Flash Measurement System: What the Configuration Determines

One exposure fixes field and resolution together

A flash measurement system images the entire field in a single exposure and extracts every dimension from that image. Because there is no scanning and no stitching, the field and the resolution are set by the optics at the moment of capture and cannot be traded later.

This is the defining property of the approach and the source of both its speed and its constraint. A system configured for a large field cannot resolve small features, and one configured for fine detail cannot cover a large part in one shot.

Field of view across the POMEAS measuring configurations spans 0.15 to 200 mm, so the range covers both small precision parts and larger assemblies, at correspondingly different resolutions.

The configuration decision therefore has to be made from the part family, not from a single sample, because changing it later means changing the optics.

Stating the largest part and the smallest feature together is what makes this decision concrete.

Illumination is part of the measurement chain

In a single-exposure system every edge in the image is detected from the same illumination condition. A lighting setup that suits one feature may hide another, and there is no opportunity to change it between measurements.

Telecentric backlighting produces silhouette edges that are stable and easy to detect, which is why it dominates dimensional work. Coaxial and ring illumination reveal surface features that backlighting cannot see, at the cost of edge stability.

Most production systems combine several illumination channels and switch between them within the capture sequence, producing several images from one part presentation.

The number and combination of channels is part of the configuration, and it should be chosen from the features to be measured rather than from a standard package.

Illumination is also the most common reason a system measures one feature well and another poorly, which makes it worth testing on real parts.

Accuracy comes from calibration, not from the optics alone

A single-exposure system measures in image coordinates and converts to part coordinates through a calibration. The calibration accounts for magnification, distortion and sensor geometry, and its quality sets the achievable accuracy.

This is why two systems with similar optics can differ in performance: the calibration method and the artefact used determine the residual error.

Calibration must cover the whole field, because distortion grows toward the edge. A calibration valid at the centre does not apply at the corner.

It also has a lifetime: changing magnification, illumination geometry or the mechanical setup invalidates it, and so does thermal drift.

For a production station a scheduled verification with a reference artefact is more useful than a one-time calibration certificate, because it catches drift.

Throughput follows from the capture, not from the software

The measurement itself is fast because the whole field is captured at once. The cycle time is dominated by part handling, focusing and the number of illumination channels, not by the image processing.

This shifts the engineering effort toward the mechanism: how the part is presented, how it is held repeatably, and how quickly it can be replaced.

It also means that adding illumination channels for additional features has a real cost in cycle time, even though each capture is short.

Systems that appear fast in a demonstration often slow down in production because the handling was not part of the demonstration.

Evaluating throughput with the actual fixture and the actual part is the only reliable way to predict it.

Where a flash system fits and where it does not

It fits parts that are essentially two-dimensional, presented flat, with features that can be seen from one direction. Stamped parts, PCB features, gaskets, glass outlines and connector profiles are typical.

It does not fit features that require a different viewing angle, deep holes, or three-dimensional geometry where the measured dimension depends on depth. Those need scanning, a different sensor principle, or multiple presentations.

It is also a poor fit for very large parts where the required field forces a coarse resolution, or for very small features on a large part, where no single configuration satisfies both.

The POMEAS measuring configurations that publish both figures span a field of view of 0.15 to 200 mm at a working distance of 55 mm, which sets the practical envelope.

Recognising these boundaries before the specification is written prevents a system that is fast and unsuitable.

Specifying a flash measurement system

State the part envelope, the smallest feature to be measured, the tolerance on each dimension, and the required cycle time. Those four define the optics, the calibration requirement and the handling budget.

Then state the features that need illumination other than backlight, since each channel affects the configuration and the cycle time.

Confirm the field covers the part envelope with margin, and that the resolution at that field meets the tolerance for every feature, including those near the edge of the field.

Confirm the accuracy requirement against the calibration method, not against the optics alone.

A specification written this way can be quoted against directly, and it makes the trade-offs visible when a requirement changes.

Frequently asked questions

How fast is a flash measurement system?

The capture is a single exposure, typically a fraction of a second. Real cycle time is dominated by part handling, focusing and the number of illumination channels, so the mechanism usually determines throughput more than the camera does.

Can a flash measurement system measure a three-dimensional feature?

Not directly. It measures what is visible from one direction in a single presentation. Features that depend on depth, such as hole depth or side-wall position, need a different sensor principle or multiple part presentations.

Why does my system measure one feature well and another poorly?

Usually because of illumination. A setup that produces a stable silhouette for one edge may hide or distort another. Multi-channel illumination addresses this, at the cost of cycle time, and the combination should be chosen from the actual features.

How accurate is a flash measurement system?

Accuracy is set by the calibration, not by the optics alone. A full-field calibration with an adequate artefact removes most of the systematic error. Two systems with similar optics can differ significantly depending on the calibration method and its maintenance.

Does a larger field mean worse accuracy?

Not automatically, but it usually means coarser resolution, because the same sensor covers more area. If the tolerance requires finer resolution than the large field allows, the options are a smaller field with multiple presentations, or a higher resolution sensor at the same field.

What parts are a poor fit for flash measurement?

Parts with features that cannot be seen from one direction, deep holes, three-dimensional dimensions, and very large parts with very small critical features. Those need scanning, a different sensor principle, or several presentations.

Single-shot dimensional measurement

• Flash Measuring Machine