The lever that matters most

Change the span before you change the profile

Almost every profile-selection conversation starts with the section and ends with a support being added anyway. Reverse that order and the whole exercise gets cheaper. The reason is the exponent: deflection is proportional to the cube of the span, and only to the first power of the load. Halving a span cuts the deflection of a given load to one eighth. Halving the load cuts it to one half.

In practice you do not get the full factor of eight, because the allowable deflection also scales with the span — a shorter beam is allowed to sag less in absolute terms. Working to L/300 throughout, halving the span gives you back a factor of about four in allowable load. That is still the largest single gain available anywhere in this decision.

Two consequences follow, and they are worth stating plainly because they contradict how most frames are specified:

• A mid-support is usually cheaper than a bigger section. One extra upright in a 40 series frame adds roughly 1.5 kg of profile and four brackets. Stepping the whole beam from 40 × 40 to 60 × 60 adds 2.6 kg per metre and forces every nut, bracket and panel fixing in that joint to change with it.
• Beam size should be the last thing you scale, not the first. If the frame is still too springy after the spans are as short as the layout allows, you have a genuine section problem and the numbers below will tell you which section solves it.

Load case

Where the load sits changes the answer by five times

Suppliers publish one load figure per profile, and it is always the mid-span central point load, because that is the worst case and the easiest to calculate. Real frames rarely load the middle. A drive unit sits at one end of a rail. A cylinder pushes at a third of the way along. Two brackets carry a panel between them. Each of those is a different beam problem, and the differences are large enough to change which section you buy.

For a simply supported span, a point load of F at distance a from one support deflects the beam at the load point by:

Put the load at the quarter point instead of the middle and the deflection falls to nine sixteenths of the mid-span value — a 44 per cent saving for free. Put it at the eighth point and it falls to under a fifth. This is why the honest answer to “what does a 4040 carry over one metre” is a range, not a number.

The budget

Set the deflection budget before you calculate anything

Deflection only becomes a pass or fail once you have decided how much sag the application tolerates. That is an application decision, not an engineering one, and it is the step most selection guides skip — which is why they end up saying “choose a larger profile” without ever saying how much larger.

The usual convention is a fraction of the span, because a short beam should be allowed less absolute sag than a long one. As a starting point:

Two notes on using this table. First, the limit applies to the worst-loaded member, not to the frame as a whole. A workbench can be a general L/300 structure while the one rail that carries a guide is held to L/500. Second, if the frame will be levelled on a floor that is itself flat only to a few millimetres, holding a beam to 0.2 mm of calculated sag is measuring the wrong thing.

Strength or stiffness

Which limit binds depends on the span

Every aluminium framing guide talks about deflection. Almost none of them mentions the other limit, and on short spans the other limit is the one that governs. An aluminium beam can fail in two distinct ways: it can bend too far (a stiffness limit, set by the geometry of the section) or it can yield at the extreme fibre (a strength limit, set by the alloy). They cross over at a specific span, and on a 40 × 40 mm standard section that crossover is almost exactly one metre.

The strength check is short. Bending stress is the moment divided by the section modulus, and the section modulus is the second moment of area divided by half the section depth:

Run both checks across the span range and the picture becomes clear:

Read the table from the left and two things happen at once. Below a metre, a 4040 does not fail because it bends too much — it fails because the aluminium at the top and bottom of the section reaches its yield point. Above a metre, the yield check becomes irrelevant and the whole design is a stiffness exercise. The crossover for this section sits at about 1,020 mm.

That single number explains a lot of field experience. It explains why short, heavily loaded brackets sometimes deform while long, lightly loaded rails simply sag. It explains why the answer to “how much does a 4040 hold” is not one number but a curve. And it explains why raising the alloy temper helps a stubby bracket but does nothing at all for a four-metre rail.

The method

The five-step selection sequence

This is the whole method, in the order that produces a decision. It takes about ten minutes with a calculator and it replaces the catalogue-browsing loop that usually consumes an afternoon.

The ladder

Turning your required I into a part number

Step four gives you a single number: the second moment of area your member needs, in cm⁴. The table below converts that number into a section, which is the last step of the method and the only step most guides describe. Read down the left column until you find your required I, then read across. Every section named here is drawn and published in our aluminium sections range, and the four-slot 4040 that most machine frames land on is here.

Three things this presentation makes obvious that a load table does not. First, the table is a lookup, not a comparison — you do not need to know what every section carries, only which one clears your number. Second, the rectangular column is often better than the square one at the same level, which is why the shape of a member is worth deciding before the size. Third, the last row exists at all: there is a point past which no single section is the right answer, and the calculation is telling you so.

Worked example

A 2,000 mm frame, sized two ways

Here is the whole method applied to a real problem. A control cabinet weighing 90 kg is to be mounted on a cross rail between two uprights 2,000 mm apart. The rail must not visibly flex when the door is opened, so the client has asked for L/300.

The two options are not close. Option A uses a 2,000 mm length of 60 × 60 heavy wall at 6.4 kg/m, plus brackets sized for a 60 series slot. Option B uses a 2,000 mm length of 40 × 40 standard at 1.7 kg/m, plus a 900 mm upright at 1.7 kg/m, plus one extra set of brackets. In kilograms of aluminium that is 12.8 kg against 4.9 kg — and the second option also keeps the entire frame in one accessory system, one nut part number and one torque figure.

The only question that decides between them is whether the middle of the frame is allowed to have an upright in it. If the cabinet needs a clear 2,000 mm opening for access, Option A is correct and you should buy the 60 × 60 heavy wall with confidence, having proved it rather than guessed it. If the upright can go in, Option B is cheaper, lighter and stiffer in every direction except the one you did not need — and it keeps the frame on the 40 × 40 four-slot section and the accessory range that goes with it.

When size is not the problem

Four symptoms that a bigger profile will not fix

Frames are often re-specified to a larger section when the real fault lies elsewhere, and the larger section then hides the fault rather than removing it. These four symptoms each have a different cause, and each one has a cheaper fix than buying aluminium.

The frame racks sideways

If the frame moves into a parallelogram when you push it from the side, you have a bracing problem. Every member in the frame is fine in bending; what is missing is a triangular load path. A diagonal, a bolted panel, or a properly fixed rigid top turns the rectangle into two triangles and removes the movement entirely. Enlarging the beams does almost nothing, because the load is not travelling down them.

The corner rotates under load

If the movement appears at a joint rather than along a member, the joint is the mechanism. Two bolts in a corner bracket resist rotation by friction and bolt bearing, and the section behind them only helps in proportion to how far apart the bolts are. A flat joining plate across the joint, or a gusset that puts the fixing further from the corner, does more than any increase in profile size.

The whole frame vibrates

Vibration is a natural-frequency problem, not a strength or stiffness problem in the usual sense. Making every member heavier lowers the natural frequency, which can move the structure closer to the excitation rather than further from it. The fixes are to stiffen the specific member that is moving, to change the mass distribution, or to damp the structure — not to upsize everything uniformly.

The load is carried by a panel, not by the frame

A frame with a thin bolted-on panel behaves as a frame plus a diaphragm. If the panel is fixed with too few fasteners it contributes nothing and the frame carries everything; if it is fixed properly it can contribute more than the members do. Getting the panel fixing right is often the difference between a frame that feels solid and one that does not, with no change of profile at all.

Questions

Frequently asked questions