Optimization of PCB Panelization Design

Panelization is the last step of a PCB design and the first step that decides what the board costs to build. How many boards go on one panel, how they are joined, and how much border is left around them set the material utilization, the placement cycle time, the number of fixtures, and the yield of the separation step. None of that shows up in the schematic, which is why it gets decided late and reviewed least.

This page is about the decision rather than the drawing. It covers what a panel actually changes, where each of the three connection methods fails, how the product stage and the soldering route change the answer, and how to compare two schemes that are both workable. For the values themselves, meaning border width, fiducial size and placement, tooling holes and spacing, see the SMT panel design rules.

PCB panel with several identical boards joined by breakaway tabs, showing the border rail and tooling area

What a Panel Actually Changes

A panel does four things at once, and each one moves a different cost:

  • It reaches the machine minimum. A board smaller than the smallest board the placer accepts cannot be run on its own, so panelization is not an optimization in that case; it is the only way the part gets assembled.
  • It reduces changeovers. Different single boards can share one panel, which turns several production runs into one and removes the setup time between them.
  • It sets the tooling count. Stencils, fixtures and test jigs are built for the panel, not the single board, so the panel is what the tooling budget is actually spent on.
  • It turns separation into a process. On a panel, depanelization is a controlled step with a known stress and a measurable edge quality. Without one, boards get snapped apart by hand and the cracks that follow appear later as field failures.

The Three Connection Methods and Where Each Breaks

V-scoring, breakaway tabs and bridging are the three routes in common use. The useful question is not which is better but how each one fails, because that is what decides whether a particular board can use it.

  • V-scoring cuts a groove on both faces and leaves roughly a third of the board thickness in the middle. It requires straight lines, so a curved outline cannot use it, and the groove means components have to stay a few millimeters clear of the score line or they crack during separation. Its failure mode is a rough or chipped board edge.
  • Breakaway tabs, often cut with castellated holes at the tab root, work on irregular outlines and curves where a score line cannot run. The tab has to be wide enough to survive handling and narrow enough to snap cleanly, and the broken edge leaves a small burr that has to be dressed. Its failure mode is a tab that tears into the board instead of breaking at the perforation.
  • Bridging joins boards with a short neck and no drilled holes, and is used for half-hole boards and thin panels where a tab would add too much material. It is the weakest of the three in handling, so it suits light boards and short runs. Its failure mode is a bridge that flexes and fatigues before the panel reaches assembly.

In practice the outline and the depanelization method choose the connection before any cost argument does. A rectangular board that will be scored cannot use tabs to save money, and a board with a curved edge has no score line to cut.

Key Factors Influencing a Panelization Scheme

Once the connection is settled, three things move the panel size and the layout: the stage the product is at, how thick the board is, and which soldering route it will run through.

Product Stage

  • Development. The design is still moving, so the panel is kept small and built to satisfy the equipment rather than to optimize material. A panel that has to be re-cut every time the outline changes costs more in tooling than it saves in laminate.
  • Small batch. The development panel can usually be carried over, but it is worth checking whether the efficiency and cost side now justifies a larger one. The fixture and jig count follows whichever way that goes, so the decision is not free.
  • Mass production. Here the panel is chosen for the total cost, which means accepting a worse material utilization if it buys a better cycle time or fewer changeovers.

Board Thickness and Warpage

Thickness sets the upper limit on panel size more often than the machine does. A thin board on a large panel sags in the middle during reflow: the panel does not sit flat, placement height varies across the panel, and the joints near the middle of the panel are the ones that open later. Below about 0.8mm, the answer is usually a smaller panel, a wider border rail to carry the stiffness, or a support bar across the panel rather than more vacuum under it.

A thick board fails the other way. At 2mm and above, a V-score does not leave enough material to separate cleanly, so boards either need tabs or need the panel to be routed, and the separation step gets slower and rougher at the same time. Thickness is therefore one of the first things to check before a scheme is drawn, not something to accommodate afterwards.

Soldering Process

Reflow and wave soldering pull the panel design in almost opposite directions, and this is the factor most often left out of a scheme review.

  • Reflow constrains warpage and thermal mass. The larger the panel, the greater the temperature difference between its center and its edge, so components on the same panel see different profiles. A double-sided build adds a second question: the panel has to survive the second pass without dropping the components placed on the first.
  • Wave soldering constrains direction and shadowing. The solder wave travels in one direction, and any component standing in front of a through-hole joint blocks it, so the panel has to be oriented so that every through-hole joint has a clear path to the wave. That orientation is fixed by the panel, not by the individual board.
  • Mixed assemblies get both, because they run through SMT assembly and then through wave or selective soldering. The two sets of requirements rarely agree, and the panel is normally built to the wave soldering constraint because it is the harder one to work around later.

This is also why the same board can come back with two different panel recommendations from two suppliers without either being wrong. The soldering route in the assembly and welding process decides the constraint, and the panel follows it.

How to Compare Two Feasible Schemes

A board of any complexity usually has several workable panels, and at that point the question stops being which one is correct and becomes which cost it trades away. Five numbers settle most of these arguments:

  1. Material utilization: the board area divided by the panel area, after the border rail is taken out. It is the number everyone quotes and the least useful on its own.
  2. Cycle time per board: placement and inspection time divided by the number of boards on the panel. A panel with more boards can have a worse number here, because the travel distance across the panel grows and the placer spends longer moving.
  3. Tooling count: how many stencils, fixtures and test jigs the scheme needs. A panel that mixes several boards reduces runs but multiplies tooling.
  4. Separation yield: the scrap from chipped edges, torn tabs and boards damaged in handling. This is the number that is missing from most comparisons and the one that shows up as a surprise in the monthly report.
  5. Changeover count: how many production runs the scheme creates over the order quantity. Fewer runs is not automatically better if it forces a panel that places badly.

These five routinely conflict. The panel with the best material utilization is often the one with the worst cycle time, because it was packed to fill area rather than to suit the machine travel. That is the point at which the decision belongs to whoever owns the total cost rather than to whoever owns the drawing, and it is the reason a panel review belongs in design for manufacturing and not in the production queue.

The one thing worth fixing before any of these numbers are compared is the connection method, because it removes options from all five at once. Get the outline, the thickness and the soldering route settled first, and the comparison usually narrows to two candidates instead of six.

Shopping Cart
Scroll to Top