Continuous Duty - Residential Dwelling

[jaggedben]

The point of Article 220 is how to calculate load. Seems like an important step.

We agree! Which is why it was mistake to delete the reference to 220 in 215.2.

 

[jaggedben]

What Wayne is suggesting is Article 230.42(A)(1) is requiring a second calc in addition to the load calc, call it a continuous load calc, which I say is unenforcable since there is no guidance in 220.

...

Well said.

 

[wwhitney]

Where is this spelled out?

The scope of Article 220. It means that anywhere in the NEC, when you see the word "load" in the sense of "what's being supplied by a conductor" (as opposed to "a load" as in a piece of utilization equipment, or "load side" in the sense of a disconnect, etc.), you see Article 220.

I see that if you don't recognize this, the questions you are posing all arise. Once you recognize the above, there's is no ambiguity.

Cheers, Wayne

 

[wwhitney]

What Wayne is suggesting is Article 230.42(A)(1) is requiring a second calc in addition to the load calc, call it a continuous load calc, which I say is unenforcable since there is no guidance in 220.

No! Article 220 has no directives on conductor sizing. All it does is say "here's how you calculate some pieces of data that need to be attached to any conductor."

Then what 230.42(A)(1) is saying is "here's how you take those pieces of data, and use them to come up with a number which is the minimum conductor ampacity." 230.42(A)(1) is not a load calculation, it is a minimum ampacity calculation, which uses the outputs of Article 220 as its inputs. There is zero conflict with Article 220.

Then Article 310 tells you how to take that minimum ampacity and come up with the physical size of a conductor.

Article 422 long has declared a waterheater continuous just like EVSE and there are no examples of applying a continuous factor to a waterheater.

That's because none of the examples in Appendix D are branch circuit calculations (other than D8 for motors). All the examples involving water heaters are feeder or service calculations, and so the water heater doesn't have to be treated as a continuous load for those calculations. [And in fact those examples don't mention whether the water heaters are storage-type water heaters, since it is moot for the feeder/service calculations.]

Cheers, Wayne

 

[wwhitney]

Another point is that the 'sum of non-continuous plus 125% of continuous' formulation is overly simplistic. As various sections make clear, the 125% factor is to help overcurrent devices function properly at full continuous load. If a sufficient amount of the load is non-continuous (and arguably depending on an appropriate diversity factor for that non-continuous load), then including the 125% factor for continuous load should not be required. Because 100% of the combined load will never be continuous.

This is a bit tricky. Non-100% rated breakers in the following:

If we have a single load that can draw a solid 100A for 3.5 hours, it is continuous, and the NEC calls for a 125A OCPD to avoid nuisance tripping.

If we have a single load that can draw a solid 100A for 2.5 hours, and then drops down to 90A for an hour, it's not a continuous 100A load, and the NEC allows a 100A OCPD. Which doesn't make much sense to me--the I²R heating in the OCPD thermal element, integrated over the full 3.5 hours, is (2.5 + 1*0.9²)/3.5 = 95% of that in the first case. While the thermal trip element is presumably calibrated at about (100/125)² = 64% of the first case.

But if that single load were broken up into a continuous 90A load and a non-continuous 10A load, then the NEC would call for a 122.5A OCPD, i.e. back to a 125A OCPD.

Anyway, given that the NEC wants to use a linear approximation to some quadratic behavior, seems like for mixed loading, 125% of the continuous portion plus 100% of the remainder is a better fit between the two extremal cases (125% continuous only, or 100% non-continuous only) than (125% if all continuous; 100% otherwise).

Cheers, Wayne

 

[tortuga]

Then lets use Example D2(a) and the implied storage type waterheater as ramsy pointed out, it would be a stretch to say a storage type water heater is not a continuous load. I cant imagine that it is not a storage type water heater for a dwelling unit in the example.

2.5kW water heater element in Annex D, Example D2(a) would also be continuous load.

2500W cold start would take 3.5 hrs for typical residential 40 Gallon heater with water @ 50 Deg.F

Water Heating Calculator for Time, Energy, and Power

The calculators on this page compute how long it takes to heat water, how much energy is consumed, and how much heating power is required.

gettopics.com

Since water heater tanks maintain temperature, cold starts require power failures, so actual use should be considered.
* Average shower is 16 Gallons of water in 8 minutes @ 75% hot = 12gal x 2 people = 24gal
* Average washing machine is 7 to 20 Gallons per load @ 50% hot = 10gal
* Average dishwasher is 4 to 6 Gallons per load @ 80% hot = 5gal

Miser shower heads & efficient appliances use less water, but dwellings may take multiple showers in the morning, and some take 20 minute showers. Standard 4800W elements handle ~60gal in 3 hrs, per above calculator link.

Example D2(a) assumes <=35gal to get <=3 hrs with 2500W elements, not related to later applied demand factor in 220.82(B).

Wayne I am not saying your wrong but continuous dwelling load is can of worms that did not start with EVSE:

• Why would the example not show a continuous factor added ? It goes thru the troubble of showing the neutral calc?
• What about residential lighting load not covered by 210.70 (as referenced in 220.41) ? Security lights, yard, can lights etc?
• Has anyone been to a class or seen a textbook that shows a continuous load in a calc for a dwelling unit?
• And why even apply a demand factor in the first place?

 

[jaggedben]

The scope of Article 220. It means that anywhere in the NEC, when you see the word "load" in the sense of "what's being supplied by a conductor" (as opposed to "a load" as in a piece of utilization equipment, or "load side" in the sense of a disconnect, etc.), you see Article 220.

I see that if you don't recognize this, the questions you are posing all arise. Once you recognize the above, there's is no ambiguity.

Cheers, Wayne

Ok but that's still not showing me where things are 'spelled out'. At the very least, 215 and 230.42 should have informational notes. Because otherwise 'load' is just a dictionary definition (no Article 100 clarification), and a key concept of the code is likely to be missed.

What actually cinches for me that you are correct is examples in Annex D. But they're examples I never looked at before because I don't do commercial load calcs, only resi ones and I use 220.8x sections. And the examples for those sections don't separate continuous and non-continuous loads.

No! Article 220 has no directives on conductor sizing. All it does is say "here's how you calculate some pieces of data that need to be attached to any conductor."

Then what 230.42(A)(1) is saying is "here's how you take those pieces of data, and use them to come up with a number which is the minimum conductor ampacity." 230.42(A)(1) is not a load calculation, it is a minimum ampacity calculation, which uses the outputs of Article 220 as its inputs. There is zero conflict with Article 220.

...

I think it's problematic that 230.42(A)(1) and 215.2(A)(1) are awfully similar to dozens of other code requirements that refer to current and not load. That's a subtle distinction (and a bit illogical even, when you consider that 220 gives answers in VA and not amps.) I know that the code is not supposed to be a how-to manual but in this case explaining now to implement them via 220 is really an element of the requirements. Especially since there's also still an open question or two. First, does 220.80's blanket 'shall be permitted' and the 220.8x sections' total lack of mention of continuous factors effectively override 215.2(A)(1) and 230.42(A)(1)? Upthread you said the code makes no distinction between residential and commercial, but perhaps that 'shall be permitted' is it. Note that none of the 220.8x examples in Annex D make a distinction between continuous and non-continuous loads. And substantively, going back to my post 31, I would argue that they shouldn't. So should 215.2(A)(1) and 230.42(A)(1) have exceptions for load calculated in accordance with 220 Part IV?

 

[wwhitney]

Ok but that's still not showing me where things are 'spelled out'. At the very least, 215 and 230.42 should have informational notes. Because otherwise 'load' is just a dictionary definition (no Article 100 clarification), and a key concept of the code is likely to be missed.

Agreed that clarity would be improved if, say, "calculated load" were defined in Article 100 to refer to the values generated by Article 220, and sections such as 230.42, 215.2, etc. were amended to use the new term "calculated load".

I think it's problematic that 230.42(A)(1) and 215.2(A)(1) are awfully similar to dozens of other code requirements that refer to current and not load.

Maybe. The fact they use the term "load" and not the term "current" is an indication that something different is going on. But a defined term like "calculated load" would be even clearer.

First, does 220.80's blanket 'shall be permitted' and the 220.8x sections' total lack of mention of continuous factors effectively override 215.2(A)(1) and 230.42(A)(1)?

No, Article 220 hardly mentions a 125% continuous factor (and in the 2017 NEC, never mentions it), so the absence in 220.8x does not signify anything. This is the main problem I see in Article 220--the other articles that are users of the output(s) of Article 220 often want a breakdown of the load into continuous and non-continuous portions, while Article 220 doesn't provide any guidance on how to do that (as mentioned in the second paragraph of my post #24.)

The "shall be permitted" is just to let you know that 220.8x is an alternative to the usual Part III algorithm.

And substantively, going back to my post 31, I would argue that they shouldn't. So should 215.2(A)(1) and 230.42(A)(1) have exceptions for load calculated in accordance with 220 Part IV?

I'm not convinced by post 31, I attempted to articulate my thoughts in post 45. Which maybe I can rephrase as follows, where "normal OCDP" means non-100% rated:

Say we have 100A total load. At one extreme it is all non-continuous, and we are allowed a 100A normal OCPD. At another extreme it is all continuous, and because of the prolonged heating in the OCPD thermal element possibly causing nuisance tripping, we must use a 125A normal OCPD. Clearly in between, the amount of heating we get in the thermal element lies in between the two extremes, and increases with an increasing proportion of continuous load. So why shouldn't the minimum OCPD size likewise lie in between the two extremes, and likewise increase with an increasing proportion of continuous load?

Cheers, Wayne

 

[jaggedben]

This is a bit tricky. ...

Of course it is, but I'm not really interested in the examples in your responses so far. I'm more interested in a realistic resi example, where perhaps there's a continuous 40A of car charging, plus some much smaller continuous amount of electronic and lighting load, and the only thing that really comes up against what the service needs to be sized for is if they run the electric oven and dryer at the same time, but those things cycle off after no more than 10 minutes. In this situation is adding a 125% factor to the 40A car charge at all meaningful to what the OCPD can handle?

This is the main problem I see in Article 220--the other articles that are users of the output(s) of Article 220 often want a breakdown of the load into continuous and non-continuous portions, while Article 220 doesn't provide any guidance on how to do that (as mentioned in the second paragraph of my post #24.)

Right, and particularly with respect to 220.82 or .83 methods I don't see how it's possible to deconflict this except to ignore the articles that want those breakdowns. Admittedly in my post #30 I made a blanket statement about Article 220 that was mistaken. But it's not mistaken at all with respect to most of Article 220 Part IV. Hence my suggestion of making explicit exceptions.

 

[jaggedben]

... In this situation is adding a 125% factor to the 40A car charge at all meaningful to what the OCPD can handle?

To make a simplified example along the lines of yours... What if the load is likely to be max 50A continuous, with an additional 50A that will never be continuous? Maybe if the continuous load is less than 50% of the total the 125% factor should not be required?

 

[tortuga]

Thats reason I say Wayne is suggesting Article 230.42(A)(1) is requiring a second calc (call it the continuous load calc) based on post 24:

Article 220 never mentions a continuous use factor (or maybe there's a small exception to that), it's just other articles that care about what fraction of the load is continuous. So whenever you apply Article 220 you have to keep track through the summations and demand factor what part of the total load is continuous.

Since there is no guidance in 220 in calculating a continuous load,
we don't apply the continuous factor to branch circuits as inputs to a 220 calc,
and 230.42(A)(1) does not have any demand factors for continuous loads
you'd just have to sum up all the continuous loads and multiply by 25% and add that to the load calc from 220 for sizing the service entrance conductors.

 

[wwhitney]

Thats reason I say Wayne is suggesting Article 230.42(A)(1) is requiring a second calc (call it the continuous load calc) based on post 24:

Since there is no guidance in 220 in calculating a continuous load,
we don't apply the continuous factor to branch circuits as inputs to a 220 calc,
and 230.42(A)(1) does not have any demand factors for continuous loads
you'd just have to sum up all the continuous loads and multiply by 25% and add that to the load calc from 220 for sizing the service entrance conductors.

There's nothing special about 230.42(A), as 215.2(A) and 210.19(A) are the same.

The idea that you'd use the total connected load for the "continuous load" input to the formulas in the above section obviously gives you non-sensical answers. E.g. take a commercial application where all the lighting is a continuous load. What's the point of Table 220.42 "Lighting Load Demand Factors" if you're just going to take the total connected load anyway?

In other words, the fact that Article 220 is silent on continuous loads means it is agnostic with respect to that question, not that it only applies to non-continuous loads.

The way I think of this is that when applying Article 220, each basic load number that you get from a nameplate or a square footage from an architectural plan or whatever comes along with a label like "continuous" or "non-continuous" or "continuous only for branch circuit calculations". And you just track that label through all your computations. [Or if you prefer a different formalism, the basic load numbers are all ordered pairs, and you're adding and scaling ordered pairs at each point.]

That method works great and is well-defined as long as the final result is a just a weighted linear combination of the inputs. Where it breaks down is when you have to use a function like max(a,b) in the calculation.

Cheers, Wayne

 

[wwhitney]

To make a simplified example along the lines of yours... What if the load is likely to be max 50A continuous, with an additional 50A that will never be continuous? Maybe if the continuous load is less than 50% of the total the 125% factor should not be required?

OK, so lets say the 220.82 load calc comes out as 100A, and you determine that 50A of that is continuous (which requires resolving the issue raised in post #24, but let's say you've resolved it somehow).

Surely for the purposes of applying 215.2, 230.42, and 210.19 we aren't supposed to consider the real world facts that this means that 99.9% of the time across all such residential services, the actual current is going to be way under 100A. We have to treat that answer as meaning the current may reach 100A at times, but that in reaching 100A, the worst case would be 50A that lasts 3 hours or more, plus another 50A that would fluctuate at some point during that 3 hours and get lower.

So with that framing, what's wrong with the argument at the end of post #48? We can compare the above situation to (a) 100A all non-continuous, and fluctuating the same way as the 50A in your example and (b) 100A all continuous. In case (a) we have confidence that a 100A OCPD won't nuisance trip. In case (b) the chance of nuisance tripping is judged to be too high, so a 125A OCPD is required.

I don't know what the chance of nuisance tripping will be in your example, but clearly it's strictly higher than case (a), and strictly lower than case (b). So it makes perfect sense to me that as it is higher than case (a), we have a requirement to upsize the OCPD; but that as it is not as high as case (b), we don't have to upsize the OCPD all the way to 125A (ignoring the granularity of breaker sizes).

Cheers, Wayne

P.S. Depending on ambient temperature, density of breakers in a panelboard enclosure, where breaker's trip curves actually land within the allowable range under the UL standard, etc, it's entirely possible that we could ignore the 125% continuous use factor altogether and never have any nuisance tripping. But this is a discussion about what the NEC requires, using a particular idealized model of how things work, so such realistic details are not relevant, as in the 2nd paragraph above.

 

[ramsy]

For purpose of the OP's inspector question the use of the calculations in 220.83 should give him the answer.

That inspector really knows how to stir the pot, ours can't tell the difference between a GFCI and AFCI.

 

[jaggedben]

Wayne, let me state my objection to including continuous factors in 220 Part IV load calcs in a much more practical way.

Given my lived experience of doing 220.82 load calcs and then seeing customers consumption data that never gets to more than about 60% of the load calc for the average over a 15min interval, let alone 3 hours ....

I don't want my Art 220 Part IV load calcs to be further penalized.

Basically as simple as that.

If I add the extra 25% of the continuous loads to the load calc, I may have to upsize my service. But if I were to add it to their real consumption ex post facto, I wouldn't. So I still have plenty of cushion. And then what is the point of parsing the math theory on the proportionality of the continuous loads if we already know that the Part IV load calcs are so conservative it may never make a difference in the real world?

Granted my sample size is small, so maybe we don't really know. Where is the data that those sections are based on? But that's where my skepticism comes from. And it's why I'm happy to keep going with tortuga's assertion that including continuous load factors in 220.82 or 220.83 load calcs is unenforceable (since we don't know how we're supposed to do it anyway) at least until some AHJ really questions me on it.

 

[wwhitney]

Given my lived experience of doing 220.82 load calcs and then seeing customers consumption data that never gets to more than about 60% of the load calc for the average over a 15min interval, let alone 3 hours ....

I don't want my Art 220 Part IV load calcs to be further penalized.

Fair enough, and that's a very practical approach. My point is that it's not very defensible if pushed by an AHJ. The NEC solution for your situation is to do a 220.87 load study.

Cheers, Wayne

 

[ramsy]

Thats reason I say Wayne is suggesting Article 230.42(A)(1) is requiring a second calc (call it the continuous load calc) based on post 24:

Since there is no guidance in 220 in calculating a continuous load,
we don't apply the continuous factor to branch circuits as inputs to a 220 calc,

I believe there is a distinction made between current rise vs IR2 rise.

Continuous inductive motor-only load does increase current, consistent with the extra 25% for 220.18 inputs with motor-only loads.
220.18(A) guides us to Art. 430 for 125% motor-only loads, and T220.102 for Farm Loads: "125% of the full load current of the largest motor, or First 60 amperes"

Continuous resistive fixed-electric heat & water heaters will only increase IR2 loss, and Article 220.82 excludes this extra 25%.

ESVE's are also inductors, perhaps with a resistive componant, but are not defined like an HVAC nameplate?

Battery chargers use inductive load inverters for DC batteries, which cause a current rise like inductive motors. Further, wireless inductive chargers cause more inductive current rise than just the inverter alone.

Nameplates have been required for combined resistive & inductive loads, such as lighting 220.18(B), Welders 630.11(B), or HVAC 220.18(A), but ESVE chargers don't require nameplate in Art. 625, or 220.

So, ESVE loads may rely on 625.42 rating "shall be considered to be continuous loads", for increased inductive current rise, since a significant inductive component exists with ESVE inverters + any wireless inductor.

and 230.42(A)(1) does not have any demand factors for continuous loads

There is a demand factor for existing buildings
220.87(2) When calculating max demand for existing buildings: "The maximum demand at 125 percent plus the new load.."

In addition to non-residential demand factors for Welders in Appendix D3(a), and Elevators in Appendix D10

 

[jaggedben]

Fair enough, and that's a very practical approach. My point is that it's not very defensible if pushed by an AHJ. The NEC solution for your situation is to do a 220.87 load study.

I can't do a 220.87 load study on a new service or one I don't have data for or one where the house is being remodeled and all the appliances are changing.
I'm saying to the extent I've been able to look at data for projects completed a year or more ago, for which I originally did a 220.82 load calc, the calcs are already conservative enough.

 

[tortuga]

Update turns out a Level 2 EVSE will not see its maximum current for 3 hours or more when charging lead acid or lithium.
The charge algorithm is constant current for a while then constant voltage, meaning the current will vary.

 

[wwhitney]

Update turns out a Level 2 EVSE will not see its maximum current for 3 hours or more when charging lead acid or lithium.
The charge algorithm is constant current for a while then constant voltage, meaning the current will vary.

Agreed on the last sentence, but not the first. The constant current regime can easily last 3 hours or more. The final constant voltage regime may only be an hour (total guess).

Level 2 EVSEs are definitively continuous loads, when the car's on-board charger is rated at least as high as the EVSE, and when the car is quite depleted so it needs well over 3 hours to charge. Which is an "expected" combination of events.

Cheers, Wayne