22

Living in the UK, where we have a different system of mains wiring, my curiosity was piqued by Is it okay to install a 15A outlet on a 20A dedicated circuit for a dishwasher? and its answers. Although there is some debate in comments as to whether a "duplex receptacle" counts one or two receptacles/outlets, the gist seems to be (from NoSparksPlease's comment):

NEC 210.21(B)(1) requires a single receptacle on a 20A be rated for 20A, 210.21(B)(3) says two or more receptacles says to refer to table 210.21(B)(3), which allows 15 and 20's

Can someone explain why a 15A single receptacle is not allowed on a 20A circuit?

(I recognise that "why" questions are often not suitable on StackExchange sites because – without documented reasons from the decision maker(s) – it can be a matter of speculation and opinion, but I feel such a question about electrical safety regulations should have an objective answer).

Under "normal operating conditions", I cannot see any problem with an appliance that can draw up to 15A, plugged into a 15A single receptacle fed by a 20A circuit. You may not be using the full capacity of the circuit, but that seems no different than if a 20A receptacle was fitted and an appliance that draws a maximum of 15A was plugged-in.

Obviously, though, electrical regulations are as much (if not more) about ensuring safety when not under "normal operating conditions". Here I can see how a 15A receptacle on a 20A circuit could be unsafe: if the appliance "goes wrong a bit" and ends up drawing 18A, you will probably be overloading the appliance's cable and definitely overloading the receptacle. But because the circuit and its breaker is 20A, it will not trip.

However, I cannot see how this danger would be ameliorated just by having a second 15A receptacle on that circuit. Assuming code doesn't require you to keep an appliance drawing 5A permanently connected to that second receptacle, simply having a second receptacle isn't going to make the 20A breaker trip any sooner, nor is it going to relieve the overloading of the receptacle that is connected to the faulty appliance drawing 18A.

What am I missing about why a 15A single receptacle cannot be the only connection to a 20A circuit?

(This answer quotes more of Code 210.21, and goes into more details about why multiple 15A receptacles on a 20A circuit are allowed and useful, but does not seem to address why a single receptacle is not allowed).

Dan A.
  • 518
  • 3
  • 15
TripeHound
  • 360
  • 2
  • 11
  • 2
    The words "outlet" and "receptacle" have specific definitions in NEC sectiion 100. An "outlet" is any point of connection of utilization equipment". This includes boxes for light fittings and hardwired appliance connections. – NoSparksPlease Dec 06 '20 at 16:13
  • @NoSparksPlease Is "receptacle" is the thing that if there is only one, must have the same rating as the circuit? If so, then I'll edit that in place of my use of "outlet". – TripeHound Dec 06 '20 at 16:24
  • 2
    Yes, receptacle devices is what the previous question was about. NEC changes are argued and adopted on a complicated three year proposal cycle. Often times we who work in the field don't take time to track proposals and do not become familiar with the reasons for the change. https://www.necanet.org/about-us/news/news-release-archive/news/2019/09/11/nfpa-releases-2020-national-electrical-code-(nec-) – NoSparksPlease Dec 06 '20 at 17:51
  • 2
    To clarify - the real distinction here is between "single receptacle" and "duplex receptacle", correct? The point being you can't install a 15A "single receptacle" anywhere on a 20A circuit, whereas you can install a 15A "duplex receptacle" on a 20A circuit even if it's the only one. I feel like even after the second edit, this should be further clarified, especially in the title. I'll suggest an edit and someone with higher rep can review. – Dan A. Dec 06 '20 at 20:30
  • @DanA.Thanks for the edit suggestion, which I've accepted. That clarification, together with the "_As a result, a 15A duplex receptacle's internal parts will be tested for handling 20A_" bit of ThreePhaseEel's answer make a lot of sense. – TripeHound Dec 06 '20 at 20:46
  • Just to make the situation worse - even if you used a single 20A outlet, you can still plug the 15A dishwasher into the 20A outlet and blow up the darn dishwasher. Part of what's missing here is the idea of *intent.* If the intended application was a 15A dishwasher, a 20A outlet with a 20A breaker should never have been installed in the first place. – JBH Dec 06 '20 at 23:42
  • 3
    @JBH: How would that work, exactly? You can't force an appliance to draw more current than it needs by plugging it into a socket that can supply more current... It doesn't work that way. – Robert Harvey Dec 06 '20 at 23:48
  • 2
    @RobertHarvey No, of course you can't force it. But the reason breakers exist at all is because accidents happen. If anything goes wrong with a 15A device on a 20A service, there's nothing to stop it from drawing 18A other than to burn up. That was the point of my comment - it's a mistake to put a 15A device on a 20A breaker for long-term permanent or semi-permanent use. It would be better to use a 15A outlet and replace the breaker with a 15A breaker. A dishwasher is a perfect example as both outlet and cord are hidden in cabinetry. – JBH Dec 07 '20 at 04:00
  • @JBH -- branch circuit breakers aren't intended to protect appliances from damaging themselves due to overload, though. (You *can* get into situations with motors where the branch breaker supplies the overload protection, but that's fairly rare) – ThreePhaseEel Dec 07 '20 at 19:01
  • The shortest nontechnical answer is "fuses in the plugs" in the UK. The shortest semi-humourous answer is that the plugs keep slipping ouf of the sockets so that children and pets can lick them so you kinda want the circuit to trip at 15A. Note that 20A receptacles and plugs are slightly better engineered than the crap they make for lamps and whatnot - they're less prone to unplugging themselves. Still possible, but better. – Rich Dec 07 '20 at 23:52
  • 1
    a very brief "why" for *any* NEC-related question is usually "because people died or burned down buildings by doing this, so we have to disallow it". The NEC and other safety codes like it are 'written in blood'. – TylerH Dec 08 '20 at 15:55
  • I don't know of any published rationale for the rule, but I think a simple answer would be that if a circuit only drives a single 15A receptacle, using a 15A breaker would be slightly safer than using a 20A breaker, *and would have no downside*. In cases where having a 20A breaker would make a circuit more useful, the safety difference may be insufficient to outweigh the usability benefits of a 20A breaker, but in the single 15A receptacle care there is no usability difference. – supercat Jan 31 '22 at 18:06

3 Answers3

29

Let me give some framework around what ThreePhaseEel said.

I can understand wondering that, since UK appliances all have fuses in their plugs, and I bet the fuses are sized to protect the appliance.

The US doesn't have those. It depends on reasonably close matching between receptacle and fuse/breaker protection. Thus, the general rule is that receptacles must match breakers: 15A must be 15A, 30A must be 30A, and 50A must be 50A.

Now there are 2 exceptions. To brush aside the other one, NEMA stubbornly refuses* to define a standard for a 40A receptacle, so 40A circuits must out of sheer necessity use 50A receps.

The other is allowing 15A receps on 20A circuits. (the reverse is out of the question). Note that it's already the case that 15A appliances with 15A plugs are allowed on 20A circuits; just take a look at the socket keying.

enter image description here

NEMA 6 is our standard for 240V appliances; the same rules apply. See how 20A sockets are keyed to accept 15A plugs. Even on a 1-socket dedicated 20A circuit, 15A appliances are allowed.

Really, we have two separate use-cases here.

General-use branch circuits. As you know, the ubiquitous American recep is the NEMA 5-15. Most of the time, those are made as duplex (two socket) receps, and are on circuits that have many sockets - typically 3 to 12 yokes (a yoke is the physical chassis that fits in a standard 1-gang box) so 6-24 sockets. And most things plugged into them are low current - fans, cell phone chargers, TVs, etc. In this "general use appliance" application, long experience is that 20A is "close enough" to 15A that such appliances are unlikely to find useful protection by being fused at 15A rather than 20A. UL wouldn't approve an appliance that senstitive, anyway, or would require it to have internal fusing. (note that US Christmas light strings have fuses in the plug - yes, our dainty little plugs).

enter image description here

Dedicated circuits. Some appliances (think: built-in microwaves; air conditioners) are such heavy draw that they cannot share a circuit - i.e. they are running near circuit limits, or if continuous, near 80% of circuit limit Yet it is reasonable not to hardwire them and have them cord-and-plug connected. (e.g. a window air conditioner that might be removed every winter, back when they were inefficient). UL requires such circuits state in their instructions that they require a dedicated circuit. That is where simplex (1-socket) receptacles are typically used. When an appliance has such heavy draw, it makes more sense to require a close match between receptacle and breaker. So effectively UL is dictating the circuit protection.

It's a bit of a tag-team operation: our Electrical Code requires you used approved equipment (110.2), approvers defer to UL, UL writes the appliance design rules, UL certifies the appliances as safe, UL certifies the instructions and labeling, and Code requires that you obey them (110.3b).

*The reason they won't is: (rewind: in Europe, they standardized on 416V "wye" or 240V leg-neutral - while it makes for a more dangerous residential power, it allows easy 3-phase distribution and is strong enough for industrial needs; this means you only need one power system) Whereas Thomas Edison married the US to this lower-voltage split-phase thing. To work around its limitations, we went to a wide variety of schemes: 208/120V "wye" 3-phase, 240V "wild-leg" delta, and 480V delta (277V leg-neutral), just to name a few. As a result we have a huge variety of socket families, and they must ALL reject each other. So finding a new pin arrangement is like trying to find a .COM domain name lol.

Harper - Reinstate Monica
  • 276,940
  • 24
  • 257
  • 671
  • Comments are not for extended discussion; this conversation has been [moved to chat](https://chat.stackexchange.com/rooms/117049/discussion-on-answer-by-harper-reinstate-monica-why-does-us-code-not-allow-a-1). – Michael Karas Dec 08 '20 at 04:14
28

The key lies in the way UL tests 15A and 20A receptacles

Answering this question requires a deep dive into the standard UL tests 15A and 20A receptacle devices to, namely UL 498. In particular, we have to look at the way these receptacles are tested for acceptable loading, which is governed by Section 113 of the standard. We start with the fact that a 15A single receptacle is required to get no more than 30°C hotter than ambient at both its socket-contacts and its wire terminations when handling 15A, as per UL 498 113.1 and 113.5, as it only has a single set of terminal screws and thus can't be used to feed power through to other devices on the circuit.

However, duplex 15A receptacles, by and large, generally have two sets of hot and neutral terminals on them, thus permitting their use for through-wiring. (This goes hands in hand with the break-off tabs, or "fins", they have that allows them to be "split" into two separate receptacles.) Some of these receptacles are tested under 113.4, which requires their contacts to be tested at 15A but their wiring terminals to be tested for feeding 20A through with no more than a 30°C rise, as per 113.2.

More commonly, though, 15A receptacle devices are made to share a common set of parts with their 20A (or T-slot) counterparts, and thus fall under the effective exception in 113.4 that allows a 15A receptacle device with a 20A counterpart to be covered by the temperature testing done on the 20A receptacle. As a result, a 15A duplex receptacle's internal parts will be tested for handling 20A through them at some level, rendering them less susceptible to overheating from a low overload than a 15A single receptacle is as the latter has only ever been tested to not overheat excessively with 15A through it, never 20A.

ThreePhaseEel
  • 79,142
  • 28
  • 127
  • 220
  • 3
    I think "_As a result, a 15A duplex receptacle's internal parts will be tested for handling 20A_" (along with Dan A's suggested edit) has nailed my main point of confusion: **Q**: how does having a second socket/receptacle/"thing" make a 15A receptacle safer? **A**: In itself it _doesn't_, it's that a 15A _duplex_ receptacle has been tested to handle 20A better. Will mull things over and probably accept this in the morning. – TripeHound Dec 06 '20 at 20:53
  • I don't have access to UL 498 - does it mean that a 15 A plug is required to have a cord rated at 15 A, even if the connected equipment draws only 1 A under normal operation (so that the cord will still be protected by the circuit breaker)? – Andrew Morton Dec 08 '20 at 16:21
  • @AndrewMorton: For some reason, it's acceptable and commonplace in the US for unfused extension leads whose plug will fit into a 15A or 20A socket to include three 15A plugs at the far end, while using wire which is too thin which would be unable to safely pass 15A, much less 20A on a continuous basis. – supercat Dec 08 '20 at 18:31
  • @supercat The difference is partly due to the fact that the ratings for free conductors are different than for conductors in insulated walls. Most light-duty extension cords use 16AWG wire (smaller than the 14AWG in the wall), but a single 16AWG wire in "free air" needs 18A of current going through it to heat to 30C above ambient. In fact, for chassis wiring, 16AWG is rated to 22A, and 14AWG to 32A[60C+amb]. It's only in insulated walls (and in multi-conductor cables) that 14AWG derates to 15A and 16AWG to 10A - because the wires cannot shed heat as efficiently in that environment. – J... Dec 08 '20 at 20:56
  • @J...: Does "free air" include sitting on (partially sunk into) shag carpeting? How much would the temperature rise in 60 seconds at the maximum current a 20A breaker would be allowed to pass for 60 seconds without tripping? – supercat Dec 08 '20 at 21:02
  • 1
    @AndrewMorton -- UL498 stops at the plug. Appliance cordage in North America is generally treated as part of the appliance and thus protected by the appliance's internal protections against overload. – ThreePhaseEel Dec 08 '20 at 23:30
  • @supercat What makes you think that I've tested the temperature rise of a 16AWG cord "partly" sunk into a shag carpet under that specific condition? Suggest you go try it to find out. – J... Dec 09 '20 at 11:36
  • @supercat As for ["free air"](https://www.electricallicenserenewal.com/Electrical-Continuing-Education-Courses/NEC-Content.php?sectionID=787.0), it is finally well defined just this year, but would generally not have been considered sunk into carpet, no. – J... Dec 09 '20 at 12:01
  • @J...: I wasn't aware of how the concept of "free air" was defined; an extension cable suspended an inch of the ground would have much higher dissipation than one in a wall, but computing heat dissipation that way would seem overly optimistic even for cables which don't have anything placed on top of them. It may be reasonable to expect people to refrain from placing a rug on top of an extension cord, but it would be hard to avoid having a cord sink partially into shag carpet when powering something in the middle of it. – supercat Dec 09 '20 at 15:06
  • @J...: Further, even if a continuous 18A would be needed to cause a 30C temperature rise, an extension cord plugged into a 20A circuit could receive 24A indefinitely without tripping a breaker, which would cause a 53C temperature rise. I wonder if the rules for extension cords were written when most residential circuits were 15A rather than 20A? – supercat Dec 09 '20 at 16:21
  • @supercat No, you'll probably find that most extension cords are rated to something like 105C, so 53C over ambient is fine as long as ambient is below 52C. NM/Romex is 90C rated, typically, but code limits, afaik, are based on a 60C cable rating. Terminals are also usually only 75C rated, etc, so the rules change for a number of reasons. In any event, it is certainly *possible* for people to abuse extension cords and to get into trouble with them. Running a 1500W power tool over 150ft of 16AWG black extension cord strung across a tar roof at noon in the desert in July, for example... – J... Dec 09 '20 at 16:36
  • @J...: Even if a cable is rated to 105C, having an exposed cable reach 93C when ambient is 40C doesn't seem very safe. The thing that seems weird is that power strips are required to have breakers even though they typically have 14ga wiring, while extension cords with lighter-duty wiring aren't. Such an exemption might make sense for an extension cords with had one receptacle at the end, but many power strips with six devices plugged in are loaded far lightly than would be an extension cord with three coffee makers plugged in. – supercat Dec 09 '20 at 16:47
  • @supercat Well, you can take it up with UL, I guess. Power bars *can* have surge protection, but this isn't required. Circuit protection (breaker, etc) is also not required on power bars. Some supply it, others don't. As for temperature, ambient doesn't matter. The cable temperature is all that matters. If the cable is 90C and ambient is 60C it's the same situation as far as the cable is concerned as the cable at 90C and ambient at 20C. The cable is still at 90C, and good to go if it's 105C rated. If the cable's 115C, that's a problem - regardles of ambient. – J... Dec 09 '20 at 17:37
  • Why can't a single receptacle be built with two sets of terminals to allow through-wiring? – Vikki Aug 07 '21 at 18:23
  • 1
    @Vikki -- I'm not aware offhand of a reason you couldn't do that, but generally speaking it isn't done as a matter of practice – ThreePhaseEel Aug 07 '21 at 18:34
  • @ThreePhaseEel: ...yes, but _why_ not? – Vikki Aug 07 '21 at 19:05
  • 1
    @Vikki -- mostly because you only generally use a single receptacle device when you are putting a dedicated receptacle on a dedicated circuit; otherwise, there's no reason to use a single receptacle instead of a duplex – ThreePhaseEel Aug 07 '21 at 19:46
  • @ThreePhaseEel: "otherwise, there's no reason to use a single receptacle instead of a duplex" - what if (say) you don't have enough room for a duplex? – Vikki Aug 07 '21 at 19:50
  • 1
    @Vikki -- that's basically considered impossible -- a single receptacle uses the same yoke mounting pattern as a duplex receptacle, and they are also equivalent for Code fill purposes – ThreePhaseEel Aug 07 '21 at 19:52
2

Does not a single outlet on a 20A circuit rather imply at least a possible 20A load on that outlet?

I would guess that either inspectors frequently found, or those writing the code were concerned they would find, 20A circuits feeding single 15A receptacles from which 20A was being drawn. Presumably due to the installer (homeowner or an under-qualified 'electrician') not wanting to spend the time and effort to obtain a 20A receptacle, or 20A receptacles may not have been widely available at the time.

Perhaps a better regulation would be to allow a 20A breaker if the single receptacle was not the ONLY outlet on the circuit? However then you'd get the original problem, except now with duplex outlets added next to the electrical panel, essentially as decoration.

Having recently found the circuit feeding the kitchen (refrigerator, electric skillet, toaster, etc.) in my mother's home run through the link between two outlet halves (discolored and corroded by heat) in a duplex outlet that was loosely mounted, my sympathies lie with the regulators.

Technophile
  • 293
  • 1
  • 6