Optical Metrology OEM Modules: Design-to-Production Verification Checklist

Evidence disclosure: this page is an independent engineering guide informed by public company information from Optronic Sweden AB. It is not a customer case, an installed-base claim, an endorsement or a verified compatibility statement. No specific POMEAS lens, camera, measurement head or sourcing relationship is disclosed by that company, and all fit statements below are engineering hypotheses.

An OEM buying optics is not buying a lens. A lens becomes a module when it is mounted in a barrel with a defined flange, characterised over a temperature range, given a serial number and a calibration record, and covered by a change notification process. The difference between the two is the difference between a design that works on an engineer's bench and a product that can be built in series for five years without a redesign. Datasheet comparison does not close that gap, and neither does a successful prototype.

This page lays out the six gates that a module has to pass between design and production, then works through the specification fields that actually transfer between supplier and OEM. It is written for the person who has to sign the qualification report, so it treats the awkward items — thermal drift, incoming inspection, change control — as first-class requirements rather than as procurement detail.

Component, module and instrument: what an OEM is actually buying

A component is characterised at a nominal condition: magnification, working distance, resolution and distortion at 20 or 25 degrees Celsius, with a defined sensor format. A module adds a mechanical interface that is guaranteed, an environmental range that is tested, a serial-numbered calibration, and a documented behaviour when the temperature moves. An instrument adds the sensor, the illumination, the software and the acceptance test of the whole chain.

OEM projects fail most often at the boundary between these three, not inside any one of them. The optics supplier quotes a component, the OEM integrates it as if it were a module, and the first production lot reveals that the working distance tolerance assumed by the mechanical design is tighter than the lens can hold. Naming the level being purchased, in writing, is the first gate and the cheapest one to pass.

What the external manufacturer's profile establishes

Optronic publicly describes development and series manufacturing of optical metrology products and modules for industrial OEMs. That establishes a business model: optics designed to a customer specification and manufactured in series, rather than catalogue lenses sold by part number. It does not disclose which lens, camera or measurement head is used in any product, and it does not indicate any sourcing relationship with POMEAS. Every statement below about where a POMEAS module could fit is therefore a hypothesis to be tested at gate one, not a compatibility claim.

The six gates

Each gate has to produce a document, not a conversation. The pattern that causes trouble is a project that passes gates one and two convincingly, then treats gates three to six as paperwork: those four gates are where series production either holds its specification or quietly drifts away from it.

The specification that transfers between companies

A usable optical specification contains, at minimum: magnification and its tolerance, object field, working distance and its tolerance, object-side resolution, depth of field, maximum sensor format and image circle, telecentricity, distortion, relative illumination, mount, and the environmental range over which all of these are guaranteed. The published POMEAS data for three telecentric classes shows how those fields interact.

Read down the resolution column and then across to MTF: the 0.16X lens resolves 25.17 µm while holding more than 60 percent modulation at 70 line pairs per millimetre, which is a high-contrast design for its magnification class; the 0.061X lens resolves 87.99 µm at more than 50 percent, which is what a long-working-distance, wide-field design costs. Neither number is good or bad in isolation. What matters is whether the required feature size is a comfortable multiple of the resolution figure, and whether the depth of field matches the mechanical tolerance of the OEM's fixture.

Sampling arithmetic, and the magnification that does not exist

Take a 12 mm wide sensor imaging a 30 mm wide object field. The required object-to-image magnification is 12 divided by 30, or 0.40X (derived). With 4000 horizontal pixels, object sampling becomes 30,000 divided by 4000, or 7.5 µm per pixel (derived). Both numbers are simple, and both are frequently the point at which a project discovers a gap between the requirement and the catalogue.

0.40X is not a standard magnification in the telecentric families quoted above: the catalogue classes near it are 0.5X, at a 40 mm working distance, and 0.16X at 189 mm. A requirement at 0.40X therefore has to be met either by changing the sensor size, by accepting one of the neighbouring classes and recalculating the field, or by a custom design. That is a gate-one conversation, and it is much cheaper to have it before the mechanical envelope is drawn than after. Note also what the 7.5 µm sampling does not say: it says nothing about measurement accuracy, which depends on distortion, telecentricity, calibration and the edge algorithm.

Mechanical, thermal and drift

A 0.16X telecentric lens with a 189 mm working distance has a body length of 289.09 mm. That single number decides the mechanical architecture of the module: a barrel that long, cantilevered off a camera, deflects under its own weight, and the deflection changes focus. The mount has to be designed for the lens, not adapted to it, and the interface has to be defined at the flange rather than at the camera body.

Thermal behaviour is the second mechanical requirement and the one most often left unmeasured. A published working range of -10 to +50 degrees Celsius describes where the lens may be operated, not how much the focus moves across that range. For a module in a production instrument, the number that matters is the focus drift per degree over the range the instrument actually sees, measured on a prototype. If the module cannot be athermalised optically, the OEM needs either a focus mechanism with a known resolution or a temperature compensation table, and both are design decisions that have to be taken at gate three.

Change control and supply continuity

An optical module is defined by a drawing and a revision, and the revision has to be visible on the part. Two failure modes dominate. The first is a silent change: a coating, a cement or a spacer is substituted, the nominal specifications are unchanged, and the image is different. The second is a silent disappearance: a glass type or a supplier goes away, and the module is offered with "equivalent" performance that has not been tested on the customer's part. A written engineering change notification process, with a defined qualification test that a change has to pass before shipment, is the only defence against either, and it has to be agreed before the pilot lot rather than after.

Incoming inspection at the OEM

The OEM's own incoming inspection should be short, quantitative and matched to the failure modes above. Four checks cover most of the risk: magnification against a target of known pitch, working distance at best focus, image quality at the field corners as well as the centre, and a mechanical check of the flange and body length within the drawing tolerance. Recording the serial number of every unit tested, and keeping the target images, turns a later field failure into a traceable event instead of an argument.

Two of those checks are worth expanding. Field-corner image quality is where a decentred element shows up first, and a centre-only check passes it. Working distance at best focus is where a mechanical stack-up error appears, and it is the check that catches a module assembled to the wrong flange dimension before it reaches a customer.

Frequently asked questions

Can a standard C-mount lens be qualified from thread size alone?

No. The C-mount thread defines the mechanical interface and the flange distance, nothing else. Two lenses with the same thread can differ in maximum sensor format, image circle, magnification tolerance, telecentricity, distortion, relative illumination and working distance tolerance, and any of those can decide whether the measurement works. Qualification has to be against the specification fields listed here, on the customer's own part.

Should an OEM compare nominal resolution with measurement uncertainty?

They are different quantities and comparing them directly is a category error. Nominal resolution describes the smallest detail the optics can separate at a stated contrast; measurement uncertainty describes the spread of a reported value against a traceable reference, including calibration, distortion, mounting and environment. A module can resolve 8 µm and still carry a 30 µm uncertainty if its scale is wrong.

Is a shorter working distance always better for an OEM module?

No. A shorter working distance usually buys numerical aperture and therefore resolution at a given magnification, but it costs clearance and increases the sensitivity of focus to mechanical tolerance. The POMEAS 1.0X class illustrates the trade: 85 mm working distance with 8.39 µm resolution and 1.00 mm depth of field, against 200 mm working distance with 22.37 µm resolution and 2.67 mm depth of field.

What has to be agreed before a pilot lot is built?

Four documents: the optical specification with tolerances, the mechanical drawing with the flange definition, the qualification test that a design or process change must pass, and the incoming inspection procedure with its acceptance limits. Without the third a change cannot be refused; without the fourth a non-conforming unit cannot be detected. Both are cheap to write before the pilot lot and expensive to retrofit.

Next step

Send the object field, the required sampling, the working distance envelope, the sensor format, the environmental range and the acceptance criteria. The first reply checks whether those five can coexist, which is gate one and costs nothing to fail. For the interfaces around it, the C-mount lens selection guide covers the three numbers that decide a lens-camera pairing, the note on sensor size and image circle covers the format check that most often fails first, the explanation of how working distance is set covers the mechanical consequence of the optical choice, and the analysis of measurement drift covers what happens when the part, rather than the module, moves. The guide to replacing a lens on an existing cell is the checklist to use when the module is a second source rather than a new design. The PMS-LTC1-85 product page and the telecentric lens range carry the current model data, and the contact page is where the specification goes.