Cable Tray Fill Ratio & Width: NEC 392.22 Worked Example

01What NEC 392.22 actually requires

Fail here and the tray either jams during the pull or fails the inspector's fill check at submittal stage. The section is short but it does not work the way most summaries describe it, because it splits into three separate tests by cable type and tray type — and only one of those three tests is a percentage.

Power and lighting cables: a table of areas, not a percentage

Where a ladder or ventilated trough tray carries multiconductor power or lighting cables, or any mixture of power, lighting, control and signal cables, 392.22(A)(1) sends you to one of three sub-rules depending on conductor size:

• All cables 4/0 AWG or larger — the sum of the cable outside diameters must not exceed the tray width, and the cables sit in a single layer. No area calculation at all.
• All cables smaller than 4/0 AWG — the sum of the cable cross-sectional areas must not exceed Column 1 of Table 392.22(A).
• A mixture of the two — the cables smaller than 4/0 AWG must fit within Column 2, which is Column 1 minus 30 mm multiplied by the sum of the diameters of the 4/0-and-larger cables. Those larger cables go in a single layer, and nothing may be laid on top of them.

Solid bottom trays repeat the same three-way split in 392.22(A)(3), with one material difference: the 4/0-and-larger diameter sum is capped at 90% of the tray width rather than 100%, and the area columns are the smaller Column 3 and Column 4.

Notice what is absent. There is no line in 392.22(A)(1) or (A)(3) that says "power cables are limited to 40% of the tray." The limit is a number of square millimetres read off a table against your tray width — 3,000 mm² at 100 mm wide, 6,000 mm² at 200 mm, 9,000 mm² at 300 mm — and it does not change with tray depth.

Control and signal only: where 50% and 40% come from

The two percentages everyone quotes do exist in 392.22, but they belong to a narrower case than most summaries imply. They apply to a tray whose contents are only multiconductor control and/or signal cables:

• 392.22(A)(2) — ladder or ventilated trough, usable inside depth 150 mm (6 in) or less: total cable area ≤ 50% of the interior cross-sectional area.
• 392.22(A)(4) — solid bottom, usable inside depth 150 mm or less: total cable area ≤ 40% of the interior cross-sectional area.

The 150 mm cap matters on deep trays. Where the usable inside depth exceeds 150 mm, both clauses direct you to compute the allowable interior area using a depth of 150 mm — extra depth buys you nothing for fill purposes when the tray carries only control and signal cable.

Single-conductor cables: a diameter sum for the small sizes

Single-conductor cables do not use the multiconductor area table. 392.22(B)(1) splits them by size:

• 1/0 through 4/0 AWG — the sum of the diameters of all single conductors must not exceed the tray width. That is the whole test. 392.22(D) adds that these conductors are installed in a single layer unless they are bound into circuit groups such as triplexed assemblies.
• 250 kcmil through 900 kcmil — the sum of the cross-sectional areas must fit Column 1 of Table 392.22(B)(1); mixed with 1000 kcmil and larger, the smaller conductors fit Column 2.
• 1000 kcmil and larger — sum of diameters must not exceed the tray width, single layer.

This is the single most common sizing error we see on submittals: a 4/0 single-conductor feeder run checked against a percentage of the tray area, when the governing clause is a diameter sum that will reject or accept the tray long before any area figure does.

Clause numbering per Article 392 as published in the 2020 edition of the NEC (adopted as the 2022 California Electrical Code). The structure of 392.22(A)(1)–(A)(6) and 392.22(B)(1)(a)–(d) is unchanged in the 2023 edition.

02The five inputs you need before sizing anything

Sizing goes wrong at the data-gathering stage far more often than at the arithmetic. Five inputs, in this order:

1. The tray type and usable inside dimensions. Ladder, ventilated trough, wire mesh, solid bottom or channel — it decides which clause you are in. "Usable" matters: side-rail intrusion and rung profile mean a 600 mm nominal ladder does not present 600 mm of cable bed.
2. The cable class for every cable in the run. Multiconductor or single-conductor, and the conductor size bracket (below 4/0 AWG, 4/0 and above, 250–900 kcmil, 1000 kcmil and above). One cable in the wrong bracket and you are reading the wrong column.
3. Finished outside diameter from the manufacturer's data sheet. Not the bare conductor diameter, and not the conduit trade size. Every Sorivo product page publishes the finished OD; for example the THHN building wire and YY/SY/CY flexible control cable pages.
4. Quantity per size. The fill calculation is a sum, so it needs counts, not counts-and-a-bit-for-spare. Add the spares as a separate line so the client can see them.
5. The support span and the load class you are designing to. Fill is only half of the tray decision — span and load come from the product standard, not from Article 392.

03Worked example 1: single-conductor 4/0 feeders

Six single-conductor 4/0 AWG THHN copper feeders, to be laid in a ladder tray, single layer. This is 392.22(B)(1)(d) territory — the diameter sum test — and the calculation takes two lines.

OD is a representative figure from the manufacturer's data sheet for finished 4/0 AWG THHN. Substitute the OD from the data sheet for the cable you are actually buying — the test and the clause do not change.

Two consequences worth carrying into the submittal. First, tray depth is irrelevant to this test — a 50 mm deep tray and a 150 mm deep tray pass identically at 150 mm width, because the clause constrains width only. Second, the spare width is not spare fill: 392.22(D) requires these conductors in a single layer, so the 96 mm of headroom at 200 mm width is headroom for future cables laid alongside, not above.

Contrast that with the same ampere rating in a different construction. Take six single-conductor 250 kcmil cables instead, which sit in the 250–900 kcmil bracket and therefore switch to the area table:

One size bracket down the list, and the governing quantity flipped from a diameter sum to an area. Reading the right row of Table 1 is the whole skill here.

04Worked example 2: mixing power and control

Now the case that trips most people: large multiconductor power cables sharing a tray with smaller control cables. One cable class is above the 4/0 AWG line, the rest below it, so 392.22(A)(1)(c) governs and the deduction term Sd appears.

OD values are representative manufacturer data-sheet figures for the constructions named. The 4/0-and-larger cables are not counted in the area column that matters — under (A)(1)(c) they are handled by the deduction term, not by an area sum.

392.22(A)(1)(c) asks one question: is the total area of the smaller cables within Column 2? And Column 2 is defined as Column 1 minus 30 mm times Sd, where Sd is the sum of the diameters of the 4/0-and-larger cables in the same tray.

Two things to take from this. The deduction term is not small: on a 200 mm tray it consumed 4,860 of the 6,000 mm² allowance, or 81% of the table value, purely because six large cables share the tray. And the arithmetic is indifferent to how you feel about the answer — the run that "looks fine" at 200 mm is short by 194 mm² of allowable area, and the fix is one width step, not a redesign.

If the same two cable groups sat in a solid bottom tray instead, you would repeat the calculation against Column 3 and Column 4: Column 3 at 225 mm is 5,100 mm², Column 4 is 5,100 − (25 × 162) = 1,050 mm², and the 1,334.4 mm² of control cable would fail at every width up to 300 mm. Solid bottom trays are materially stricter, and that is the only sense in which a percentage such as 40% describes power trays — loosely, and by coincidence of the numbers.

05The lookup tables, in the units you work in

Both tables are published in the NEC with a millimetre column and an inch column side by side. They are separate tabulations rather than conversions of one another — the inch figure for a 100 mm tray is 4.5 in², while the millimetre figure of 3,000 mm² converts to about 4.65 in². Work from the column that matches your drawings and do not convert between them. The tables below reproduce the millimetre columns, which is how SORIVO trays and most export project drawings are specified.

Sd = the sum, in millimetres, of the diameters of all 4/0 AWG and larger multiconductor cables sharing the tray. Column 1 applies under 392.22(A)(1)(b); Column 2 under (A)(1)(c); Column 3 under (A)(3)(b); Column 4 under (A)(3)(c). The allowable area does not vary with tray depth. Column 1 follows the tray width at roughly 30 mm² per millimetre of width, but read the table rather than the ratio — the published values are the enforceable ones.

Sd here is the sum, in millimetres, of the diameters of all 1000 kcmil and larger single-conductor cables in the same tray. Single conductors of 1/0–4/0 AWG do not appear in this table at all — they are governed by the diameter sum test in Table 1.

06Ampacity derating once the tray is sized

Fill is the geometric check. The electrical check is 392.80, and it is where a passing fill calculation can still produce a cable that runs hot. The factors depend on whether the cables are multiconductor or single-conductor, and on whether the tray is covered.

The four derating percentages are quoted from 392.80(A)(1)(b) and (A)(2)(a)(b) as published. Note that (A)(2)(a)–(b) apply to single-conductor installations following 392.22(B) — which is exactly the case in worked example 1, where six 4/0 THHN feeders fall in the 1/0–500 kcmil bracket and are therefore held to 65% of the free-air table in an uncovered tray unless they are laid single-layer with at least one cable diameter of spacing between conductors.

Ambient temperature and grouping corrections still apply on top, and the tray is not special in that respect — the same correction factors that apply in conduit and direct burial apply here. We work through the stacking of temperature, grouping and altitude corrections in the ampacity derating factors guide, and the installation-method comparison for tray versus conduit versus burial is in cable ampacity by installation method.

07NEMA VE 1 vs IEC 61537: which load number are you reading?

Article 392 tells you how full the tray may be. It says nothing about whether the tray can carry the load. That comes from the product standard, and the two families in common use derive their numbers differently — which is why a load figure quoted without its span is not a comparison at all.

NEMA VE 1 figures from the type-test description published in manufacturer engineering guides; IEC 61537 structure per the 2023 edition's published scope and change list. The two are not convertible: a NEMA class and an IEC declaration describe different tests of the same physical tray.

08What separates a 20-year tray from a 3-year tray

Fill and load get the attention; coating and traceability decide whether the installation is still sound in year fifteen. Steel gauge, galvanizing quality and published load data vary widely between suppliers, and the differences are visible in the documents rather than in the catalogue photograph.

How to read a tray load table

A usable load table carries three numbers, not one: the safe working load, the support span it is valid at, and the deflection criterion at that load. A figure quoted without the span cannot be compared with your route, and a figure quoted without a deflection criterion cannot be compared with a project that specifies one. When you hold all three constant between two suppliers, the comparison becomes arithmetic instead of impression.

Two checks catch most submittal problems. Confirm the span in the table matches your actual support spacing — load capacity falls quickly as span increases, so a table based on a 1.5 m span tells you nothing about a 3 m run. Then ask for the underlying test report rather than the summary figure; a supplier with certified tables and batch traceability produces both within a day.

Galvanizing: what the coating numbers mean

Both ASTM A123 and ISO 1461 set minimum coating thickness by steel thickness rather than as a single site-wide figure. ASTM A123 Table 1 grades for structural shapes step up through 45, 65, 75 and 100 µm as the steel gets thicker; ISO 1461 works in minimum mean thickness bands, in the region of 45–55 µm for steel under 3 mm and 70–85 µm at and above 3 mm. That is why a specification written as "galvanized to 65 µm" is incomplete: on a thin rail it may be generous, and on a 3 mm side rail it may be below the grade the standard assigns.

Hot-dip galvanizing after fabrication is what makes the coating work at the edges. Every sheared edge and punched hole on a pre-galvanized (coil-coated) tray is bare steel — typically around 18–20 µm of zinc per side on the coil, with nothing on the cut face — which is why thin pre-galvanized trays rust at the cut edges and splice plates first, not in the middle of the rail. Match the finish to the environment: dry conditioned interiors tolerate the thinner coatings, while outdoor, coastal, wastewater and chemical areas justify stainless 304/316 or aluminium despite the premium. When a specification simply says galvanized to ASTM A123 or ISO 1461, ask for the coating thickness grade per steel gauge on the shop drawing. It turns site acceptance into a measurement rather than an argument.

09Three things that change the sum

Covered and solid-cover trays

A solid cover traps heat and triggers 392.80(A)(1)(b): covered for more than 1.8 m, multiconductor cables are held to 95% of the ampacity table, and single-conductor cables generally to a lower figure still. A tray that passes the fill check at 40% of the available area can still fail the ampacity check once the cover is on. Size the cables, not just the tray, with the cover in the calculation.

Multi-layer arrangement

Ladder trays physically accept multiple layers, and Article 392 constrains the ones that matter: cables 4/0 AWG and larger — and in mixed trays, all of them — must sit in a single layer with nothing laid on top, and 1/0–4/0 AWG single conductors must be single-layer as well. Each extra layer of the cables that are permitted to stack reduces airflow and pushes you toward the derating figures above.

Firestop at rated penetrations

Firestop pillows, wrap strips and putty pads occupy tray volume at the crossing. Where the tray passes through a fire-rated wall or floor, size the penetration so the listed firestop system can be installed around the cables without compressing them against the rungs, and carry the extra width through the rest of the run rather than narrowing it immediately either side of the wall.

10What we publish, and what we do not

This page quotes clause-level fill and derating requirements from Article 392 as published, and gives two worked examples with the arithmetic visible so a reviewer can re-run them. It does not interpret your local amendment cycle, and it does not replace the engineer of record's check against the edition your jurisdiction has adopted — adoption of a new NEC edition varies by state and by project, and the clause structure is what has to be confirmed, not the article number.

On the tray side, we publish what we build and test: tray types, materials, finishes and load data, with certificates where a project requires them. We do not publish a fill calculator that silently mixes the three tests in Table 1, because the most common sizing failure we see is not arithmetic — it is applying the control-cable percentage to a power-cable tray.

Short version. Identify the cable class per cable, not per tray. Read the governing clause off Table 1. Compute the area or the diameter sum with finished outside diameters. Then apply 392.80 before you accept the ampacity column. Four steps, and the tray stops being the reason the inspection fails.