Also searched as aluminium profile standard sizes, 40x40 aluminium extrusion dimensions, T-slot profile size chart, aluminium profile series chart, standard aluminium extrusion sizes mm, aluminium profile catalogue dimensions.

What the number in the name actually tells you

The naming convention is the easiest thing to get right and the most commonly misread. Take 4590. It is not a model number, a strength class or a load rating. It is a cross-section measurement: 45 mm on one side, 90 mm on the other. The same rule gives you 2020, 2040, 3030, 4040, 4545, 6060, 8080 and the deep beam sections such as 30120 and 50100.

Two consequences follow immediately, and both catch buyers out.

First, the name is silent about the wall. The overwhelming majority of aluminium T-slot extrusions are sold under a name that describes only the outer envelope. A 40 × 40 heavy wall and a 40 × 40 light wall are both sold as “4040”. They take the same brackets, the same end caps and the same slot nuts, they look the same in a photograph, and they behave differently by a factor of two under load. If your drawing says 4040 and nothing else, the supplier will ship whatever wall their stock happens to be.

Second, the name is silent about the slot. The slot mouth is the opening at the face of the profile, and it is the feature every accessory engages with. It is set by the series, not by the envelope, and the series is not always what the name suggests. A 40 series profile has an 8 mm mouth on most European grids; a 45 series has a 10 mm mouth. Nothing crosses over — a 40 series T-nut will not seat in a 45 series slot, and the outer dimensions being 5 mm apart makes that mistake easy to make and expensive to discover.

The series matrix

Below is the matrix we work to. Every figure is a nominal value for a standard-wall section of that series in 6063-T5, and the inertia column is the second moment of area about the strong axis, which is the number that governs deflection. Treat the inertia and weight figures as typical values for the series rather than as a promise for any single item — wall thickness moves them, which is the subject of the next section.

Read the table across the middle and the structure of the market becomes clear. Between the 30 series and the 45 series the inertia goes up by a factor of roughly five while the weight only goes up by a factor of two and a half. That gap is where most of the interesting design decisions live, and it is why “just go one size up” is often better value than “order the heavy wall”.

The deep beam rows deserve a note. A 30 × 120 is not a big frame member; it is a shallow beam, and its 115 cm⁴ about the strong axis comes from depth rather than from material. It is the right answer when a worktop or a conveyor rail has to span 2 m without sagging, and the wrong answer when the load arrives from the side, where its weak-axis inertia is closer to 12 cm⁴.

One envelope, three different members

This is the part of the matrix that costs people the most money, because the difference is invisible. Three sections below are all sold as 40 × 40 four-slot. They have the same outside dimensions to the millimetre, the same slot pitch and the same slot class. Only the wall changes.

Two things in that table are worth pausing on.

The stiffness gain is real but sub-linear. Going from light to heavy wall adds 81 per cent more aluminium (1.30 to 2.35 kg/m) and buys 103 per cent more inertia. In a purely stiffness-driven design that is a good trade. But it is a worse trade than it looks, because the same material spent on a larger envelope buys far more. Moving from a 40 × 40 standard to a 45 × 45 standard adds 46 per cent more weight and buys 68 per cent more inertia — and it also brings a larger slot class, a bigger centre bore and a stiffer joint. Wall thickness is the upgrade you take when the envelope is fixed by an existing frame or a panel cut-out. It is not the upgrade you take by default.

The bore is a hidden specification. Look at the centre bore column. The light-wall 40 × 40 has a 6.8 mm bore that taps M8; the standard and heavy have an 8.2 mm bore that taps M14. If your design puts a levelling foot or a castor into the end of the profile, the wall grade decides which thread you can have — and the light wall cannot be upgraded by asking the machine shop nicely, because there is not enough metal around the bore.

Where “standard” stops, and the seven lines a PO needs

Suppliers publish catalogues of standard sections, and the word “standard” in that context means one thing only: the die already exists, so there is no tooling charge. It does not mean the section is standardised across the industry, and it does not mean two suppliers’ 4040 will interchange.

The practical fix is to stop treating the name as a specification and to write the specification out. Seven lines are enough for a supplier to quote unambiguously and for you to reject a delivery that does not match.

1. Envelope. 40 × 40 mm nominal, plus the drawing. The drawing is what settles corner radius, slot shape and bore position, and it is the only defence against a supplier whose “4040” is a different die.
2. Series and slot class. State the slot mouth, not just the series: “8.0 mm mouth, M6/M8 class”. This is the line that prevents the cross-series hardware mistake.
3. Wall thickness, with a tolerance. Nominal wall plus the tolerance you will accept. “2.0 mm nominal, EN 755-9 group I” is a specification; “standard wall” is a hope.
4. Alloy and temper. 6063-T5 for general framing. Write it out even when you expect it, because a substitution to a softer temper is undetectable at goods-in without a tensile test.
5. Tolerance class. EN 755-9 for general engineering, or the tighter class your assembly actually needs. Do not leave it blank and do not ask for the tightest class “just to be safe” — you will pay for capability you cannot use.
6. Cut length and end condition. Length, length tolerance, squareness of cut ends, and whether the ends are deburred or tapped.
7. Inspection evidence. Which certificate you want — mill certificate for the material, dimensional report for the cut length, or both — and which dimensions are critical to function.

Line three and line five are the ones that get left out, and they are the two that decide whether the profiles you receive will assemble. The next two sections give you the numbers to put on them.

Tolerance follows the circumscribing circle, not the side length

Almost every buyer who looks at an extrusion tolerance table expects it to be indexed by the profile size. It is not. EN 755-9 indexes the cross-sectional tolerance by two things: the dimension being toleranced, and the diameter of the smallest circle that will contain the whole cross-section — the circumscribing circle, usually written CD.

That second index is the one people miss. For a square section the circumscribing circle is the diagonal, not the side. A 40 × 40 has a 40 mm side but a 56.6 mm circumscribing circle. A 60 × 60 has an 84.9 mm circle. A 90 × 90 crosses the 100 mm boundary and lands in the next tolerance bucket, even though no single dimension on the section is anywhere near 100 mm.

Here are the actual figures for alloy group I, which is where 6063, 6060, 6005 and 6063A sit. Group II — 6061 and 6082 — gets roughly 40 to 60 per cent more tolerance for the same geometry, which is a useful reminder that the alloy you specify changes what you can hold.

Apply that to the sections in the matrix and you get the tolerance you should actually expect on the width, without asking anyone.

Two practical readings come out of that table. The first is that a 30, 40 and 45 series section all carry the same ±0.50 mm on their main dimension — the envelope size barely changes the tolerance until the profile gets genuinely large. The second is that a supplier quoting “±0.1 mm on the profile” is not quoting an extrusion tolerance at all; they are quoting a machining tolerance, and if they mean it they are going to cut or mill the section to get there.

Choosing by span, not by size

The single most useful thing you can do with the matrix is to stop asking “what size do I need” and start asking “what span do I have to cross”. For a machine frame the governing criterion is almost never material failure; it is how far the member bends. A simply supported member with a load at mid-span is limited by:

Run that across the series and you get a table you can use directly in a design review. The figures below are kilograms at mid-span.

The shape of that table is the whole lesson. Doubling the span costs you a factor of four in allowable load, because deflection goes with the cube of span. Going from a 40 × 40 light to a 40 × 40 heavy, at the same span, buys you roughly a factor of two. Going from a 40 × 40 light to a 45 × 45 heavy buys you about 2.2 — for less than twice the weight.

Which is why the useful first question in a frame design is not “40 or 45” but “can I add a support at the middle”. Halving an unsupported span multiplies the allowable load by four and costs one extra member and two brackets. No section change comes close to that return.

Frequently asked questions