editions. The required locations are a topic in their own right; confirm the list against the adopted edition rather than the one you learned years ago, because it keeps growing.
AFCI protection
An arc-fault circuit interrupter watches for the electrical signature of an arcing fault, the kind a damaged cord, a stapled-through cable, or a loose connection makes, and opens the circuit before that arc starts a fire. Where a GFCI is about shock, the AFCI is about ignition.
It usually lives at the breaker as a combination AFCI breaker that protects the whole branch circuit, which is the cleanest way to cover the run. There is also an outlet-branch-circuit AFCI device for retrofits where a breaker swap is not practical, installed at the first outlet. Both are listed to the arc-fault standard, UL 1699.
AFCI protection is required for most dwelling living-area circuits in recent editions, commonly cited around 210.12, and like the GFCI list it has broadened over the cycles, so confirm the rooms against the adopted code. The field reality worth knowing: AFCIs nuisance-trip on certain motor loads and on shared-neutral wiring that is not handled right, and a real arcing connection, often a back-stab or a loose screw, will trip them too. When an AFCI keeps tripping, check the terminations before you blame the breaker.
Tamper-resistant and weather-resistant devices
Two device features are required by location, not by preference, and using the wrong one is a failed inspection even when the wiring is perfect.
Tamper-resistant (TR) receptacles have internal spring-loaded shutters that only open when both blades push in at once, so a child cannot jam a single object into one slot. Recent code requires TR receptacles at most 15 A and 20 A, 125 V and 250 V nonlocking outlets in dwellings, commonly cited around 406.12, and the covered areas have grown to include garages and accessory buildings. A standard receptacle in a dwelling is the wrong device now.
Weather-resistant (WR) receptacles are built to take damp and wet locations without corroding, and outdoors they pair with an in-use cover, the bubble cover that keeps the receptacle protected while a cord is plugged in, which in current editions has to be the extra-duty type. A plain indoor receptacle behind a flat while-closed cover outdoors is the common mistake. Outdoors and in wet or damp spots, the device is WR and the cover is an in-use extra-duty cover. Confirm the exact required device types and locations against the adopted edition, because both lists have moved.
Matching the receptacle to a 15 A or 20 A circuit
A 20 A receptacle has a T-shaped neutral slot so a 20 A plug can land in it; a 15 A receptacle has two straight parallel slots. The device has to match the circuit it is on, and the rule is not symmetric.
On a 20 A branch circuit you may install 15 A receptacles when there is more than one receptacle on the circuit, which is how a normal multi-outlet 20 A small-appliance or general circuit gets wired with ordinary 15 A duplexes. What you cannot do is put a 20 A receptacle on a 15 A circuit, because the T-slot tells the user the outlet can deliver 20 A when the breaker will trip first. And a single receptacle on an individual 20 A branch circuit has to be a 20 A receptacle rated for the full circuit. These allowances are commonly cited around 210.21(B).
The practical mistake is reaching for whatever 20 A device is in the van and landing it on a 15 A home run, or running a 12 AWG circuit, protecting it at 20 A, and then installing devices and a breaker that do not agree. Match the breaker, the conductor, and the device rating to each other.
What is a multi-wire branch circuit?
A multi-wire branch circuit (MWBC) is two hot conductors on opposite phases sharing one neutral. Because the two hots are out of phase, the neutral carries only the difference between the two loads, not the sum, so one neutral can serve two circuits instead of two separate neutrals. It saves a conductor and is common on kitchen small-appliance circuits and in commercial work.
Two rules make it safe, and both are about the fact that those two hots are full circuits sharing infrastructure. First, the two hots must be on opposite phases, or legs, so the neutral currents subtract instead of add; land both on the same leg and the neutral carries the sum of both circuits and overheats. Second, the code requires a means to disconnect all the ungrounded conductors of the MWBC simultaneously, commonly cited around 210.4, which in practice means a two-pole common-trip breaker or two single-pole breakers joined with a listed handle tie. That way you cannot kill one half and work on the box thinking it is dead while the shared neutral still carries the other hot's return current.
The MWBC is efficient and completely fine when wired right. It is also genuinely dangerous when the neutral is handled wrong, which is the next section.
Box fill counts the device too
The device is not free space in the box. Box fill, the count of what a box can legally hold, includes an allowance for each device on its mounting yoke, and that allowance is two conductors of the largest size connected to the device. A duplex receptacle in a box already wired full of conductors, clamps, and grounds can push the box over its volume, and an overstuffed box is hard to fold conductors into, hard to seat the device square, and harder on the terminations as you cram it shut.
The full counting method, conductors plus clamps plus devices plus grounds against the box's cubic-inch rating, is commonly cited around 314.16 and is a topic in itself. At the device, the thing to carry is that the device costs you volume, so the box that looked big enough for the wires may not be big enough once the device goes in. When the install is a fight to close, the answer is a deeper box or a box extension, not muscle. Forcing a device into a packed box damages the terminations you just made.
Wire prep and seating the device in the box
How the conductor is prepped decides how the termination holds. Strip to the length the device's strip gauge shows, no more, so there is no bare conductor outside the screw and no insulation under it. Do not nick the copper when you strip; a nick is a stress riser, and that is where the conductor breaks later when it gets flexed during install or service. For a screw termination, bend a hook and wrap it clockwise around the screw, so that tightening the screw pulls the loop closed instead of pushing it open. Wrap it counterclockwise and the screw spits the conductor out as you tighten.
Seating the device matters as much as the connection. Fold the conductors back into the box in an accordion, not a wad, so the device slides in without straining a termination. Do not over-tighten the 6-32 mounting screws; cranking them bows the strap, cracks the device, or pulls it crooked, and a device under that kind of stress can crack a terminal months later. Use the breakaway ears and a level so the device sits plumb in the wall, because the cover plate will telegraph every degree it is off. Set the plate flush, not crushed into the wall and not standing off it. None of this is glamorous. All of it is the difference between an install that holds and one you see again.
Commercial and data-center device wiring
Commercial and data-center work raises the bar on the same fundamentals. The devices are spec-grade or hospital-grade, built with stronger contacts and back-wire clamps instead of stab connections, because the duty cycle and the cost of a failure are both higher. The terminations get torqued to spec as a matter of routine, not as an exception, and the inspection expects it.
Two device-wiring patterns show up more on this work. Isolated-ground receptacles, marked with an orange face and a triangle, run a separate insulated grounding conductor back to the source to keep noise off the ground reference for sensitive electronics, which is a different conductor path than the normal box bond. And shared neutrals get scrutiny because nonlinear loads, computers and switching power supplies, put harmonic current on the neutral that does not cancel between phases the way the fundamental does, so a shared neutral that is fine on linear loads can run hot on a bank of servers. Many data-center designs avoid shared neutrals on those circuits for exactly that reason. Treat the specifics as project-and-design driven, and follow the contract documents, because this is where the spec gets particular.
Testing the device after you wire it
Wire it, then prove it. A plug-in receptacle tester, the three-light kind, drops into the outlet and shows the common wiring faults at a glance: correct wiring, open ground, open neutral, open hot, and reversed polarity where the hot and neutral are swapped. Many include a GFCI test button that trips a protected outlet so you confirm the GFCI actually works from the load side.
Know the tester's blind spots, because they are real. A three-light tester cannot reliably catch a bootleg ground, where someone jumped the neutral to the ground screw to fake a ground on an old two-wire circuit, and it can be fooled by certain reverse combinations. For the cases that matter, confirm with a meter: hot to neutral should read line voltage, hot to ground should read about the same, and neutral to ground should read close to zero. A neutral-to-ground voltage that climbs under load points to a loose neutral or an overloaded shared neutral, which is exactly the MWBC failure you want to catch before it becomes a callback.
On a switch, confirm it breaks the hot and that the load is dead with the switch off, verified with a meter on a known-live reference, not with the light. On a 3-way pair, operate every switch through every position and watch it follow, because a common-and-traveler error often works in some positions and not others.
How device wiring actually fails
The failure modes on device wiring are a short, repeating list, and knowing them by their symptom is how you troubleshoot fast.
The back-stab burnout is the most common: a push-in connection relaxes, heats, and you get a dead or intermittent outlet, a burning smell, and brown scorching on the device when you pull it. Reversed polarity, hot and neutral swapped, leaves the device working but energizes the part of a plugged-in appliance that is supposed to be neutral, a shock hazard the tester catches instantly. An open ground leaves metal that should be bonded floating, so a fault has no path to trip the breaker. The open neutral on a multi-wire circuit is the dangerous one, putting 240 V across 120 V loads when a through-device neutral opens. And the plain loose screw is the quiet one: a termination that was never torqued, running warm for years until it finally cooks.
Notice that most of this list traces back to two decisions made at the device: using a back-stab instead of a screw, and running the circuit through the device instead of pigtailing it. Make those two choices right and you have eliminated the majority of the failures before they start.
What to document
Device wiring rarely gets a paper trail, and that is a missed chance, because the device is where the troubleshooting starts later. On work where it matters, on commercial circuits, on MWBCs, on anything torqued to spec, capture the device, how it was wired, and the note that the next person needs.
The records worth keeping: which circuits share a neutral and that the neutral was pigtailed, that the MWBC hots are on opposite legs and tied at the breaker, the torque applied where a calibrated value was required, the GFCI and AFCI locations and their line-load arrangement, and any isolated-ground or special device. If the install deviates from the obvious, a split-wired receptacle with the tab broken, a switched receptacle, a re-identified white conductor used as a hot, write it down so the next electrician does not have to reverse-engineer it live.
| Device or detail | How it was wired | Note for the record |
|---|---|---|
| Receptacle on a string | Pigtailed, not daisy-chained | Device removal leaves downstream live |
| MWBC neutral | Spliced and pigtailed at each device | Continuity not through any device |
| MWBC hots | Opposite legs, handle-tied or 2-pole | Simultaneous disconnect provided |
| GFCI | Supply on LINE, downstream on LOAD | Confirm reset and TEST works |
| Torqued terminals | To the device value | Calibrated tool where required |
| Re-identified conductor | White used as a switched hot | Marked with tape at both ends |
Common mistakes
- Using back-stab push-in terminations, which loosen, heat, and burn the device.
- Daisy-chaining the circuit through the device instead of pigtailing, so a device failure or removal kills everything downstream.
- Running a multi-wire branch-circuit neutral through the device, which opens the neutral and puts 240 V across 120 V loads when the device is pulled.
- Swapping the hot and neutral, energizing the wrong side of plugged-in devices (reversed polarity).
- Reversing LINE and LOAD on a GFCI, so it protects nothing downstream and often will not reset.
- Landing a traveler on the common or the hot on a traveler in a 3-way, so it works from one switch but not the other.
- Leaving no neutral in a switch box where the adopted code now requires one for electronic switches.
- Installing a plain device where a tamper-resistant or weather-resistant device is required, or skipping the in-use cover outdoors.
- Putting a 20 A receptacle on a 15 A circuit, or mismatching the breaker, conductor, and device rating.
- Leaving terminal screws loose or un-torqued, the quiet failure that cooks years later.
Field checklist
Want this checklist to run itself on every job — with photo proof and a signed record crews can hand the customer? That's FieldOS.
Standards and references
The NEC, NFPA 70, governs device wiring across several articles, and the section numbers move between cycles, so confirm them against the edition the jurisdiction has actually adopted and any local amendments before you cite them. Terminations and the requirement to make them up to the manufacturer's torque, including the calibrated-tool language, live around 110.14. The receptacle and circuit requirements sit in Article 210 and 406: GFCI protection commonly around 210.8, AFCI protection around 210.12, the 15 A and 20 A device matching around 210.21(B), multi-wire branch circuits around 210.4, and tamper-resistant receptacles around 406.12 with weather-resistant and in-use covers around 406.9.
Two safety rules are worth knowing by name and topic rather than by number alone. The continuity of a multi-wire branch-circuit neutral must not depend on a device, which is the pigtail-the-neutral rule commonly cited around 300.13(B). And the grounding continuity in a box must not depend on the device either. The neutral-at-the-switch requirement for electronic switches is commonly cited around 404.2(C) and carries several edition-dependent exceptions.
The devices themselves are listed: receptacles to UL 498, general-use snap switches to UL 20, GFCIs to UL 943, and AFCIs to UL 1699. The listing and the manufacturer's instructions can impose requirements tighter than the general code, and where they do, the listing governs. The three things to hold above all the section numbers: use the screws and not the stabs, pigtail the multi-wire neutral, and torque the terminations.
Units and terms
Device wiring uses a small vocabulary that shows up across drawings, device packaging, and the code with slightly different names for the same thing.
The ungrounded conductor is the hot, on the brass screw. The grounded conductor is the neutral, on the silver screw, and it is not the same as the grounding conductor. The equipment grounding conductor, the EGC or just the ground, is the bare or green-insulated fault path on the green screw. Terminal torque is given in inch-pounds (in-lb) or newton-meters (N-m), printed on the device or in its instructions. A pigtail is a short conductor spliced to the through wiring that lands on the device so the device is not in the through path.
- Pigtail
- A short tail spliced to the circuit conductors so the device is not in the through-current path
- Back-stab vs back-wire clamp
- A spring-clip push-in (avoid) versus a pressure plate driven by the terminal screw (acceptable)
- Line / load
- On a protective device, line is the incoming supply, load is the protected downstream wiring
- Common / traveler
- On a 3-way switch, the common connects to hot or load, the travelers run between switches
- MWBC
- Multi-wire branch circuit: two opposite-phase hots sharing one neutral, with a simultaneous disconnect
- TR / WR
- Tamper-resistant (shuttered, required in dwellings) and weather-resistant (damp and wet locations)
FAQ
Should you use back-stab or screw terminals on a receptacle?
Use the screw terminals. Back-stab push-ins rely on a thin spring clip that relaxes under heating and cooling, raises resistance, and burns the device, which is the most common residential outlet failure. A true back-wire clamp, where the terminal screw drives a pressure plate onto the conductor, is fine. The spring-clip stab is not.
Why pigtail a receptacle instead of daisy-chaining through it?
Pigtailing splices the circuit conductors together and runs a short tail to the device, so the device only carries its own outlet's load. Daisy-chaining feeds the downstream circuit through the device terminals, so a loose screw or a removed device kills everything downstream. Pigtailing also lets you change the device with the rest of the run still live.
How do you wire a 3-way switch?
Land the incoming hot on the first switch's common, run the two travelers between both switches' traveler terminals, and take the second switch's common out to the light. The neutral runs straight to the fixture. The common is the colored screw. Mixing up the common and travelers is the number one 3-way callback.
What is a multi-wire branch circuit?
A multi-wire branch circuit is two hot conductors on opposite phases sharing one neutral, so the neutral carries only the difference between the two loads. It needs a simultaneous disconnect, a two-pole or handle-tied breaker, commonly cited around NEC 210.4, and the shared neutral must be pigtailed at every device, never run through it.
Why must the neutral on a multi-wire branch circuit be pigtailed?
Because if the neutral runs through a device and that device is removed or its terminal fails, the neutral opens while both hots stay live. The 120 V loads downstream end up in series across 240 V, and the light loads can see most of it and burn out. The continuity rule is commonly cited around NEC 300.13(B).
Which screw is hot, neutral, and ground on a receptacle?
Hot, the ungrounded conductor, lands on the brass screw. Neutral, the grounded conductor, lands on the silver screw. The equipment grounding conductor lands on the green screw. Swapping brass and silver creates reversed polarity, which energizes the wrong side of every plugged-in device and is a shock hazard a plug-in tester catches instantly.
What happens if you reverse line and load on a GFCI?
If you land the incoming supply on the LOAD terminals instead of LINE, the GFCI protects nothing downstream, and most modern self-testing GFCIs will not reset at all, which is the device flagging the miswire. LINE is the supply from the panel; LOAD is the downstream wiring you want the GFCI to also protect.
Do you need a neutral in a switch box?
Recent code editions require a neutral at most lighting switch locations so electronic and smart switches have a return path for their electronics, commonly cited around NEC 404.2(C). It is usually capped for future use, not connected to a plain switch. Exceptions exist for raceway-fed and accessible boxes, so confirm the adopted edition.
Can you put a 20 amp receptacle on a 15 amp circuit?
No. A 20 A receptacle has a T-slot that tells the user it delivers 20 A, but a 15 A breaker would trip first. You can install 15 A receptacles on a 20 A multi-outlet circuit, and a single receptacle on an individual 20 A circuit must be 20 A. Match the breaker, conductor, and device, commonly cited around 210.21(B).
What torque do receptacle and switch screws need?
Torque to the value the device lists, printed on the device or in its packaging, which is small, often in the low single digits to teens of inch-pounds. Recent NEC editions around 110.14 push using a calibrated torque tool to actually hit it. Do not eyeball it, and do not loosen and re-torque a set conductor without re-seating it.
People also ask
Codes cited in this guide
This guide is written and reviewed against the published standards below. Always confirm the current adopted edition with the authority having jurisdiction.