Strain Gauge: Design Rules and Process Limits

A board can be perfectly soldered and still fail later because the assembly process bent it. Strain measurement is the tool that turns that suspicion into a number. It is used to qualify a new fixture, to compare two depaneling methods, to set a conveyor support layout and to prove that a change has reduced the load on the joints that matter. This article explains what a gauge measures, how to place it, what limits are reasonable, and how to make the result change something on the line.

Why Strain Is Measured on an Assembly

Printed circuit boards are made of materials that are stiff but not infinitely so, and many of the components mounted on them are brittle. A small deflection of the board becomes a large local strain at a solder joint, particularly at the corner of a large ceramic package. Strain measurement quantifies that deflection at defined points so that the process can be judged against a limit instead of against opinion. A board bend that is invisible to the eye can still exceed the limit at the joint that carries the load.

The technique is most valuable at transitions. A new fixture, a new board support, a change of conveyor width, a different depaneling method or a heavier component all change the load, and each of them is a good reason to measure rather than assume.

What a Strain Gauge Actually Reports

A strain gauge is a resistive foil bonded to the surface, and its resistance changes as the surface stretches or compresses. The signal chain converts that change into microstrain, which is a dimensionless measure: one thousand microstrain corresponds to a change in length of one part in a thousand.

The gauge reports the strain on the surface it is bonded to, at that location and in that direction. It does not report the strain inside the board, the stress at a solder joint below the surface, or the load anywhere else on the assembly. Interpreting a reading as a general property of the board is one of the common mistakes in this kind of testing.

Gauge Selection and Placement

Small gauges with a short grid length suit the tight spaces around a component, while a longer grid averages over a larger area and is less sensitive to local variation. A single gauge measures strain in one direction, so a rosette is needed when the principal direction is unknown and the shear component matters.

Placement should follow the question. If the concern is a ceramic capacitor near a breakaway tab, the gauge belongs next to that capacitor on the axis of the bend. If the concern is the board handling in a machine, the gauge belongs where the board is supported and where it is free to flex. Our component tolerance notes explain why the placement of a large package decides how much strain it can tolerate.

Attachment and Wiring Practice

Bonding is where most measurement error is introduced. The adhesive has to be thin and uniform, the gauge has to be aligned with the axis of interest and the cure has to follow the adhesive supplier instructions. A gauge with a bubble or a thick glue line reports a distorted value and often debonds during handling.

The lead wires are part of the mechanical system, not just the electrical one. A stiff wire tied to the board locally stiffens the very area being measured, and a wire that tugs at the gauge changes the reading when the board moves. Strain relief at the board edge and a flexible routing path are the usual fixes.

Reading the Data: Microstrain and Rate

The magnitude of the strain is only half of the story. The rate at which it is applied matters as much, because a slow bend that allows the solder to creep is far less damaging than a fast one of the same amplitude. Data should therefore be captured with enough sample rate to show the rise time of the event.

A typical trace recorded through a conveyor passes through several events: loading, support engagement, a soldering or reflow thermal step, and unloading. Attributing a peak to the correct event requires a record of what the line was doing at that moment, which is why a test is usually run with a video or a machine signal log alongside the strain data.

Typical Limits and Where They Come From

Published limits for board handling exist and are widely quoted, but they are generic and they assume a particular component population. The limit that matters for a given assembly is the one derived from the components on it, and that limit is usually set by the most brittle part rather than by the board.

A component supplier may publish a strain limit for a specific package, and a process that is comfortably inside that figure for one package can be marginal for another mounted nearby. Where no published data exists, a step-stress test that bends a sample until joints crack gives a starting point, with a safety factor applied. In many field returns a solder joint crack that appears weeks after build is the delayed result of exactly this kind of loading.

Sources of Strain on a Production Line

The conveyor is a frequent source. A board that spans the rails without support deflections under its own weight and under the load of a stencil squeegee or a placement head. The classic damage case is a chip capacitor cracked by a combination of board flexure and a large thermal expansion mismatch, and it often appears long after the line that caused it.

Depaneling, screw fastening, connector insertion, test fixture clamping and even manual handling of a large panel are all significant. Screw fastening deserves particular attention, because an operator driving a screw into a standoff can bend a board well beyond what the machine does.

Using the Result to Change a Process

A strain measurement becomes useful when it identifies the worst event and that event is fixed. Adding support pins under a heavy area, widening the conveyor support, reducing the screw torque, adding a washer to spread the load or changing the depaneling method are all interventions that can be verified by repeating the measurement with the same gauge layout.

The before and after comparison is what makes the case for the change. Without a repeated measurement, a modification is as unproven as the original design was, and the improvement may exist only in the intention.

Ongoing Monitoring in Production

Strain gauges are rarely fitted to every board, but the measurement can be periodic. A reference board with gauges bonded at fixed locations, run through the line at intervals, detects drift in a fixture or a support before it produces a defect. The same discipline that governs the production process flow treats that reference board as a calibrated tool rather than as a sample.

At gopcb the manufacturing tolerances that determine how much the board can flex are held in our fabrication notes, and the failure modes that appear when those limits are exceeded are described in our solder defect guide. Our quality guide covers how the resulting defects are classified.

Additional Considerations for This Build

Practical attention to board bend pays for itself here, because it decides whether the finished board behaves as the drawing intended. Where the requirement is not stated on the fabrication drawing or in the assembly notes, the shop has to assume a default, and that default is rarely the value the design was simulated with. Stating board bend explicitly, together with the tolerance that applies, removes the assumption and keeps the result predictable from batch to batch.

Deliberate attention to solder joint crack pays for itself here, because it decides whether the finished board behaves as the drawing intended. Where the requirement is not stated on the fabrication drawing or in the assembly notes, the shop has to assume a default, and that default is rarely the value the design was simulated with. Stating solder joint crack explicitly, together with the tolerance that applies, removes the assumption and keeps the result predictable from batch to batch.

FAQ

How many gauges are needed for a valid measurement? Enough to cover the locations where the risk is concentrated, which for many assemblies means one near the most brittle component and one at the point of maximum deflection. A single gauge can easily miss the peak.

Can strain be measured on a bare board instead of a populated one? Yes for characterising the handling equipment, because the mass of the assembly changes the deflection only slightly in most cases. For a component limit, the populated assembly is the correct subject.

What sample rate should be used? Fast enough to resolve the rise time of the fastest event expected, which for a machine handling step is usually a few hundred hertz or more. A slow sample rate hides the peak that causes the damage.