Reading McNichols Bar Grating Load Tables: Span, Deflection, and Material in One Decision

A bar grating load table is not a product rating; it is a lookup in which four inputs — bearing bar depth, bearing bar spacing, clear span, and load type — produce one allowable number under one deflection limit. The distinction matters because most walkway and platform spans are governed by deflection rather than by stress: a bar that passes a stress check can still be rejected because the walkway flexes underfoot. So read the table from the limit backward. Declare the design load first, whether uniform, concentrated, or rolling; declare the deflection you will accept at that span; and only then look up which depth and spacing combination satisfies both. The number in the cell is a consequence of those declarations, not a property of the product, which is why one span can legitimately show two different allowable loads and why a value lifted from one row cannot be carried to another material.

A Load Table Is a Lookup Result, Not a Product Rating

Every bar grating load table answers one narrow question: at a given clear span, with a given bearing bar depth and spacing, what uniformly distributed load produces an acceptable amount of bending — and in most pedestrian and light industrial cases the ceiling is set by how much the grating is allowed to sag, not by the load at which it fails. That framing changes the order of work. A specifier who starts at the load column is reading the table backwards, because the figure printed in a cell was computed for one deflection criterion and one support condition. Before any cell means anything, two declarations have to be written down: what kind of load the grating actually carries, whether a uniform distributed load, a concentrated load from equipment feet, or a rolling load from a cart, and how much deflection the application tolerates. An electromechanical mezzanine and a maintenance walkway can share the same span and the same bar, yet need different numbers, because the acceptable bounce differs.

Table headers look like product identifiers, but each column encodes geometry: bearing bar depth, meaning the vertical depth of the load-carrying bar; bearing bar spacing; and the clear span between supports. Two gratings in the same nominal style can sit in different rows because their depth and spacing differ, and once depth changes, section modulus and moment of inertia change with it — the inertia steeply, as depth increases. Material enters the same calculation through the modulus of elasticity, which is why a comparison cannot be lifted from one column into another. A technical manual for aluminum products makes that point from the other direction: aluminum is about one third the density of steel, copper, and brass, resists common atmospheric and marine atmospheres, gains further corrosion and scratch resistance from anodizing, and in some alloys can match or exceed the strength of common construction steel. All of that is true of the material, and none of it is a load rating for a specific bar. The comparison stops being useful exactly where the geometry changes, because stiffness follows depth first and material second. The figure therefore has to come from the manufacturer's own table for the section being specified.

If the allowable figure is treated as a fixed product rating, the error propagates quietly. A specifier picks the row that looked strongest, writes that bar into the structural notes, and stops checking; then the span shifts by a few inches during a field revision, or the grating is substituted with a lighter material, and the number in the notes no longer describes what is being installed. Category-level documents actively encourage the mistake, because they are built to be used as filters. An aluminum product guide, for example, sorts material by thickness band — thin-gauge sheet runs from 0.2mm–6.0mm — and by alloy-family use case, which is the right way to narrow a candidate list and the wrong way to certify a member for a load. The consequence is hard to see: nothing fails on paper, because the paper still shows a number, just a number computed for a different combination of geometry, span, and deflection limit than the one in the field.

Why the Same Span Carries Two Different Allowable Loads

An engineer checking a platform against what looks like a single span often finds two or three allowable loads printed against it, and the first instinct is that the highest value must be the safest choice. It is not. The rows differ because the table is banded, not because the manufacturer is offering a range of quality. Material catalogues band products the same way: a supplier guide to aluminum sheet, plate, and coil splits the family by form and thickness band, and each band carries its own property set, so a comparison drawn across bands only means something when the same criterion is applied to both sides. Bar grating tables do exactly this with depth and spacing. The question of which value is the real one therefore has no single answer: the value belongs to the band, and the band is defined by geometry plus the deflection limit the table was built on.

The mechanism behind the double value is section geometry working with material stiffness. Bearing bar depth sets how far the material sits from the neutral axis: increase depth and the moment of inertia rises steeply, so the same load produces less sag and, within the stress limit, more allowable load. Spacing sets how much of the walking surface each bar carries, so tightening the pitch shares the load across more bars and raises the allowable value per unit area without changing depth. Material enters as the modulus of elasticity in the same deflection equation — a change of alloy shifts deflection in proportion, and a change of family shifts it by a much larger factor. This is why a guide to choosing an aluminum alloy and gauge begins with alloy family and temper rather than with a single strength number, separating 3003 for general workability, 5052 for corrosion resistance and marine work, and 6061 as a stronger structural alloy. Ranking by strength is not ranking by stiffness, and neither ranking is a rating for a grating bar; the table value is computed from the combination, so an alloy that looks stronger on a data sheet can still produce a lower allowable load at a given span.

So can a value computed for carbon steel simply be carried over to aluminum or stainless? No, and the reason goes past the modulus of elasticity. Any load table embeds a structural model: the bar acting as a simple beam over its supports, with the end connections transferring shear without adding stiffness. That assumption belongs to the fastening detail, not to the metal. Where grating attaches to a dissimilar metal, such as an aluminum unit landing on a galvanized steel frame, the joint itself has to be qualified, which is why published work exists on dissimilar spot welding of aluminum alloy to galvannealed steel at the interface level. A supplier category guide would list both materials as available, and that listing says nothing about what happens at the connection. The same caution applies to every cell in a load table: the number holds for the support and fastening condition it was derived under, and not outside it.

Where Load Tables Get Misread in the Field

The most expensive misreadings do not happen in the load column at all; they happen one column over, where a deflection-driven selection quietly turns into a corrosion decision. A documented case makes the arithmetic plain: a chemical processing facility installed 304 stainless cooling headers in a seawater heat exchange system, choosing them over 316 to save roughly 12,000 in material cost, and the pipe walls then suffered chloride attack — the failure mode that 316's molybdenum content exists to resist. Nothing in that decision was a load-table error, and no deflection check would have caught it; the stiffness margin was fine. A grating selected for a chloride-bearing environment inherits the same trap, because the purchase-order saving is real only if the exposure assumption holds for the life of the installation rather than for the first inspection. Whether the exposure is salt spray, brackish wash-down, or de-icing chloride, the material decision has to be made before the geometry is finalized, not after.

The dividing line between the two austenitic grades is chloride exposure rather than general severity, and the stainless material guide states it as a chemistry difference: 316 and 316L carry 2 to 3 percent molybdenum, which is what buys resistance to chloride-based attack, while 304 and 304L are intended for atmospheric conditions and low-level chemical exposure. Two consequences follow for anyone selecting grating for a platform. The price gap is structural rather than incidental, since 316 typically runs 30 to 40 percent above 304, so the choice is a real budget line that should be justified by a named exposure instead of a general preference for the better grade. The decision also arrives with documentation attached. The same sourcing guide references ASTM A240 / A480 and EN 10088-2 / EN 10028-7 behind the chemistry and property values, and describes those values as typical, for guidance only, which is the posture a specifier should take toward any grade comparison not yet confirmed against a mill certificate or the manufacturer's own table.

Comparisons hold only when the basis is stated. A machining guide for copper alloys sorts the family by the criterion that matters to the application — C11000 copper for electrical and thermal conductivity at 101% IACS, C36000 brass for high-speed machining and cost efficiency, C93200 bronze for anti-friction and wear resistance — and it warns that confusing them produces overheating connectors and saltwater-seized marine fittings. Each of those figures is quoted against a defined basis, which is the discipline a load table demands as well. A claim that a material is strong, or conducts well, with no statement of per unit area, per unit weight, or per unit cost, cannot be compared with a competing claim. The same is true of allowable load: one bar can look better or worse than another depending on whether the comparison is drawn per unit width, per unit weight, or per unit of deflection. If the basis is not written down, the comparison is not a decision input.

From Table Row to Spec Decision: An Ordered Rule

The rule that survives field revision has four steps, and the order carries the argument. First, declare the load type and the acceptable deflection; until both are in writing, no cell in the table means anything. Second, fix the span, meaning the clear distance between supports as built rather than the nominal bay dimension on the drawing. Third, choose the material and finish, naming the exposure explicitly if the installation sees chlorides, wash-down chemicals, or marine air. Fourth, verify the selected depth and spacing against the load table published for that material and finish, and record which table edition was used. Deflection before load is deliberate: on a walkway span the limit usually exists to control how the surface feels underfoot and how it behaves under a rolling cart, and that limit is normally reached before the stress in the bar becomes critical. Working in this order makes it structurally impossible for the largest number in a row to become the default, because the row must satisfy the deflection declaration first.

Two pieces of physics justify giving deflection priority. Bending deflection grows with the cube of the span while bending stress grows only with the square, so as spans lengthen the stiffness criterion arrives first; and the limit that matters to a person walking is a distance, measured in fractions of an inch, not a stress ratio. This is also where the weight objective enters the trade honestly. A shallower bar at a tighter spacing can carry the same uniform load as a deeper bar at a wider pitch while weighing less, but for the same depth-to-span ratio it will sag more. If light weight is the goal, whether a panel lifted into place by hand or a structure with a tight dead-load budget, the practical moves are to accept a tighter deflection declaration or to shorten the span, not to select a bar whose allowable load looks adequate while its stiffness does not. Weight and stiffness pull against each other, and the declaration made in step one is the only place that conflict can be settled without arguing about the table.

The rule fails in one predictable way: when its numbers come from a general material category rather than the manufacturer's own table. Category documents are organized to shorten a search — an aluminum product guide, for instance, groups its coil and plate by alloy family and gauge so a buyer can find candidates quickly — and their own framing says the values are typical, for guidance only, and to be confirmed against primary specifications. Brand-level allowable loads, span limits, and finish recommendations deserve the same treatment: confirm each one in the McNichols load table for the exact bearing bar depth, spacing, material, and span before it enters a project specification. The ordered rule is a decision procedure, not a source of numbers; the numbers come from the table that applies to the product actually being installed, and if that table has been revised since the last comparable project, the earlier figure no longer governs.

Read the table last, not first. The sequence that holds up under revision is to state the load type and the deflection you will accept, fix the clear span as built, name the material and finish with the exposure written beside it, and only then look up the depth and spacing that satisfy all three, confirming every figure against the McNichols load table for that exact combination. A row that passes a stress check but does not meet the declared deflection limit is not the conservative choice; it is the wrong choice with a comfortable margin.