Telecentric Lens Sensor Size: Matching Format to Field and Accuracy

Sensor format fixes the field before anything else

A telecentric lens holds magnification substantially constant across its working range, which makes the relationship between sensor and field unusually simple. Field of view equals sensor size divided by magnification. There is no perspective term and no distance dependence inside the telecentric range, so the field is set the moment the format is chosen.

That simplicity is also the constraint. Because the field cannot be adjusted by moving the camera, changing the field means changing either the sensor or the magnification. A layout that assumes the field can be trimmed on site will not work with a telecentric lens.

The practical order is therefore: fix the field from the part, choose a magnification that maps that field onto a standard sensor, then confirm the chosen format is available with the required interface and readout speed.

In the POMEAS telecentric range magnification spans 0.061 to 12.5 X which means most field requirements can be met at more than one format. The choice of format then becomes a resolution decision rather than a coverage decision.

Getting this order right removes most of the rework in a telecentric selection, because field errors are the hardest to correct after the machine is built.

The image circle is a hard limit, not a guideline

Every lens projects a corrected image over a finite circle. Outside that circle the image falls off in brightness and the correction for distortion and telecentricity no longer holds. A sensor larger than the image circle does not gain field: it gains vignetting in the corners and degraded accuracy at the edge of the frame.

Manufacturers state the maximum supported format, and that statement should be treated as a ceiling rather than a suggestion. Mounting a larger sensor behind a smaller image circle produces images that look acceptable in the centre and fail at the edges, which is exactly where a measurement application is least able to tolerate it.

The failure mode is subtle in visual inspection and obvious in metrology. A part measured near the frame edge with a mismatched sensor will show an error that changes with position and does not reproduce, and the cause is easy to misattribute to the stage or the calibration.

For measurement work the usable area is slightly smaller than the stated image circle, because telecentricity and distortion are specified over a defined region. Treating the stated format as the working format leaves no margin.

The safe procedure is to select the format first and then confirm the lens states support for it, rather than selecting a lens and hoping the sensor fits inside the circle.

Field and resolution pull in opposite directions

For a fixed sensor, increasing magnification shrinks the field and improves object-space resolution by the same factor. Decreasing magnification does the reverse. The two requirements therefore have to be resolved together rather than sequentially.

A useful way to state the requirement is as a resolution per field ratio: what is the smallest feature that must be resolved at the edge of the largest part. That single number, combined with the sensor's pixel count, determines whether the application is feasible at all with a given format.

In the POMEAS telecentric range field of view spans 0.4 to 265 mm and object-space resolution spans 1.1 to 398 um. Those ranges do not overlap arbitrarily: the models with the largest fields are not the models with the finest resolution, because both are governed by the same optical magnification.

This is why a specification that asks for a large field and a fine resolution at the same time has to be checked against the sensor pixel count before any model is quoted. If the required resolution per field exceeds what the sensor can sample, no lens will deliver it.

Stating the requirement in this form early prevents a long search for a model that does not exist.

Depth of field interacts with the format choice

Depth of field in a telecentric system is the height range over which the magnification and the focus remain within tolerance. It is specified per model and shrinks quickly as magnification rises, so a format chosen for fine resolution can leave too little depth for a part with height variation.

Across the POMEAS telecentric range depth of field spans 0.017 to 172 mm, which is a spread of four orders of magnitude. The high magnification models sit at the narrow end, and that is a property of the optics rather than a defect of any particular design.

The consequence for format selection is that a part with significant height variation, such as a stamped component with burrs or a casting with draft, may need a lower magnification and therefore a larger sensor to keep the field. The two constraints have to be satisfied simultaneously.

Where the height variation is unavoidable and the magnification cannot be reduced, the alternative is to measure at several focus heights and combine the results, which costs cycle time. That trade is worth making explicitly rather than discovering it during commissioning.

Checking depth of field against the real height tolerance, instead of the nominal flatness, is what keeps this constraint from surfacing late.

Connecting format to the measurement tolerance

The measurement tolerance determines how much of the sensor can be used. A tight tolerance requires the part to stay near the optical axis where telecentricity and distortion are best, which effectively reduces the usable field and therefore pushes the choice toward a larger format at lower magnification.

Telecentricity across the POMEAS range is specified between 0.02 to 0.04 deg, and distortion between 0.002 to 0.06 %. Both are edge-of-field properties, so the fraction of the frame actually used is part of the accuracy budget.

A practical rule is to keep the measured feature inside the central portion of the frame when the tolerance is tight, and to accept the full frame only when the tolerance allows it. This is a layout decision that follows from the format choice, not a correction applied afterwards.

If the tolerance cannot be met anywhere in the frame at the required field, the application needs a different magnification and a larger sensor, or a different measurement principle. Recognising that early avoids a specification that no telecentric lens can satisfy.

Writing the tolerance next to the format in the requirement document makes this relationship visible to everyone reviewing the design.

A format selection sequence that avoids rework

First, state the field as the bounding box of the largest part plus a margin for positioning error. Second, state the smallest feature that must be resolved and the measurement tolerance. Third, choose a magnification that maps the field onto a standard format.

Fourth, check that the chosen magnification gives enough depth of field for the real height variation. Fifth, confirm the format is inside the lens image circle. Sixth, check the interface and readout rate of the camera at that format.

Each step can send the design back to an earlier one, and that is expected. What matters is that the loop is closed before a model is ordered, rather than during commissioning.

The sequence also produces a specification document that a supplier can quote against unambiguously, because every requirement is expressed as a number with a condition attached.

Following it consistently is what turns a telecentric selection from an exercise in comparing datasheets into a short, checkable decision.

Frequently asked questions

Can I use a larger sensor than the lens image circle?

No, not usefully. Outside the image circle the illumination falls off and the telecentricity and distortion correction no longer hold, so the extra sensor area produces vignetted, inaccurate image data. The stated maximum format is a ceiling for the optical performance, not just for brightness.

Does changing the working distance change the field of view?

Not within the telecentric range. That is the defining property of the design: magnification stays substantially constant, so field equals sensor size divided by magnification regardless of distance. Outside the range, or with a partial telecentric design, the field does start to change.

How do I get both a large field and fine resolution?

By using a sensor with more pixels at the same format, not by changing the lens. The lens fixes the resolution per field ratio. If the required ratio exceeds what the sensor can sample, no lens will deliver it, and the application needs either a larger sensor or a different measurement approach.

Why do high magnification telecentric lenses have so little depth of field?

Because depth of field falls roughly with the square of magnification for a given aperture. A 5X model has a far narrower usable height range than a 0.5X model by design. Parts with height variation therefore usually need lower magnification and a larger sensor to keep the field.

Is 2/3 inch still a reasonable format for telecentric work?

It remains common where the field is small and the tolerance is moderate. For larger fields or tighter tolerances the trend is toward 1 inch and above, because the same field at the same resolution needs more pixels. The right answer follows from the field and resolution requirement, not from convention.

What happens if the part extends beyond the field of view?

The measurement is incomplete, and with a telecentric lens the field cannot be expanded by moving the camera. The options are a lower magnification with the same sensor, a larger sensor at the same magnification, or stitching several fields, which costs cycle time and introduces seam handling.

Lighting and camera lens selection

• Machine Vision Lighting Selection Basics
• Industrial Camera Lens Selection