Bow and Twist Measurement in PCB Panels

Bow and twist are the two ways a panel loses its flatness, and they are measured differently, caused differently and corrected differently. They are grouped together in most specifications, which is a source of confusion when a lot is rejected and nobody is sure which property failed. Separating them makes both the measurement and the corrective action clearer.

Bow and Twist Defined

Bow is a curvature of the board about a single axis, so the panel takes on a cylindrical shape. Twist is a warping about a diagonal, so one corner lifts relative to the others and the panel takes on a saddle shape. A panel can exhibit both at once.

The distinction matters because the measurement differs. Bow is measured against a flat plane using the maximum deviation of the surface, while twist is measured by holding three corners down and measuring how far the fourth lifts.

How Each Is Quantified

Bow is expressed as the maximum distance between the surface and a reference plane, divided by the length over which it is measured, and reported as a percentage. Twist uses the same form of expression but the reference is the corner displacement across the diagonal.

The measurement requires a flat reference, a defined support arrangement and a stable temperature. A panel measured at one temperature and specified at another will not agree, which is why the test conditions belong in the specification along with the number.

What Causes Them

The dominant cause is asymmetry in the construction. Copper distribution that is unbalanced between layers, a stackup with an odd number of layers built from symmetric prepreg, or a difference in resin content across the panel all produce internal stresses that appear as flatness error when the board is released from the press.

Thermal history also matters. Every lamination cycle and every reflow adds to the accumulated stress, so a board that is flat after fabrication can bow after assembly. That progression is why the requirement should relate to the condition at the point of use rather than only to the delivered panel.

Controlling It in Fabrication

Balanced copper keeps the shrinkage on each side of the neutral axis similar, and it is the single most effective control. Where a design needs a large copper area on one layer, a compensating pattern on the corresponding layer opposite reduces the imbalance.

Lamination parameters matter as well. A press cycle that cures the resin too quickly creates a gradient through the thickness, and the resulting stress appears later. The parameters that control this are set by the fabricator, which is why the design should state the requirement rather than the method. Our article on PCB warpage control covers the design side in more detail.

Assembly and Handling Contributions

Reflow is the largest single thermal event a bare board sees, and it can move a panel that was within limits on delivery. The effect is worst for thin boards with large copper areas and for boards processed in a panel that is not supported.

Storage contributes as well. A panel stored under load, leaning against a rack or stacked unevenly, can take a permanent set over time. Flat storage on a shelf with support across the full area removes that contribution at negligible cost. The outline tolerance aspects are covered in our note on outline tolerance.

Specifying a Workable Limit

The limit should be set by what the assembly process can tolerate rather than by a default. A board that has to be placed on a vacuum table and printed with fine pitch paste is far less tolerant of bow than one that is hand assembled.

A common figure is 0.75 per cent for surface mount boards and 1.0 per cent for others, measured after reflow. Tightening below what the process needs increases cost and yield loss without improving the assembly.

Corrective Action

Where a lot fails, the first question is whether the cause is in the construction or in the process. A construction problem, such as an unbalanced stackup, will recur on every order until the design is changed, while a process problem may be a single press cycle.

Re-baking a board under weight sometimes restores flatness, but it is a temporary measure and the board may return to its distorted shape after the next thermal cycle. The durable answer is a design and process change, and the evidence needed to justify it follows the same pattern as any quality decision described in our article on PCB quality assessment.

Checks Before Release

Documentation exists so that a person who was not present can reproduce the work and reach the same conclusion. The acceptance criteria should be written before the work starts, so that the decision is made by the specification rather than by the person inspecting.

The environment around the process, including temperature, humidity and cleanliness, sets limits on what the process can hold. A result that cannot be reproduced is not a result, and reproducibility should be demonstrated rather than assumed.

The tooling, the material and the profile form one system, and a change to any of them should be assessed against the other two before it is released. Handling between operations is part of the process, and the damage it causes is often attributed to the operation that preceded it.

Points to Confirm at First Article

Where a process is at the edge of its capability, the margin should be bought deliberately rather than discovered during production.

FAQ

Does baking fix a bowed board? It can temporarily, if the board is restrained while it cools. The effect is not permanent and the board may distort again after assembly.

Is twist or bow more serious? Twist is usually worse for assembly, because a lifted corner prevents the board from seating on a flat table.

Should the limit apply before or after reflow? After reflow, if the requirement is to assemble the board. A board that meets the limit on delivery but not after reflow will still cause problems.