Electrical
Electrical load calculation per NEC Article 220
Total the connected load, apply the demand factors, and the calculation returns the minimum service or feeder size, not the impossible sum of every nameplate at once.
Direct answer
An electrical load calculation totals the connected load, applies the NEC demand factors, and returns the minimum service or feeder size in amps. It works because no building runs every load at once, so the diversified demand, not the impossible sum of all nameplates, sizes the gear. The adopted code edition and the AHJ control.
Key takeaways
- An NEC Article 220 load calculation totals connected load, applies demand factors, and returns the minimum service or feeder size in amps.
- Dwelling general-lighting demand counts the first 3000 VA at 100 percent, 3001 to 120,000 VA at 35 percent, and the rest at 25 percent.
- Count only the larger of the heating or air-conditioning load, never both, because they are noncoincident under NEC 220.60.
- A single household range rated 12 kW or less counts at 8 kW from Column C of Table 220.55, not its nameplate.
- The 2026 NEC moved Article 220 into Article 120 and cut the dwelling lighting load from 3 VA to 2 VA per square foot; confirm the adopted edition.
The load calculation, and why it sizes the service
An electrical load calculation is the procedure that totals a building's connected load, applies the code's demand factors, and gives you the minimum size of the service or feeder that has to supply it. The result is a number in amps. That number sets the service conductors, the main breaker, the panel rating, and the size of the gear you order.
The reason the calculation exists at all is that you cannot just add up every nameplate and size the service to the sum. Nobody runs the range, the dryer, the water heater, the air conditioner, and every receptacle at full tilt at the same instant. The code recognizes that with demand factors, which let you count a fraction of certain loads instead of the whole thing. Size the service to the impossible all-at-once total and you pay for copper and gear that will never carry the current. Size it to the diversified demand and it fits the real load with margin to spare.
Where this goes wrong is rarely the arithmetic. It is feeding the calculation the wrong inputs: a nameplate where a demand figure belonged, the running current of a continuous load instead of 125 percent of it, or the smaller of the heat and the air conditioner when the rule wants the larger. Then the service is undersized on paper and oversized in the wallet, and nobody can reconstruct how the number was reached.
The NEC Article 220 framework: standard and optional methods
Load calculations live in the NEC under Article 220, Branch-Circuit, Feeder, and Service Load Calculations. It gives two routes to a dwelling service size. The standard method, the long Part III procedure, takes each category of load on its own, applies its own demand factor, and adds the results. The optional method, in Part IV at 220.82 for a single dwelling, lumps most of the load together and applies one pair of percentages. They are two roads to the same kind of answer, and the optional method usually lands lower.
Everything in the calculation is in volt-amperes, not watts. VA is voltage times current without the power-factor correction, which is what sizes a conductor, because the conductor carries the current regardless of how much of it does real work. You build a total in VA, then divide by the system voltage to get the amps the service has to carry.
One note that matters in 2026 and after: the NEC reorganized this material. Through the 2023 edition it is Article 220. The 2026 edition moved branch-circuit, feeder, and service load calculations into a new Article 120 in Chapter 1. The math is the same, but the numbers shift. Most jurisdictions are still enforcing the 2020 or 2023 NEC, so confirm both the adopted edition and any local amendments before you cite a section on a submittal.
The general lighting load
The general lighting load is a flat unit load applied to the floor area, and it covers the general lighting and the general-purpose receptacles together. You do not count fixtures. You take the square footage and multiply by the VA per square foot the code assigns to that occupancy. For a dwelling that figure has long been 3 VA per square foot, computed from the outside dimensions and excluding open porches, garages, and unfinished space not adaptable for future use.
Watch the edition here. The 2026 NEC cut the dwelling unit figure from 3 VA per square foot to 2, on the reasoning that LED lighting draws a fraction of what incandescent did. So a 2000 sq ft house calculates at 6000 VA under the 2020 and 2023 editions and at 4000 VA under the 2026 edition. That is a real difference in the service size, and which one is right depends entirely on the code your AHJ has adopted.
Other occupancies carry their own unit loads in the general-lighting table, and they vary by building type. Pull the figure from the table for the occupancy you are calculating, in the edition that governs the job, rather than from memory, because the values have moved between cycles.
VAlighting = Area (sq ft) × unit load (VA/sq ft)What is a demand factor?
A demand factor is the fraction of a connected load you are permitted to count toward the service or feeder, because that load will not run at full output at the same time as everything else. It is the code's way of pricing in diversity. A demand factor of 35 percent on a block of load means you carry 35 percent of it into the total and leave the rest out.
The demand factors are not guesses. They are in the Article 220 tables, set per category from decades of metered data on how buildings actually draw. General lighting has a demand table. Household ranges have one. Dryers, fixed appliances, commercial kitchen equipment, multifamily feeders all have their own. The dwelling general-lighting demand, for example, takes the first 3000 VA at 100 percent, the portion from 3001 to 120,000 VA at 35 percent, and anything above that at 25 percent.
The discipline is knowing which loads get a demand factor and which do not. Some loads you carry at 100 percent because they genuinely can run full and continuous: a continuous load gets no diversity break, it gets the opposite, a 125 percent adder. The mistake that runs through bad calculations is applying a demand factor where the code wants the full load, or carrying a load full where a demand factor was allowed and the service came out a size too big.
Continuous loads and the 125 percent rule
A continuous load is one expected to run at its maximum for 3 hours or more. The long-standing rule across the NEC is that the conductor and the overcurrent device serving a continuous load are sized at 125 percent of the continuous current, plus 100 percent of the noncontinuous. That 125 percent is in the feeder and service minimums at 215.2 and 230.42, and it lands in the load calculation the same way: the continuous portion goes in at 125 percent.
The reason is heat. A load that comes and goes lets the conductor cool between draws. A load that sits there for hours holds the conductor at temperature, so the code derates by sizing the conductor and breaker up. This is the same physics that drives the continuous-load adder on voltage drop, where the steady current is what builds the drop, and the same 125 percent that sizes a motor branch conductor off its full-load current.
The 2026 edition complicates the blanket statement. Reports on the reorganized Article 120 indicate the 2026 NEC changed how continuous loads enter the load calculation, easing the across-the-board 125 percent in the calculation itself. The 125 percent continuous rule still governs branch-circuit, feeder, and OCPD sizing broadly, so do not drop it on instinct. Confirm the exact treatment against the edition your jurisdiction has adopted.
Receptacle, small-appliance, and laundry loads
In a dwelling, the general-purpose receptacles are already inside the general lighting unit load, so you do not count them again. What you do add are the small-appliance and laundry branch circuits, which the code carves out separately. A dwelling requires at least two small-appliance branch circuits for the kitchen and dining areas and at least one laundry circuit, and each of those is counted at 1500 VA in the calculation.
So two small-appliance circuits and one laundry circuit add 4500 VA. Those go on top of the general lighting figure, and then the whole block of general lighting plus small-appliance plus laundry gets the general-lighting demand factor together. That grouping matters. The small-appliance and laundry loads ride the same demand table as the lighting, not their own.
Nondwelling receptacles are a different animal. In a commercial building you do not get the lighting unit load to absorb the receptacles. Each receptacle outlet is counted at not less than 180 VA per single or multiple device on one yoke, and that total carries its own demand factor for the portion above a threshold. The dwelling lumps receptacles in. The commercial calc counts them one at a time.
Appliances, ranges, and dryers
Fixed appliances and cooking equipment each follow their own rule, and the trap is treating them all alike. Fastened-in-place appliances other than the range, dryer, space heating, and air conditioning, things like the water heater, dishwasher, and disposal, go in at nameplate. Where there are four or more of them on a dwelling service, the code lets you take a demand factor on the group, commonly 75 percent at 220.53.
Household electric ranges have their own table, 220.55, and it is one of the least intuitive in the book. One range rated 12 kW or less is not counted at 12 kW. It is counted at a maximum demand of 8 kW from Column C. Larger ranges and multiple ranges follow the columns and notes, but the single-range answer most crews need is 8 kW.
Electric clothes dryers run off 220.54: count 5000 VA or the nameplate, whichever is larger. A 5.5 kW dryer goes in at 5500 VA because the nameplate beats the 5000 floor. Multiple dryers in a multifamily building get a decreasing demand factor as the count climbs. The point across all three categories is the same: the nameplate is the starting input, but the table, not the nameplate, is usually the number that lands in the total.
| Load | How it enters the dwelling calc | Reference |
|---|---|---|
| General lighting + receptacles | Area times unit load, then general-lighting demand | 220.12, 220.42 |
| Small-appliance circuits | 1500 VA each, with the lighting block | 220.52 |
| Laundry circuit | 1500 VA, with the lighting block | 220.52 |
| Fixed appliances (4 or more) | Nameplate, then 75 percent demand | 220.53 |
| Household range, 12 kW or less | 8 kW from Column C | Table 220.55 |
| Electric dryer | 5000 VA or nameplate, larger of the two | 220.54 |
How do you handle heating versus air conditioning?
You count the larger of the heating load or the air-conditioning load, not both, because they are noncoincident: the building does not heat and cool at the same instant. The rule is 220.60, noncoincident loads, which lets you carry only the largest of loads that cannot run together. Take the electric heat at its load, take the air conditioning at its load, compare them, and put the bigger one in the total. The smaller one drops out entirely.
Get the comparison right and it is the single biggest demand break in many calculations. A 10 kW electric furnace beats a 5 kW air conditioner, so you carry 10 kW and ignore the AC. Reverse the building and the AC wins. The error that doubles a service is adding both, which counts a condition that physically cannot occur.
One catch from recent editions: if the air-conditioning side contains the largest motor in the building, the hermetic compressor gets its 125 percent adder before you compare it to the heat. So you are not always comparing two raw nameplates. You compare the heat load against the air conditioning sized the way the code requires it to be sized, then keep the larger. Confirm the motor treatment against the adopted edition.
Motor loads in the service calculation
Motors enter the feeder and service calculation through 430.24, and the rule has a twist that separates it from a plain sum. You take 125 percent of the full-load current of the largest motor, plus 100 percent of every other motor on the feeder, plus the rest of the load. Only the single largest motor gets the 125 percent adder. The others go in at full load, once each.
The 125 percent on the largest motor is headroom for starting and continuous running of the one machine most likely to be the limiting case. It is not a per-motor multiplier. Apply 125 percent to every motor and you oversize the feeder, which is a common and expensive misread of the rule.
The full-load current you use for the largest motor is the table value from Article 430, not the nameplate, the same distinction that governs sizing the motor branch circuit itself: the conductor and the branch short-circuit device run off the NEC table full-load current, while only the overload runs off the nameplate. On a building with a dominant motor load, a chiller plant or a large pump, this rule and the largest-motor adder can drive the service more than the lighting and receptacles combined.
What is the optional method for a single dwelling?
The optional method is a shorter dwelling calculation that gathers most of the load into one pile and applies a single pair of percentages. It is in 220.82, and it is allowed for a dwelling unit served by a single 120/240 V or 120/208 V 3-wire service rated 100 A or larger. When the house qualifies, you may use either the standard method or the optional method, and most estimators run both and take the smaller legal number.
Here is the move. Add the general lighting at the unit load, the two small-appliance circuits and the laundry at 1500 VA each, and the nameplate of every permanently connected appliance, the range at its nameplate, the dryer, water heater, dishwasher, and so on. Apply 100 percent to the first 10,000 VA of that pile and 40 percent to everything above 10,000 VA. Heating and air conditioning are handled separately under their own percentages, taking the largest applicable.
The optional method usually wins because that flat 40 percent on the bulk of the load is a deeper diversity break than the category-by-category demand factors of the standard method. For a typical 2000 sq ft all-electric house the optional method can land near 110 A where the standard method lands around 145 A. Same house, same loads, smaller service, because the method takes more diversity. The condition is the 100 A and single-service requirement, so a large or unusual dwelling may not qualify.
Service, feeder, and branch: where the calculation lands
The same machinery runs at three levels, and what changes is the scope of load. The service calculation totals the whole building and sizes the service-entrance conductors and the main. A feeder calculation totals only the load downstream of that feeder, a subpanel or a wing, and sizes the feeder conductors and its overcurrent device. A branch-circuit calculation sizes a single circuit to its connected load. Smaller scope, same demand factors where they apply.
Where the levels meet is the demand factor question. A demand factor taken at the service is taken on the whole building's load. Take it again on a feeder that only serves part of the building and you may be double-dipping diversity that does not exist at the smaller scope. The cleaner habit is to apply the demand factors at the level the table is written for, then check that the sum of the feeders does not understate what the service has to carry.
Feeders carry their own minimum at 215.2, services at 230.42, and both pick up the continuous-load and ampacity rules. A feeder sized exactly to its calculated load with no headroom is a feeder that gets re-pulled the first time a subpanel grows, which is why feeders are where spare capacity earns its keep.
The feeder and service neutral load
The neutral does not have to be sized to the full calculated load. It carries only the unbalanced load, the maximum difference between the neutral and any one ungrounded conductor, which is 220.61. On a 120/240 V single-phase service, a pure 240 V load like a water heater or a straight electric range element puts nothing on the neutral, so it drops out of the neutral calculation even though it sized the ungrounded conductors.
Two reductions sharpen the neutral. The code permits a 70 percent demand factor on the neutral load contributed by household ranges, wall ovens, cooktops, and dryers, because their neutral current is a fraction of their line current. And for the portion of the unbalanced load above 200 A on the qualifying systems, you may take a 70 percent demand factor on that excess. So a large service can run a neutral a size or two smaller than the phase conductors.
The hard stop is nonlinear load. On a 3-phase, 4-wire wye system feeding electronic and other nonlinear loads, you cannot take the 70 percent reduction on the part that is nonlinear, because the harmonic currents add on the neutral instead of canceling. On a building full of computers, drives, and LED drivers, the neutral can carry as much as a phase, and undersizing it is how you cook a neutral that the simple unbalance math said was fine.
Commercial and other nondwelling calculations
Nondwelling calculations drop the dwelling shortcuts and count more things explicitly. Receptacles come in at not less than 180 VA each and carry their own demand factor above a threshold. Show-window lighting is counted at not less than 200 VA per linear foot measured along the base. Sign and outline lighting carries a minimum of 1200 VA for the required branch circuit. None of these has a dwelling equivalent, because the dwelling unit load swept them up.
Commercial kitchen equipment gets a genuine break. Table 220.56 applies a demand factor to electric cooking and kitchen equipment by the number of units, from 100 percent at one or two units down to 65 percent at six or more, with the floor that the result cannot be less than the sum of the two largest pieces. A restaurant with a long equipment list is where that demand factor pays for itself, and where forgetting it doubles the kitchen feeder.
High-density loads break the table-driven mindset. A data hall does not diversify the way an office does, because the racks run flat out around the clock. There the connected load and the design load converge, the continuous treatment governs, and the calculation leans on the actual rack power and redundancy scheme rather than on a generous demand factor. When the load is engineered to run continuous, the diversity assumption that makes the demand tables work no longer holds, and you size close to the connected load.
How do you add an EV charger to an existing service?
You do not re-derive the whole building from nameplates. You find the real existing demand and check what the service has left. For an existing service, 220.87 lets you use the actual maximum demand measured over a continuous one-year period, or, if a year of data is not available, the maximum demand recorded over at least 30 days with a recording meter on the highest-loaded phase. That measured demand, taken at 125 percent, is your existing load.
The arithmetic from there is spare capacity. Subtract the existing load from the service rating and what remains is the headroom available for the new load. An EV charger is a continuous load under Article 625, so it enters at 125 percent of its rating. The new total, existing plus 125 percent of the charger, has to fit inside the service and the feeders without overloading either.
Worked quickly: a 200 A service meters a 96 A annual peak, so the existing load is 96 times 1.25, about 120 A. That leaves roughly 80 A of headroom. A 48 A charger at 125 percent is 60 A, which fits inside the 80. Push to a second 48 A charger and you are over, and the answer is a panel upgrade, load management that sheds the charger when the building peaks, or a smaller charger. This is where the metered method beats a paper recalc: the paper number is almost always more conservative than the building's real demand, and on a retrofit that conservatism is what forces an unnecessary service upgrade.
From calculated amps to the gear and the conductor
The calculation ends in VA. Divide by the system voltage to get amps: for single-phase, total VA over the voltage; for three-phase, total VA over the voltage times 1.732. That amperage is the minimum the service or feeder has to carry, and it drives four things at once: the conductor ampacity, the overcurrent device, the panel and gear rating, and the next standard size you round up to.
The conductor is sized so its ampacity, from the 310.16 table with the temperature and bundling corrections applied, meets or exceeds the calculated load. The overcurrent device is then selected, usually rounded up to the next standard rating where the calculated load does not land on one. And the conductor still has to clear two more checks the load calculation does not perform: it has to fit the raceway under the conduit-fill percentages, and on a long run it has to hold the voltage-drop target, which can drive the conductor larger than ampacity alone.
So the load calculation gives you the floor, not the final conductor. A feeder that passes the load calc can still fail the fill table or arrive at the far end low. Run the load calc for the size, then the ampacity table for the heat, then the fill check for the pipe, then the voltage-drop check for the distance. The largest conductor any of those four demands is the one you pull.
I = VAtotal / VI = VAtotal / (V × 1.732)Field example: a single-family dwelling, standard method
Take a 2000 sq ft all-electric house on a 120/240 V single-phase service, calculated under the 2023 NEC at 3 VA per square foot. Walk the standard method category by category and the service size falls out.
General lighting is 2000 times 3, which is 6000 VA. Add two small-appliance circuits and one laundry at 1500 VA each, another 4500 VA, for 10,500 VA in that block. Apply the general-lighting demand: the first 3000 VA at 100 percent and the remaining 7500 at 35 percent gives 3000 plus 2625, which is 5625 VA. The range, 12 kW or less, comes in at 8000 VA from Column C. The dryer is 5000 VA. Four fixed appliances, the water heater, dishwasher, disposal, and a fastened microwave, total 8100 VA at nameplate, and at the 75 percent demand for four or more that is 6075 VA. Heating beats air conditioning here, a 10 kW furnace against a 5 kW condenser, so carry 10,000 VA and drop the AC.
Add the demand figures: 5625 plus 8000 plus 5000 plus 6075 plus 10,000 is 34,700 VA. Divide by 240 V and the service load is about 145 A. The next standard size up is a 150 A service, though most builders install 200 A for the headroom and the future EV or heat-pump load. Run the same house through the optional method and it lands near 110 A, which is why the method you choose is itself a design decision. Under the 2026 edition at 2 VA per square foot, the general lighting drops by 2000 VA and the service comes out smaller again.
| Load | Connected VA | Demand factor | Demand VA |
|---|---|---|---|
| General lighting (2000 sq ft x 3 VA) | 6000 | see below | |
| Small-appliance (2) + laundry (1) | 4500 | see below | |
| Lighting block subtotal | 10,500 | 100% / 35% split | 5625 |
| Range (12 kW or less) | 12,000 | Column C | 8000 |
| Dryer | 5000 | 100% | 5000 |
| Fixed appliances (4) | 8100 | 75% | 6075 |
| Heat 10 kW vs AC 5 kW (larger) | 10,000 | 100% | 10,000 |
| Total demand | 34,700 VA | ||
| Service load (34,700 / 240 V) | ~145 A, use 150 A |
Spare capacity and sizing for future load
The calculated minimum is a floor, and building right to it is a decision you usually regret. Electrification keeps adding load to existing buildings: EV charging, heat pumps replacing gas, induction ranges, battery storage. A service sized to the bare calculation on the day it was installed is a service that needs an upgrade the first time one of those shows up.
Designing with headroom is not padding. It is reading where the load is going. On a house, the gap between a 150 A calculated service and a 200 A installed service is small money up front and a panel swap avoided later. On a commercial feeder, leaving room in the panel and the feeder for a known tenant fit-out or a future equipment line is cheaper than re-pulling and re-terminating once the walls are closed.
The honest version of headroom is sizing for load you can actually name, not a blanket round-up. If the EV charger and the heat pump are coming, calculate them in now, even as a future line, and size the service to the building you will have in five years. The metered existing-load method at 220.87 is the tool for proving how much room a service really has before you spend on an upgrade you may not need.
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.
What to document
A load calculation that cannot be reconstructed is a load calculation the inspector will not accept and the next engineer cannot trust. The record has to show not just the answer but every load, the VA that went in, the demand factor applied, and the demand VA that came out, so a reviewer can add the column themselves and land on your number.
Capture the occupancy and floor area, the code edition used, each load category with its connected VA and demand factor, the heating-versus-air-conditioning decision and which one you kept, the largest-motor treatment, the total demand VA, the system voltage, the resulting amps, the standard size selected, the method used if a dwelling, and who calculated and checked it. If you used the metered method on an existing service, attach the demand data and the recording period. The table below is the minimum a defensible calculation carries.
| Load type | Connected VA | Demand factor | Demand VA |
|---|---|---|---|
| General lighting + small-appliance + laundry | sum the block | 100% / 35% / 25% by tier | calculated |
| Household range | nameplate | Table 220.55 column | calculated |
| Electric dryer | 5000 or nameplate | 100% (single) | calculated |
| Fixed appliances | sum of nameplates | 75% if 4 or more | calculated |
| Heating or AC (larger only) | the larger load | 100% / per method | calculated |
| Largest motor + others | FLC values | 125% largest, 100% rest | calculated |
| Total | service amps at system V |
Common mistakes
- Adding every load at full nameplate with no demand factor, which oversizes the service and the cost.
- Counting both the heating and the air-conditioning load instead of only the larger noncoincident one.
- Applying the 125 percent largest-motor adder to every motor instead of just the single largest.
- Dropping the 125 percent on continuous loads where the adopted edition still requires it.
- Using the wrong VA per square foot for the edition, 3 versus the 2026 figure of 2 for dwellings.
- Putting a nameplate in where a demand table figure belonged, a 12 kW range counted at 12 instead of 8 kW.
- Sizing the neutral to the full load, or taking the 70 percent reduction on nonlinear wye load.
- Counting dwelling general-purpose receptacles a second time when the unit load already includes them.
- Recalculating an existing service from nameplates when the metered method would prove the real spare capacity.
Standards and references
The framework is NEC, NFPA 70, Article 220 through the 2023 edition, reorganized into Article 120 in Chapter 1 in the 2026 edition. General lighting unit loads and demand are at 220.12 and 220.42. Small-appliance and laundry loads are at 220.52, fixed appliances at 220.53, dryers at 220.54, ranges at Table 220.55, and commercial kitchen equipment at 220.56. Noncoincident heating versus air conditioning is 220.60, the neutral load is 220.61, show windows and signs are at 220.43, and nondwelling receptacles at 180 VA are in 220.14.
The optional dwelling method is 220.82, and the existing-load metered method is 220.87. Feeder and service minimums are at 215.2 and 230.42, motor feeder inclusion is at 430.24, EV supply equipment as a continuous load is in Article 625, and the ampacity that has to hold up alongside the calculated load comes from Table 310.16 with the corrections in 310.15.
These section numbers are accurate for the 2017 through 2023 editions and shifted in 2026 with the move to Article 120, and they vary further by local amendment. Confirm the article and section against the edition the jurisdiction has adopted before you put a citation on a submittal, and let the project specification and the AHJ govern where they are stricter than the code minimum.
Units and terms
Load calculations run in volt-amperes, and the same quantities show up under a few names across a drawing set, a spec, and an equipment sheet.
Volt-amperes (VA) is apparent power, voltage times current, and it is what sizes a conductor because the conductor carries the current regardless of power factor. Larger totals are written in kVA, thousands of VA. Connected load is the sum of all nameplates; demand load is what remains after the demand factors; the service or feeder is sized to the demand load. A demand factor is the fraction of a load counted toward that demand.
- VA / kVA
- Volt-amperes, the apparent power that sizes conductors and gear; kVA is thousands of VA
- Connected load
- The full sum of all load nameplates before any demand factor is applied
- Demand load
- The load remaining after demand factors, which sizes the service or feeder
- Demand factor
- The fraction of a connected load counted toward the service, set by the Article 220 tables
- Continuous load
- A load expected to run at maximum for 3 hours or more, generally taken at 125 percent
- Noncoincident loads
- Loads that cannot run at once, such as heat and AC, where only the larger is counted (220.60)
- Neutral load
- The maximum unbalanced load the neutral carries, sized under 220.61
- Optional method
- The 220.82 dwelling calculation, 100 percent of the first 10,000 VA and 40 percent of the rest
FAQ
What is an electrical load calculation?
An electrical load calculation totals a building's connected load, applies the NEC Article 220 demand factors, and returns the minimum service or feeder size in amps. It accounts for the fact that no building runs every load at once, so the diversified demand sizes the gear, not the sum of all nameplates.
What is a demand factor in a load calculation?
A demand factor is the fraction of a connected load you may count toward the service, because that load will not run at full output with everything else. The Article 220 tables set them by category from metered data. The dwelling general-lighting demand, for example, takes the first 3000 VA at 100 percent and the rest at 35 percent.
What is the optional method for a dwelling load calculation?
The optional method, NEC 220.82, lumps general lighting, small-appliance, laundry, and appliance nameplates together, then takes 100 percent of the first 10,000 VA and 40 percent of the remainder. It is allowed for a single 120/240 V service rated 100 A or more and usually yields a smaller service than the standard method.
How do you add an EV charger to an existing service?
Use NEC 220.87 to find the existing load from the metered annual peak, or a 30-day recording, taken at 125 percent. Subtract that from the service rating for the spare capacity. The EV charger, a continuous load under Article 625, enters at 125 percent of its rating and has to fit inside the headroom.
Do you count both heating and air conditioning in a load calculation?
No. You count only the larger of the two, because heating and air conditioning are noncoincident under NEC 220.60 and cannot run at the same time. Compare the heat load against the AC load and carry the larger into the total. Counting both is a common error that oversizes the service.
How much is a household range counted at in a load calculation?
A single household range rated 12 kW or less is counted at 8 kW, not 12, from Column C of NEC Table 220.55. The nameplate is the starting input, but the demand table sets the figure that lands in the total. Larger and multiple ranges follow the other columns and notes in the table.
How do you size the neutral in a service load calculation?
Size the neutral to the maximum unbalanced load under NEC 220.61, not the full load. A 70 percent demand factor is allowed on the neutral from ranges and dryers, and on the unbalanced portion above 200 A. Do not take the reduction on nonlinear wye load, where harmonic current adds on the neutral.
Why is my calculated service larger than the building seems to need?
Usually a demand factor was missed or a nameplate was used where a demand figure belonged. Check that you counted only the larger of heat or AC, applied the range and appliance demand tables, and added 125 percent to only the largest motor. On an existing building, the metered 220.87 method often proves a smaller real load.
What is the difference between connected load and demand load?
Connected load is the sum of every nameplate, the impossible all-at-once total. Demand load is what remains after the Article 220 demand factors are applied, reflecting how the building actually draws. The service and feeder are sized to the demand load, which is why adding raw nameplates oversizes the gear and the cost.
Did the 2026 NEC change load calculations?
Yes. The 2026 NEC moved Article 220 into a new Article 120 in Chapter 1 and cut the dwelling general-lighting unit load from 3 VA to 2 VA per square foot. Reports also indicate it revised the continuous-load treatment in the calculation. Most jurisdictions still enforce the 2020 or 2023 edition, so confirm the adopted code.
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.