Placement count and changeover: the two numbers behind an SMT price

Two numbers explain most of what an SMT line charges: **how many parts get placed**, and **how often the line gets set up**. The first is per board, the second is per job, and treating them as one number is why quotes for small batches so often surprise people.

Machine cost tracks placements

A placement machine works at a rate in placements per hour. Cost per board is therefore placement count divided by that rate, times the machine rate — not a flat time per board.

Use your line's **actual** achieved rate, not the machine's headline figure. The headline is measured under ideal conditions with ideal parts; your rate includes the odd-shaped parts, the fine-pitch parts that need slower placement, and everything else that makes real boards real.

Reflow is measured differently

An oven takes a board through a thermal profile in a fixed time, regardless of how many components are on it. So reflow is costed **per board**, or per panel if you run arrays — while placement is costed per placement.

This is why panelization pays: several boards in one array pass through the oven as one item. It is also why a sparse board is proportionally more expensive to reflow than a dense one, which is the opposite of the placement effect.

Changeover: feeder setup and first article

Between jobs the line is reloaded — feeders changed, program loaded, stencil swapped, first article built and checked. That cost happens once per job and barely varies with quantity.

Divide it by the boards in the batch. Which means the same board costs measurably more in a small batch, and that is not a pricing trick — it is the actual cost behavior, and the sheet should be able to show it to a customer asking why.

Three improvements the same two numbers prove

1. Reduce placement count through design — fewer, larger-value parts where the circuit allows. This one keeps paying on every board forever.
2. Panelize so more boards pass through the oven at once, cutting reflow cost per board directly.
3. Group jobs that share feeder setups, so one changeover serves several — scheduling, not capital.