Reading a Sheet-Metal Fault

A fault board is the most useful thing you will make in this batch, and it takes an afternoon. Make each of these deliberately, on scrap, and label every one in permanent marker with what you did to cause it. Keep them. 1. **Wrinkled flange** — draw a cup with the blank holder barely finger tight. 2. **Split wall** — draw an identical cup with the blank holder bolted down hard. 3. **Cracked bend** — bend a coupon with the sheared burr on the *outside* of the bend, along the grain, at a radius below one thickness. 4. **Good bend, same radius** — the same coupon deburred and bent across the grain. This is the control and it is the one that makes the board convincing. 5. **Orange peel** — stretch a coarse-grained coupon hard over a dome. 6. **Eared rim** — draw a cup right through and leave the rim untrimmed. 7. **Galled wall** — draw a cup dry, with no lubricant at all, on a die you do not mind marking. 8. **Oil-canning panel** — a flat 150 mm square of thin sheet with no bead in it. Press the middle and listen. 9. **Unrolled seam** — the plain lock seam from rung 5, peeled. Nine samples, nine labels. Then put them somewhere you will see them, because the value of the board is recognition speed: the point is to look at a bad part six months from now and know within a second which of these it is.

Two failures look similar at arm's length and mean opposite things. Under the microscope they are unmistakable, and telling them apart is the single most valuable skill in this rung. **A split** is a *ductile* failure in metal that was stretched too far. Look for: - **local thinning** right at the failure — measure it with the micrometer against the metal 10 mm away, and it will be markedly thinner; - a **neck**, a visible narrowing that preceded the opening; - a rough, fibrous fracture face; - often an **orange-peel** surface nearby, because that metal was heavily strained. **A tear from an edge** is a crack that started at the blank's edge and ran. Look for: - **no thinning** — the metal at the failure is the same thickness as its neighbours; - the crack starting exactly at the boundary, usually at a burr or in rung 1's fracture zone; - a flat, straighter fracture face. The distinction matters because the fixes are opposite. A split says *the metal was asked for too much* — reduce the demand: less blank holder force, more lubricant, a softer temper, a bigger radius. A tear says *the metal was fine, the edge was not* — fix the edge, and changing the press settings will do nothing at all. Go through your fault board with the micrometer and sort every failure into one of the two before you read step 4's table. Getting this wrong is why people spend a day adjusting a press that was never the problem.

The graph on this page is punch force against punch travel for a draw. If you can log force, you can see a fault before the part exists. The simplest rig that works: a load cell or a bathroom-scale strain gauge between the ram and the punch, read by the Arduino or a meter, with travel read off a dial indicator or just marked every 2 mm on the ram and read by eye. Two people, one calling travel and one calling force, is a perfectly good instrument for this. Draw four cups and record force every 2 mm: - **normal**, properly held and lubricated; - **blank holder loose**; - **blank holder tight**; - **dry, no lubricant**. Plot all four on one sheet of graph paper. The shapes in step 4 are what you are looking for, and they are distinguishable by eye long before they are distinguishable in numbers: - a **rounded peak that falls away** is a healthy draw; - a **cliff** is a split; - a **saw-tooth** on the rise is wrinkles being dragged into the die gap one at a time; - a **peak much higher than the others** is friction — too tight, or no lubricant. The most useful thing on the plot is *where* the peak is. The force peaks when the flange is small and still being pulled, and a draw that is going to split splits just before the peak. If you are watching a gauge, that is your warning.

**Attribution, stated honestly.** No patent is claimed for this rung and none exists to claim: reading a fault is a craft skill that every press shop develops and writes down in its own words. What *was* invented, and is worth naming, is the framework that makes those words comparable — the **forming limit diagram**, developed by Stuart Keeler and Walter Backofen and by Gorton Goodwin in the mid-1960s, which puts the major and minor strains from rung 6's circle grid on one chart and draws a line between what survives and what splits. Every modern press shop's diagnosis sits on that chart. **Why the faults are so few.** Eleven named faults in step 4's table, and almost all of them reduce to three causes: asked to stretch too far, not held enough, or a bad starting edge. That is not a coincidence — sheet metal has only a few degrees of freedom. It can thin, it can buckle, or it can crack from a flaw, and every named fault is one of those three wearing a location. **Why this rung closes the batch.** The nine rungs here are one argument: a flat sheet has no shape and no stiffness, and every operation in the discipline is a way of giving it one or the other. Cut it (rung 1), fold it (2), corrugate it (3), roll it into a joint (4, 5), measure what it will take (6), press it against rubber (7), draw a fastener out of it (8). This rung is what happens when one of those is asked for slightly more than it can give, and it is where a maker actually spends their time. **Honest limits.** A fault board is a recognition aid, not a diagnosis. Two different causes can produce faults that look identical — a wall split from an over-tight blank holder and one from a die radius that is too sharp look the same on the part, and the only way to separate them is the micrometer test in step 2 plus a look at the tooling. The process window in step 4 assumes a round cup and a uniform blank holder; a real part is neither, and its window is different at every point round the die. And none of this substitutes for keeping the tooling clean: a large fraction of the faults blamed on material are metal picked up on a die radius.