Why Do Punches Break When an NC Servo Feeder Misfeeds?

A Misfeed Is a Timing Event

Engineers tend to describe misfeeds in spatial terms: the strip ended up 0.4 mm short. The die sees it in temporal terms: the strip was still settling when the press came down. Those two descriptions lead to different fixes, and only the second one is useful.

Think about the sequence inside a single stroke on a progressive die running at 180 SPM. That is 333 milliseconds per stroke. Of that, the press spends perhaps 120 milliseconds with the ram below the point where material contact matters. The feeder has the rest of the cycle to accelerate the strip, move it the programmed length, decelerate it to a stop and hold position. On a well-tuned NC servo feeder, the move occupies 90 to 140 milliseconds and the remaining time is margin. When anything eats into that margin, the strip is still moving, or still vibrating, when the pilot pins enter.

Pilot pins are the tell. On a die with pilots, the pin pulls the strip into register before the cutting edges engage, which hides small positional errors but transfers them into pilot wear and strip distortion. On a die without pilots, there is no second chance: whatever position the strip held at contact is the position it gets cut at.

Four Ways a Misfeed Reaches the Die

Not every misfeed looks the same, and the remedy differs by mechanism. Here are the four I see most often, with the signature that identifies each.

Read the failure pattern before you touch a setting. Random failures point at a mechanical contact problem. Consistent failures point at a profile or thermal problem. Speed-dependent failures point at the cycle budget. A punch that breaks on the very first stroke after a tool change points at setup, not at the feeder.

What Happens in the Last 30 Milliseconds

The window that decides whether a punch survives is smaller than most people assume. On a servo feed, the move command ends with a position hold, but the strip does not stop dead. It has mass, it has strip tension on both sides, and it has the elastic behaviour of a long strip span between the feed rolls and the die.

Three things continue to move after the servo reports "in position." The first is the strip span itself, which behaves like a stretched spring and can oscillate at 15 to 40 Hz depending on span length and tension. The second is roll compliance: the polyurethane or composite coating on the feed rolls compresses under clamping force and relaxes over a few milliseconds. The third is the loop on the decoiler side, which changes the effective back tension the strip sees at the rolls.

That is why raising roll pressure fixes some misfeeds and makes others worse. More pressure reduces slip and increases roll compliance simultaneously. There is a window, and finding it is a matter of testing rather than guessing. On a line we commissioned last year running 0.9 mm CRS at 210 SPM, the optimum roll pressure band was 12 percent wide. Below it, slip; above it, visible roll marking and a slight increase in feed scatter as the coating compressed.

Why Short Feed Lengths Hide the Problem

Short feeds are the most dangerous configuration on a coil line, and they are the hardest to diagnose because they look healthy on the gauge.

When the programmed feed length is short, the servo spends less time accelerating and more time settled. Scatter on the finished part looks excellent. But the strip span between the feed rolls and the die is still long, and the loop dynamics have not changed. The result is a line that holds 0.05 mm on the part for months, then breaks a punch the week the shop runs a longer pitch or adds a station.

Two habits catch this before it costs a die. First, verify position at the die entry, not at the feeder exit, by mounting a dial indicator or a laser sensor on the strip just before the entry guide. The difference between feeder exit and die entry is the span error, and on a 900 mm span it can reach 0.25 mm at production speed even when the feeder reports perfect position. Second, run the longest pitch the tool will ever use during setup, not the pitch of the first job. Set the profile against the worst case and the short pitches take care of themselves.

Detection You Can Install This Week

Preventing die damage does not require a new machine. It requires knowing that a misfeed has occurred before the ram reaches the bottom. These measures pay for themselves the first time they fire.

• Enable the servo following-error window and set the trip threshold at twice your normal peak error, not at the drive default
• Add a proximity or laser sensor on the strip at die entry and wire it to the press stop circuit, not just to an alarm
• Log peak current on the feed servo and review the trend weekly; a rising peak current predicts roll slip before parts go short
• Fit a mechanical strip-end and buckle detector, which catches the failure that no electronic sensor can see
• Write the trip thresholds into the setup sheet so a new operator does not disable them to keep the press running

The last point matters more than the hardware. On three separate shop visits I have found the misfeed detector bypassed with a jumper because it was tripping too often. In every case the detector was right and the setup was wrong. A detection system that gets disabled is worse than none, because it creates the illusion of protection. Any line we ship carries CE marking and the interlocks are designed so that bypassing them requires a deliberate, logged action rather than a jumper wire.

Repairing the Die Is the Cheap Part

When a punch breaks, the instinct is to replace it and get running. That is the correct short-term move and the wrong long-term one, because the conditions that broke it are still in the line.

Before restarting, capture four things: the strip position error at the moment of failure if the control logged it, the feed length of the last fifty parts, the loop height on the decoiler side, and the roll pressure setting. Then compare those against the values recorded when the tool last ran clean. Nine times out of ten, one number has drifted, and finding it takes twenty minutes against a die repair that takes a shift.

There is also a structural lesson in the failure itself. Punches break in the same place because the same load path concentrates stress there. If you have replaced the same punch three times, the die is telling you that the clearance is wrong, the stripper pressure is uneven, or the misfeed protection is inadequate. Replacing the punch a fourth time is a decision to keep paying for the same lesson.