SMT Line Balancing and Takt Time in Production

A surface mount line is a chain of machines, and a chain moves at the speed of its slowest link. Line balancing is the practice of matching the work at each station so that no machine waits and no machine becomes a wall of work in progress. It is the difference between a line that runs at its rated speed and a line that runs at whatever pace the bottleneck allows. This guide explains how gopcb approaches line balancing on its SMT lines.

What Line Balancing Solves

An unbalanced line wastes capacity in two directions at once. Stations ahead of the slowest machine finish early and then sit idle, while stations behind it are starved. The line output is set by the slowest step, so the investment in every other machine is partly idle regardless of how capable those machines are.

Balancing also controls work in progress. When stations run at different rates, boards accumulate in front of the slowest one, which hides defects, consumes floor space and makes traceability harder. A balanced line keeps flow smooth, which makes problems visible as soon as they occur. Balance also affects morale, because operators at the slowest station spend a full shift under pressure while others wait for work.

<img src="https://www.gopcba.com/wp-content/uploads/2024/09/Automotive_PCBs.webp" alt="SMT line balancing chart showing takt time and station cycle time” />

Takt Time and Customer Demand

Takt time is the pace a line must sustain to meet customer demand. It is calculated by dividing available production time by the number of units required in that period. If a line must ship five hundred boards in a twenty hour production window, the takt time is one board every two point four minutes.

Takt time is a target derived from demand, not a measure of the line itself. It tells you what the line has to achieve, while cycle time tells you what each station actually achieves. Comparing the two shows immediately whether the line can meet the schedule with the equipment and staffing it has.

Cycle Time and the Bottleneck

The bottleneck is the station whose cycle time is longest. On an SMT line it is often the placement machine handling the largest number of components, or the printer if the product needs a slow print stroke. Identifying it requires measured data, because the constraint changes with the product mix and is not always where it appears to be.

Improving anything other than the bottleneck does not increase output. This is the central rule of balancing, and it is frequently ignored because non bottleneck stations are easier to improve. Time spent speeding up a station that already waits is time that produces no additional boards. The exception is a station that feeds the bottleneck indirectly, such as a printer that determines whether the placement machine has work waiting.

Buffers, Work in Progress and Starvation

Small buffers between machines absorb short interruptions, such as a feeder change or a quick inspection, without stopping the line. Without buffers, any pause at one station immediately starves the next. With buffers that are too large, defects are hidden for longer and the line becomes harder to control.

The practical approach is a deliberately small buffer at each interface, sized to cover a normal interruption but not a real breakdown. When buffers are consistently full or consistently empty, that is a signal that the stations around them are out of balance and need to be re-examined. Buffer size should also be reviewed whenever the product mix changes, because a buffer that suited one family can hide problems in another.

Balancing Placement Machines

Placement is usually where balancing has the most leverage, because components can be distributed between machines. Feeders can be moved, components reassigned and programs rewritten so that each machine carries a similar number of placements. The goal is not equal component counts but equal cycle times.

Component mix complicates this. A machine placing twenty large connectors may take longer than one placing two hundred chip components, so the distribution must be based on time, not count. Placement data from the machine itself gives the most reliable basis for the split. Reassignment also moves feeders, so the travel distance per head belongs in the calculation alongside the number of placements.

Changeover and Mixed Model Lines

Changeover is part of the balancing problem on any high mix line. Setup time at the printer or the placement machines reduces available production time, which lengthens the effective takt time. Reducing changeover and balancing the line are therefore the same project seen from two directions.

Mixed model lines add a further constraint, because different products have different bottlenecks. Balancing for the average product can leave the largest product constrained and the smallest product over-served. The usual answer is to balance for the dominant product family and manage the others with buffers.

Measuring and Iterating

Balancing is a measurement activity before it is an engineering activity. Machine cycle times, queue lengths, wait states and stop reasons should be logged automatically and reviewed weekly. The data should be read against the schedule, because a line balanced for a twenty hour production window is out of balance as soon as that window shrinks. Manual stopwatch studies are useful occasionally, but they cannot capture the variation that occurs across a full shift.

The review should end with one change at a time and a re-measurement. Moving feeders at three machines and rewriting four programs at once makes the result impossible to attribute. The same discipline that applies to process improvement, described in our cost reduction notes, applies here.

Balancing Quality Steps as Well

Inspection stations are part of the line and must be balanced too. An inspection step that takes longer than the takt time becomes the new bottleneck, no matter how fast the machines around it are. Optical inspection cycle time depends on board size, component count and the number of images required.

Test strategy should be reviewed alongside balancing. Where every board passes through a slow inspection step, sampling or a reduced inspection set may release capacity without weakening coverage. Our notes on judging PCB quality discuss the trade-off between thoroughness and flow.

Practical Limits of Balancing

Perfect balance is not achievable, because cycle times vary with the product and machines do not divide as neatly as the arithmetic suggests. The realistic goal is to reduce the gap between the slowest and fastest stations to a level that small buffers can absorb, rather than to chase a mathematical ideal that no real product mix will reproduce.

Line control systems make this easier by giving real time visibility of station status and queue length. The monitoring approach used on automated lines, described in our notes on industrial SCADA systems, can be adapted to an SMT line to show the bottleneck as it moves through the shift.

FAQ

What is the difference between takt time and cycle time? Takt time is the pace required by customer demand, calculated from available time divided by units needed. Cycle time is how long a station actually takes to complete its work. Comparing them shows whether the line can meet the schedule.

How do I find the bottleneck on an SMT line? Measure cycle time at every station and look for the longest one, then confirm it with queue data. Boards accumulating in front of a machine, or that machine running continuously while others wait, are strong indications of where the constraint sits.

Does balancing matter on a high mix line? Yes, but the constraint moves with the product. Most high mix lines balance for the dominant product family and absorb the rest with buffers and flexible staffing, rather than attempting a single balance point that suits every product.