Short answer: IEEE 1584-2018 models an enclosure up to 49 in [1245 mm] wide and high, and study software holds anything larger at that ceiling. The equations are normalized to a 20 x 20 x 20 in [508 mm] box, which puts a default run at the small, conservative end of the range. So a switchboard entered at 84 x 91.5 x 19.5 in comes back lower than the default, 16.1 against 21.4 cal/cm², because the calculation sees the capped opening of 49 x 49 in with 19.5 in of depth, not the full cabinet. Enter your measured dimensions whenever they sit inside the range, and record the effective size beside the physical one.
The 20 x 20 in box the equations are built around
The 2018 edition made enclosure size a calculated input, alongside electrode configuration, and it retired the 125 kVA screen that used to exempt small low-voltage transformers. The model was fitted to arc tests in metal boxes of recorded size, with the results normalized to a typical 20 x 20 x 20 in enclosure. The incident energy you read is that normalized value, scaled by a size factor built from height, width and depth.
The fit covers the boxes the tests covered, and the standard prints the recommended ranges with their ceiling:
| Input | Recommended range in IEEE 1584-2018 | Above the range |
|---|---|---|
| Height | 14 to 49 in [356 to 1245 mm] | Held at the ceiling |
| Width | 4 x bus gap, up to the ceiling | Held at the ceiling |
| Opening area | Up to 2401 in² | Saturates at the maximum opening |
| Depth | Enter the measured value | Uses the entry as given |
The ceiling is the whole point: 49 x 49 in is 2401 in², the largest opening the test data supports. Beyond it the equations would be extrapolating, so the model draws the line there and your software enforces it. The floor on width is four times the bus gap; for example, a 32 mm [1.26 in] gap calls for at least 5 in [127 mm] of width, which any real compartment clears.
Electrode configuration matters as much as size on switchgear. Model the arrangement the bus actually presents, then check the worst-case option; using VCB and checking the worst case for switchboard electrodes walks through that choice.
Default dimensions or measured ones
Leave the size at Default and the tool picks one of two paths. Some programs calculate against the 20 x 20 x 20 in normalized box, the small end of the range. Others load the typical dimensions attached to the equipment class, the same tables that carry the typical bus gaps. Both paths usually land higher than a measured box, because a small opening concentrates the arc toward the working position.
Read the input summary before you trust either number. The word "Default" tells you nothing until you see the dimensions printed beside it; when they read 20 x 20 in, your result is the normalized worst case. The switchboard worked below runs 21.4 cal/cm² on default and 16.1 cal/cm² on entered dimensions, a gap of about 25 % from geometry alone.
Measured dimensions inside the range are the better answer, because the study exists to predict the energy at the working distance. A box that is genuinely 30 x 40 x 18 in goes in as 30 x 40 x 18 in. Widening the box on the input screen to shave the label is the one move on the wrong side of this rule.
Two sections: one enclosure or two
The width the model wants belongs to the enclosure around the arcing location, not the lineup. Adjacent sections share one volume only where nothing interrupts it: an open transition leaves the combined internal width in play, while a metal barrier at the modeled bus means you enter the compartment alone. Two equal sections making an 84 in lineup give 42 in [1067 mm] each, and 42 in sits inside the model's range on its own.
Depth gets the same honesty. The model wants the front-to-back dimension the arc sees, not the cabinet outline, so door recesses and mounting pans can turn a 24 in outline into roughly 19.5 in [495 mm] of modeled depth. Depth carries no cap, and it moves the result in its own direction; why shallow enclosures calculate lower incident energy covers the mechanism.
Compartment-level geometry is the same thinking MCC buckets get in IEEE 1584, where the bucket, not the cabinet, sets the size.
Reading an 84 x 91.5 x 19.5 in switchboard entry
The inputs, from the study file and the standard's limits:
- Enclosure entered: 84 in wide, 91.5 in high, 19.5 in deep, two sections with no barrier at the modeled bus.
- Default run at that bus: 21.4 cal/cm².
- Entered-dimensions run: 16.1 cal/cm².
- Model ceiling: the cap on width and height; depth as entered.
The rule in words: the 2018 size factor is fitted to openings up to 49 in square, and study software clamps larger entries at that ceiling. The arithmetic written out:
- Width: 84 in entered, capped. Height: 91.5 in entered, capped.
- Effective opening: 49 x 49 = 2401 in², the standard's stated maximum.
- Saving: 21.4 - 16.1 = 5.3 cal/cm².
- Share of the default: 5.3 / 21.4 x 100 = 24.8 %, call it 25 %.
Prove the clamp on your file before the label prints. Run a third case at the ceiling in width and height with 19.5 in of depth and every other input unchanged; a match at 16.1 cal/cm² confirms the entered run used the ceiling, not the cabinet. A mismatch means the tool handles oversize entries its own way, so open its effective-dimensions report and take the number that report supports.
The decision: keep the entered case once the geometry question is settled. With no barrier between the sections, 84 in of shared width is the honest input and the clamp is the standard's doing. Both results sit inside NFPA 70E Category 3, which runs above 8 up to 25 cal/cm², so the PPE stays the same and only the label number moves. The boundary worth chasing is 25 cal/cm², where Category 4 starts on its way to 40 cal/cm². The label governs energized work; de-energize and lock out whenever the task allows it.
What to write in the study record
Six lines defend the label long after the study closes:
- Physical dimensions measured: 84 x 91.5 x 19.5 in, and where each number came from.
- Effective dimensions the run used: the capped opening with 19.5 in of depth.
- Equipment class, the bus gap it loaded, and the electrode configuration modeled.
- Software name and version.
- The barrier decision: sections combined or split, with the reason.
- The proof-run result at the ceiling.
A reviewer who sees 16.1 cal/cm² on a 91.5 in board will ask which dimensions reached the model. This record answers the question before it is asked, and it protects the number when your file is reopened years later.
Where the incident energy must come down, geometry is a real lever used honestly: shorter bus gaps, barriers that create smaller modeled compartments, and equipment rebuilt behind them. Enclosing equipment to reach Category 2 works that lever with 8 cal/cm² as the line to beat. A standard stainless box carries no arc rating of its own, and what an arc flash rated enclosure actually is is a separate question. When the redesign lands on a size no catalog carries, custom enclosures come built to your width, height and depth, delivered in 10 business days from your approval of the drawing. Physical sizing follows its own rules; the enclosure sizing guide covers them.
Enclosures for this job
Catalog sizes that suit the job this article answers, picked from its subject; open one for its full specifications.
| Enclosure | Material | Outside H x W x D / display | Rating |
|---|---|---|---|
| 12 × 10 × 6 in | 304 stainless | 12 x 10 x 6 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 20 × 16 × 8 in | 304 stainless | 20 x 16 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 24 × 20 × 8 in | 304 stainless | 24 x 20 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
Questions
What happens when you enter enclosure width above 49 inches?
The model stops at the ceiling of 49 in [1245 mm]. Width and height are recommended only to that value, and study tools hold larger entries there, so 84 in of width and 91.5 in of height both calculate at the cap. Confirm it on your file with a clamped run at the ceiling with 19.5 in of depth: if it returns the same incident energy to the decimal, the clamp is active.
Why did the default size calculate higher than the entered dimensions?
Defaults calculate against a small opening. The 2018 equations are normalized to a 20 x 20 x 20 in box, and some tools use exactly that; others load the typical size for the selected equipment class. A small opening concentrates the arc toward the working position, which is why the default run gave 21.4 cal/cm² against 16.1 cal/cm² entered, about 25 % apart.
Should two adjacent switchboard sections be entered as one enclosure?
Only when no barrier interrupts the enclosure at the arcing location. An open transition means the combined internal width is the box; a metal partition means you enter the compartment alone. Two equal sections in an 84 in lineup give 42 in each, and the smaller opening calculates higher, so the split case usually governs.
Does the drop from 21.4 to 16.1 cal/cm2 change the PPE?
No. NFPA 70E PPE Category 3 covers incident energies above 8 up to 25 cal/cm², and both values sit inside it, so the clothing stays the same and only the label number moves. The thresholds that change the work are 25 cal/cm², where Category 4 begins, and 40 cal/cm², where Category 4 ends.
Can enclosure size be used on purpose to lower incident energy?
Enter the box as built and take the number. Inside the model's range the size factor is physics: larger openings genuinely spread the arc, up to the ceiling. If the target is Category 2, redesign honestly with shorter bus gaps and barriers that create smaller modeled compartments, with 8 cal/cm² as the line to beat.
