A hem is a bend that went all the way round
There is no separate hem formula. Take the bend allowance expression, which is the angle in radians times the quantity inside radius plus K-factor times thickness, and put 180 degrees in for the angle. The radians work out to pi, so the arc is pi times radius plus K times thickness, and the blank a hem consumes is that arc plus the flat return that folds back over the part.
The reason it feels different is that the return disappears. On an ordinary flange the leg is a dimension somebody inspects. On a hem the return lies against the part and the finished piece measures exactly what it measured before, so every bit of the allowance looks like waste from the outside. It is not: it is the edge that stops the panel cutting the person installing it.
| Edge on 0.0478 in material, K 0.44 | Radius | Return | Blank consumed per edge |
|---|---|---|---|
| Open hem, generous | 0.0500 | 0.2500 | 0.4732 in |
| Open hem, tight | 0.0250 | 0.2500 | 0.3946 in |
| Closed hem | 0.0200 | 0.2500 | 0.3789 in |
| Raw edge | - | - | 0 |
Note how little the radius moves the answer compared with the return length. On thin material the arc is a small part of the total, so the return is the number to argue about when metal is tight.
Open, closed, and why the difference is not cosmetic
An open hem leaves a deliberate gap between the return and the part, set by the radius. It forms easily, it does not work the outside fibre very hard, and it stays open, which means it also collects whatever is in the air. A closed or flat hem is squeezed until the return lies against the part. It is stiffer, it looks finished, and it takes far more force at the fold because the material at the outside of the bend has been stretched as far as that alloy will go.
That last point is where closed hems go wrong. The drawing says zero radius; the metal declines. There is always some radius left at the fold, and pressing toward zero cracks the outside of the hem. Aluminum in a hard temper does it readily, and so does anything already cold worked. If a closed hem is splitting, the answers are a slightly larger radius, a softer temper, or bending across the rolling grain instead of along it.
How short a return can be
Short returns are a tooling limit rather than a metal one. The fold is made in two operations on most equipment: an acute bend to somewhere around thirty degrees, then flattening. Both need enough flange left to hold, and below roughly four material thicknesses there is not much to hold. The exact minimum belongs to the machine and the tooling, and it is worth finding out before designing a part around a sixteenth inch return in eighteen gauge.
Long returns have their own problem. Anything much past ten thicknesses tends to bow along its length as it closes, because the fold is closing on a strip that is now stiff enough to fight back. Between four and ten thicknesses is where most hems live without anyone thinking about it.
Lock seams get measured, not looked up
The last option in each selector is deliberately empty. Pittsburgh locks, grooved seams, pocket locks and drive cleats all consume blank width, and the figures published for them vary between shops, between machines and between gauges, because what the seam takes depends on how the rolls are set. Spending ten minutes with a test strip is worth more than any table: cut a piece of the gauge you are running, form the seam on the machine that will make the job, and measure how much narrower it finished. That figure then holds until somebody adjusts the machine.
Applying those figures across a whole run of duct, rather than to one part, is the duct fabrication take-off calculator.
Where this fits with the other layout pages
Edge treatments are the last thing added to a blank and the first thing forgotten. The order that works is: develop the part with its bends through the bend allowance calculator or the multi-bend flat pattern calculator, add the hems and locks here, and only then nest the result with the sheet nesting calculator. A quarter inch of hem allowance on four edges is half an inch on each dimension, which is quite enough to turn three parts across a sheet into two. For a four sided tray, the walls come from the box and pan flat blank calculator and any hem on the top edge from here.
Questions people ask
How much material does a hem add to the blank?
The flat return length plus the arc of a 180 degree bend, which is pi times the quantity inside radius plus K-factor times thickness. For a quarter inch return with a 0.05 inch radius in 18 gauge at a K of 0.44, that is 0.25 plus 0.22, so about 0.47 inches per hemmed edge. Two hemmed edges on the same dimension nearly add an inch, which is why hems are worth putting into the nest before the sheet is bought rather than after.
What is the difference between an open hem and a closed hem?
The gap. An open hem keeps a deliberate space between the return and the part, set by the bend radius, and forms without working the outside of the fold especially hard. A closed or flat hem is flattened until the return lies against the part, which is stiffer and neater and asks a great deal more of the material at the fold. Closed hems crack on hard tempers and on aluminum, and the fix is usually a slightly more generous radius rather than more force.
What is the minimum hem return length?
It is a tooling question rather than a material one, and roughly four material thicknesses is the point below which most equipment starts to struggle. The fold is normally made as an acute bend followed by flattening, and both steps need enough flange left for the tooling to hold. Whether your machine can go shorter is worth establishing on scrap before a part is designed around it, because a return that cannot be formed turns into a deburring operation instead.
Do I use the same K-factor for a hem as for a 90 degree bend?
Not reliably. K-factor describes where the neutral axis ended up, and a hem works the material much harder than a gentle bend does, so the neutral axis shifts further toward the inside and K comes down. Using a 90 degree value on a hem tends to make the allowance slightly too large, which shows up as a part that finishes a few thousandths over. As with every other K-factor question the answer is a test fold measured with a caliper, not a table.
Why does the calculator not know what a Pittsburgh lock takes?
Because it varies enough that supplying a number would be worse than supplying none. The blank a lock seam consumes depends on the seam type, on the gauge, and on how the lock former is set up, and two shops running the same nominal seam can differ by an eighth of an inch. Over a duct job that is several sheets. Forming a sample on your own machine and measuring it takes minutes and gives a figure that is actually true for your work.