Line Changeover: Preparation, Placement and Process Control

Line changeover is the interval between the last good board of one product and the first good board of the next, and it is one of the largest hidden costs on an SMT line. It is not only the machine time; it is the paste that has to be replaced, the boards that are printed and inspected before the line stabilises and the attention of everyone involved. Those costs rarely appear in a production report, which is why the interval is often accepted as a fixed property of the line rather than managed as a process.

The way to shorten a changeover is to move work off the line. Every task that can be done while the previous product is still running, from kitting to feeder setup to programming, removes time from the interval where the line is producing nothing.

What Changeover Really Costs

The visible cost is the stopped line, and the invisible cost is the ramp. A changeover that ends when the first board is printed still leaves the line in a period where paste volume, placement and profile have not yet settled, and the boards produced in that period carry higher defect risk.

Counting both gives a more useful figure. A line that spends twenty minutes changing over and forty minutes ramping to full yield has a changeover of an hour, and shortening the ramp through preparation usually delivers more than shaving minutes off the mechanical setup. Preparation also spreads the work across people who would otherwise be idle, which is why the saving is larger than the arithmetic suggests.

Preparation: Kitting and Stencils

Kitting means gathering everything the product needs before the line stops: reels, trays, the correct stencil, the assembly drawing, the profile and the paste lot. A kit that is complete and verified removes searching from the changeover window, where it costs the most. A short checklist attached to the kit, signed off by the person who built it, converts the task from memory to procedure.

The stencil is the item most often missing or wrong. It should be checked for part number, revision, condition and tension as part of kitting, so that a stencil that has lost tension is identified while the previous product is still running rather than after the line has stopped.

<img src="https://www.gopcba.com/wp-content/uploads/2026/08/low-volume-pcb-aoi-xray-testing-early-defect-detection.jpg.webp" alt="First article board checked after an SMT line changeover and setup” />

Feeder Setup Offline

Preparing feeders away from the machine is the single largest saving available on most lines. Reels can be loaded, spliced and verified on a bench cart, and the correct pick positions can be taught and documented before the cart reaches the machine.

The rules that make offline setup reliable are the same ones that make feeder work reliable at any time, and they are covered in the usual feeder setup checks. A cart built to a verified list can be exchanged in minutes, while feeders built at the machine take far longer and introduce errors under time pressure.

Program, Recipe and Paste Changes

Placement programs and machine recipes should be loaded and verified before the changeover begins, including component databases, nozzle assignments and vision settings. A program that has to be edited during the changeover is a program that will be edited in a hurry.

Paste changes need their own decision. Where the new product uses the same paste, the existing material can often continue; where it uses a different paste, the stencil and the printer have to be cleaned and the new paste brought to temperature. That brings the thaw timing into the changeover plan, which is easy to forget. A jar taken from the refrigerator at the moment the line stops will not be ready when the printer is.

First Article and the Decision to Run

The first article is the gate. Printing, placement and reflow are checked against the drawing, and the inspection should include the smallest aperture, the finest pitch part and the joint that is hardest to heat. A first article that passes on a convenient area proves very little.

Whoever signs it off has to be able to stop the run. Where the first article is treated as a formality, the line continues with a defect that will be found later in volume, and the cost of the changeover is repaid many times over in rework. The first article should also be repeated after any adjustment, because a setup that has been touched is no longer the setup that was verified.

Cleaning and Residue Between Products

Cleaning is part of the changeover, not a separate task. Paste left on the stencil and the printer from the previous product is the wrong paste for the new one, and residue on the conveyor, in the nozzles and on the reflow fixtures carries over into the next build.

The stencil deserves particular attention because cleaning chemistry and cleaning method both affect its tension and its mesh adhesive. A fast clean that damages the tool shortens its life and moves the cost from the changeover to the stencil budget.

Line Balance and Batch Strategy

How work is batched affects how often the line changes over. Grouping products with similar paste, stencil thickness and component families reduces the number of changes, and sequencing them so that the largest batch runs between the small ones keeps the number of setups down.

The machine configuration matters too. A line with more placement capacity can absorb a wider component range without feeder changes, which reduces the number of times the machine has to be rebuilt.

Measuring Changeover Time

Measurement should start when the last good board leaves the line and end when the first article is accepted, not when the machine starts running again. Recording the two figures separately, mechanical setup and ramp to acceptance, shows which one is actually the problem.

The record should also capture what went wrong. A short note on missing parts, program edits or a stencil that had to be changed gives the next changeover a chance to be better, and it turns an operational annoyance into a list of fixable items. Reviews of that list every month usually identify two or three recurring causes that can be removed with a small procedural change.

Common Causes of a Long Changeover

The recurring causes are predictable: an incomplete kit, a feeder cart that was not prepared, a program that needed editing, a first article that failed and a paste change that was not planned. Each of them is a preparation failure rather than a machine limitation.

Addressing them in that order usually halves the interval without any capital spend. The remaining time is the genuine mechanical work, and it is much smaller than most lines assume before they start measuring. The first measurement is often a surprise, and it is the beginning of the improvement rather than a judgement on the team.

FAQ

What is a reasonable changeover time? It depends on the number of feeders, the paste and stencil change and the experience of the team, so a comparison with another line is rarely useful. The meaningful figure is your own trend over time.

Should the line be cleaned during or before the changeover? Cleaning belongs at the start of the changeover, because it is part of what makes the new product run cleanly. Deferring it to the end of the shift leaves residue in the machine while the new product is being built.

Can changeover be eliminated altogether? Not entirely, but it can be reduced by grouping products and by building more of each product per run. Where the demand does not support a long run, the goal is to make the changeover short and predictable rather than to remove it.