PCB Strain Gauge Monitoring

Board flexure is one of the few assembly stresses that can be measured directly, and measuring it changes what a factory believes about its own processes. A strain gauge bonded to the board surface records the deformation as the panel is routed, broken, handled, pressed into a fixture or pushed into a housing, and the peak value can be compared against a threshold that correlates with damaged joints and cracked components. The data is often surprising.

What Strain Does to Joints

Solder joints are brittle and tolerate very little bending before a crack begins. A crack may be partial at first, producing an intermittent connection that appears only when the assembly warms up, and it is exactly the kind of defect that survives every test on the line and fails in the field. Ceramic capacitors are similarly sensitive, and a crack in the body may not be visible externally at all, which makes the failure appear inexplicable when it eventually occurs.

The thresholds quoted in industry guidance are usually expressed in microstrain, and they are lower than most people expect. The consequence is that a handling or separation process that looks gentle can exceed the limit on a thin board with components near the edge. Warpage and stress control addresses the same physics from the design side, and the two activities should inform each other.

Where Strain Comes From

Depanelising is the largest single source. Routing and breaking both bend the board, and the stress is concentrated at the break line where components may be closest. Handling is the second: picking a board up by one corner, pushing it into a fixture, snapping a connector into place and stacking panels on top of each other all produce measurable flexure. Test fixtures that press a board onto probes are a third, particularly where the support is sparse.

Screw fastening and enclosure assembly are often overlooked. Driving a screw into a standoff that is not perfectly aligned bends the board, and the strain can be high enough to crack a joint near the mounting hole. Depaneling practice and mechanical assembly should both be assessed, because eliminating strain at one station while leaving another unmeasured gives a false sense of security.

Installing Gauges

The gauge is bonded to the board surface close to the components of concern, with the orientation chosen to capture the bending direction that matters. A single gauge measures strain along one axis, so a location that bends in two directions needs a rosette or two gauges. The bond itself must be stiff enough to transfer the strain, which means a thin, uniform layer of adhesive and a cure that follows the instructions.

The instrumentation is as important as the gauge. The signal conditioning must have enough bandwidth to capture the peak, because strain events are short and a slow system reports an average that hides them. The sampling rate and the recording window should be set from the event being studied, and the whole system should be checked on a known deflection before it is used to judge a process. Depaneling methods differ enough that each may need its own measurement.

Setting Thresholds

The threshold should come from the component supplier’s guidance and from your own failure experience rather than from a convenient round number. Where a product has failed in the field with a cracked joint, the strain recorded at the equivalent location during production is the evidence that connects the process to the failure. Where no such data exists, published guidance provides a starting point that should be tightened as experience accumulates.

Thresholds should be set per location and per process step, because the acceptable strain near a mounting hole differs from that near a fine-pitch component. The measurement also needs a rule about what happens when a threshold is exceeded: the affected units should be identified, the cause investigated and the process changed. A threshold without a reaction is a number that will be ignored.

Monitoring During Production

Continuous monitoring is rarely necessary; periodic surveys are usually enough. Measuring strain during a changeover, after a fixture change and at intervals through a shift shows whether the process is stable and whether the operators are handling the boards as the procedure requires. A short survey with a portable system, repeated whenever the process changes, is a practical and inexpensive programme.

The most useful application is comparing methods. Measuring the same panel with a router and with a score-and-break, or with and without support under the board, produces a direct comparison that justifies a process change on evidence rather than preference. That is the point at which strain monitoring stops being a compliance activity and starts paying for itself. Edge and clearance rules determine how much room the improved method has to work with.

Interpreting the Data

The peak value matters most, but its duration and the number of cycles also contribute to damage. Two events with the same peak but different durations do not do the same amount of harm, and a process that produces many small events can accumulate damage that a single large one would also cause. The data should therefore be reported as a waveform with the peak marked, not as a single number.

Where the strain at a location is consistently high, the fix is usually mechanical: better support, a different separation method, a change in handling or a design change that moves the sensitive component away from the stress concentration. Design changes are the most durable, which is why the data should be shared with the layout engineer rather than kept in the process department.

Building It into the Process

Strain measurement is most valuable when it is built into the qualification of a process rather than applied after a failure. Measuring during the first build of a new product establishes a baseline, and repeating the measurement whenever the panel, the method or the fixture changes catches the regressions. The result is the same discipline applied to thermal profiling, and it is justified in the same way, by the cost of the failures it prevents.

The records should be kept with the process documentation, including the gauge location, the threshold and the measured values. When a similar product arrives later, the previous data shortens the work considerably. Over several products, the factory accumulates a picture of which operations are inherently stressful and which methods reduce that stress, which is knowledge that survives the people who gathered it.

FAQ

How many gauges are needed? At least one at the location of highest stress, chosen from a visual assessment of the process. A second gauge at a reference location helps distinguish global bending from a local event.

When is strain measured? During process qualification, after any change to the panel or the fixture, and periodically during production as a check on handling.

What happens if the threshold is exceeded? Identify the units affected, investigate the cause and change the process. A threshold without a reaction is useless.

Does strain monitoring replace visual inspection? No. It identifies the processes that damage joints, which visual inspection cannot do reliably, but inspection remains the check on the product.