The team spends 45 minutes changing over the machine. The boss says implement SMED. The team buys quick-change clamps, paints the tools, and cuts the changeover to 35 minutes. The boss asks why it’s not 10. Because the internal steps are still internal.

Internal vs external

SMED (Single-Minute Exchange of Die) separates setup tasks into internal and external. Internal tasks are done while the machine is stopped: loosening bolts, removing the die, fitting the new one. External tasks are done while the machine is running: getting the new die, prepping tools, warming up materials.

The core idea is to convert as many internal steps to external as possible. If you can stage the next die, set the tools, and pre-heat the material while the last part is running, the stopped time drops to just the physical exchange.

The mistake everyone makes

Teams buy quick clamps and call it SMED. Quick clamps are external; they speed up the bolt torque. But the 20 minutes spent measuring and setting the die height is still internal. The clamps cut 10 minutes; the measurement still takes 20.

The real savings come from eliminating adjustments. If the die height is set by shims that never change, you don’t adjust it. If the stop positions are set by mechanical locators that don’t move, you don’t reset them. Every adjustment you can replace with a fixed locator is internal time you remove.

Standardize the die

If every die uses the same height, the same mounting pattern, and the same tooling, the changeover becomes a repeatable routine. The operator doesn’t think; they swap. Custom dies that need unique mounting each time kill SMED. Spend the money on standardized die holders, not on quick clamps.

Prefabricated die carts help. The next die sits on a cart outside the machine, fully loaded with the right tools and inserts. The operator rolls the cart over, swaps the die, and sends the old one back. No walking to the tool room, no searching for the right hex key.

The first-part run

Most of the changeover time isn’t the physical swap. It’s the first part: test, measure, adjust, run another, adjust again. This is internal time that never gets measured as changeover. The machine is stopped while the operator tunes the process.

Eliminate this by standardizing the process parameters. If the recipe is stored in the machine controller and the die is repeatable, the first part should be good. If it isn’t, the problem is the die or the process, not the operator. SMED measures the time from last good part to first good part, not just the bolt swap.

Measure before you optimize

Film a changeover. Watch the video and time each step. You’ll find that 60 percent of the time is spent walking, searching for tools, and waiting. These are external tasks that should happen while the machine runs. The actual die swap might be 10 minutes. The other 35 is logistics.

Once you’ve timed it, attack the biggest chunk first. If the operator walks to the tool room for 15 minutes, stage the tools at the machine. If they spend 10 minutes finding the right insert, label the trays. Don’t start with quick clamps; start with the walk.

Bottom line

SMED isn’t about faster bolts. It’s about converting internal steps to external, and eliminating adjustments. Quick clamps are the last 10 percent. The first 90 percent is staging, standardization, and process parameters that don’t need tuning. Measure the whole changeover, not just the wrench time.