210.21(B)(3) 20amp duplex receptacle on 15amp circuit

Scurry

Member
Location
Florida
Occupation
Retired Electrician
I asked this question on one of Mike Holt's recent facebook posts and he instructed me to ask here.

Table 210.21(B)(3) used to say "not over 15" for receptacle ratings on 15 amp circuit ratings. Which meant that 20 amp receptacles could not be used on 15 amp circuits with multiple receptacles/outlets.

NOW, Table 210.21(B)(3) simply says "15" for receptacle ratings on 15 amp circuit ratings. And Code 210.21(B)(3) says "receptacle ratings shall not be less than the values listed in Table 210.21(B)(3)". Does that mean that 20 amp receptacles are now permitted on 15 circuits? I no longer see anything prohibiting it. I know the intention was to keep people from plugging a 20 amp load into a 15 amp circuit, so maybe there's another code which now prevents it?
 
This was a change in the 2023 NEC. 210.21(B)(3) used to say that receptacle ratings "shall conform to" the table values. Now it says that they "shall not be less than" the table values.

Since "conform to" was generally interpreted to mean "match", this represents a change to what is allowed. Whether the CMP in charge of this code section meant "match" originally or not is unclear, but it is clear that under the 2023 NEC you can put a 20A receptacle on a 15A circuit supplying two or more receptacles.

Edit: Hmm, 210.24 was not updated, so there is a conflict within the 2023 NEC. Still 210.24 is meant to summarize the earlier sections, so I would think the actual text of 210.21(B)(3) would take precedence. Oddly, no one picked up on this discrepancy, so it does not appear that the 2026 NEC will fix it. Good topic for a PI for the 2029 NEC.

Cheers, Wayne
 
Last edited:
Edit: Hmm, 210.24 was not updated, so there is a conflict within the 2023 NEC. Still 210.24 is meant to summarize the earlier sections, so I would think the actual text of 210.21(B)(3) would take precedence. Oddly, no one picked up on this discrepancy, so it does not appear that the 2026 NEC will fix it. Good topic for a PI for the 2029 NEC.
Wayne, any idea why this change was made in the first place?
 
Wayne, any idea why this change was made in the first place?
No, IIRC the PI I read that led to this more or less said "a 30A single receptacle is allowed on a 15A individual branch circuit, so obviously multiple such receptacles on a general purpose branch circuit should also be allowed."

Cheers, Wayne
 
This is a stupid code change. Am I missing something or are the CMP members so incompetent that they could figure this one out? Why even make 15 amp duplex receptacles if you can put 20 amp duplex receptacles on a 15 amp circuit?

For a single receptacle on an individual branch circuit this makes sense because you don't have a receptacle configuration to match every standard branch circuit OCPD size. For example there is no 40 amp receptacle so for a 40 amp range circuit you would need to use a 50 amp receptacle with a 50 amp cord and plug.
 
It appears that the CMP did not really understand the results of the accepted new language. The PI was not attempting to do this.
I don't think any of the code changes books for the 2023 code caught this. I first saw this in a facebook post a month or two ago, maybe a post by the OP in this thread.
Here is the substantiation and panel statement for the 2023 change.
Statement of Problem and Substantiation for Public Input

This public input, along with the others I am providing, is a direct consequence of my investigation of a fire in a GFCI receptacle mounted on the outside of my house which could have been catastrophic, but thankfully wasn’t. I am a licensed engineer. A detailed report was prepared. The existing Table 210.21(B)(3) has an entry for “not over 15” and an entry for “40 or 50,” both under the column heading “Receptacle Rating.” There are no NEMA blade configurations for receptacles rated less than 15 amps or for 40 amp receptacles. The existing table by showing “not over 15” suggests that receptacles less than 15 amps are acceptable. This understanding is not correct. In addition, 406.3(B) specifically states that receptacles shall be rated not less than 15 amps. Receptacles rated 50 amps shall be limited to 50 amp circuits. In general, the ampacity of receptacles should correspond with the ampacity of circuits. A receptacle undersized for the circuit can become a safety problem through overloading which doesn’t trip the circuit breaker, while a receptacle oversized for the circuit can result in nuisance tripping which could result in a safety problem in terms of operation of equipment, or at least an inconvenience. The unique problem of 15 amp receptacles on a 20 amp circuit is addressed separately in another public input I am making but for the purposes of matching receptacles with circuits, it is necessary that the rating of a receptacle correspond with the blade configuration of the receptacle. However, this does not prevent a NEMA 5-15R receptacle, for example, from having an appropriate internal design to allow placement in a 20 amp circuit with full protection from a 20 amp breaker. The usual method of circuit sizing is to determine load, then the wire size based on the sum of the non-continuous load and 125% of continuous load (with consideration for adjustment factors). Then the breaker size is selected to match wire size, except the breaker size can be bumped up to the next standard size if the calculated breaker size would otherwise fall between two standard sizes. Therefore, it should follow that the receptacle size is chosen first, and then the circuit size matched accordingly. It is understood that accepted practice often determines the number of receptacles that are known to be workable for a particular circuit rating. This approach is often the determining factor for use of 20 amp circuits with 15 amp receptacles, the approach the existing table takes. A line for a 60 amp receptacle, with a 60 amp circuit rating, was added to complete the listing of the available amperage ratings of receptacles in accordance with NEMA blade configurations. It is acknowledged that Part II. Branch Circuit Ratings, 210.18 Rating., specifically recognizes 15, 20, 30, 40, and 50 amp branch circuits. The new table proposed eliminates the problem of having no receptacle rating available for a 40 amp circuit. Lastly, the present table Table 210.21(B)(3) duplicates information presented in Table 210.24 and therefore it is redundant.
Committee Statement
The text in the table has been modified to match existing receptacles that are available. The title of the table has been changed to make it clear that this table applies to circuits with more than one receptacle or receptacle outlet as the parent language of the table states.[/quote

There was a Public Comment for the 2026 that would have fixed this, but the second draft meeting minutes show this as reject so it would not have gone to the formal written ballot. The reject statement in the meeting minutes said:
Insufficient substantiation was provided to required receptacle ratings to conform to Table 210.21(B)(3). The branch circuit should be
protected by the proper sized OCPD.
 
What's everyone's opinion on how 406.4(A) applies here?

"Except as provided in 406.4(D), receptacles installed on 15- and 20-ampere branch circuits shall be of the grounding type. Grounding-type receptacles shall be installed only on circuits of the voltage class and current for which they are rated, except as provided in 210.21(B)(1) for single receptacles or Table 210.21(B)(2) and Table 210.21(B)(3) for two or more receptacles."

This code references the table only, which gives exact ratings and does not include the text "shall not be less than". Since this code is stricter than the text 210.21(B)(3), would it not override it?
 
This code references the table only, which gives exact ratings and does not include the text "shall not be less than". Since this code is stricter than the text 210.21(B)(3), would it not override it?
I would agree that the unchanged text in 406.4(A) does prevent you from using the new flexibility of the wording change in 210.21(B)(3). Which makes the code change to 210.21(B)(3) even stupider.

Cheers, Wayne
 
Top