What Is a Telecentric Lens?

A telecentric lens is a compound lens system engineered so that its chief rays travel parallel to the optical axis. Because magnification is determined by chief-ray geometry rather than by object distance, the image of a part keeps the same size even when the part moves closer to or farther from the lens. In machine vision, this constant-magnification behavior eliminates the perspective error that makes ordinary lenses unreliable for dimensional measurement.

That single paragraph is the whole idea — but the engineering behind it, and the trade-offs that come with it, are what you need to understand before specifying one. This article explains how a telecentric lens works, how object-space, image-space and bi-telecentric designs differ, what specifications actually matter, and when a telecentric lens is worth the extra cost compared with a standard FA lens.

Part of the POMEAS Machine Vision Optics Technical Guides series — see the Machine Vision Optics Technical Guides index for all 20 articles in this set, grouped by topic.

Why Ordinary Lenses Fail at Measurement

With a conventional photographic or FA lens, the chief rays from each object point converge toward the aperture stop at an angle. Move the object 1 mm closer to the lens, and its image becomes slightly larger. This is perspective (parallax) error, and it produces three very practical problems on a production line:

• Size drift. The same part measures differently depending on where it sits along the Z axis. A 10.00 mm shaft can read 10.03 mm one moment and 9.98 mm the next, purely because of fixturing variation.
• Shape distortion. Features at the edge of the field of view are seen obliquely. A round hole images as an ellipse, and the side wall of a deep bore becomes visible when the inspection really needs to see the bottom.
• Unstable calibration. Any magnification drift between parts or between fixtures forces recalibration — and recalibration costs line downtime.

Software cannot fully fix these effects. Sub-pixel edge detection and distortion correction help, but they cannot recover information that the lens never captured in the first place. That is why measurement-grade systems start with the optics, not the algorithm.

How a Telecentric Lens Works

The chief ray of an object point is the ray that passes through the center of the aperture stop. In a telecentric design, the aperture stop is positioned at the focal plane of the lens group in front of it. Any ray aimed at the center of that stop is refracted so that it emerges parallel to the optical axis. Once all chief rays are parallel, two things happen:

• Magnification is locked. Moving the object along the Z axis may blur the image, but it does not change the image size. Focus errors no longer become measurement errors.
• Viewing angle is fixed. Every feature in the field of view is seen from the same direction — the lens behaves like an orthographic measuring probe rather than a camera looking at a scene in perspective.

Which side of the lens is telecentric depends on where the stop sits, and that leads directly to the three telecentric types.

The Three Types: Object-Space, Image-Space, and Bi-Telecentric

For most factory-floor measurement tasks, an object-space telecentric design is sufficient and more affordable. A bi-telecentric design is the choice when the application also needs coaxial (through-the-lens) illumination, or when the highest measurement accuracy is required. POMEAS covers both families: the LTC15-110 1.5× object-space telecentric lens for compact measurement stations, the 1.0× bi-telecentric lens with 110 mm working distance for 1.1-inch-sensor metrology, and the LTCM2-110C 2× coaxial telecentric lens for reflective-surface inspection.

Bi-Telecentric vs Object-Side Telecentric: What Changes at 1X

Direct answer: at 1X both designs keep the scale factor constant when the part moves toward or away from the lens. The difference is which side of the lens is corrected. An object-side telecentric lens makes the chief rays parallel in object space only, so a part at any height inside the telecentric range measures the same. A bi-telecentric (double telecentric) lens also makes them parallel in image space, so the scale survives the image plane moving as well — a sensor that is re-focused after a filter change, a camera that is not perfectly square to the axis, or focus drift over a long production run.

What the second telecentric group buys

Image-space telecentricity is the less obvious half of the design. With the aperture stop at the common focal point of the front and rear groups, chief rays reach the sensor parallel to the axis across the whole image circle. Three things follow from that. Refocusing no longer changes magnification, so a filter swapped between products does not silently rescale every measurement taken after it. Illumination falls off less toward the corners, which starts to matter once the sensor is 1.1″ and the F-number is short. And the barrel grows: the 1X F-mount double telecentric PMS-LTFM1-127 is 297.35 mm long to reach a 127 mm working distance, and the extra rear elements are why a bi-telecentric lens costs more than an object-side design of the same magnification.

Which errors each design actually removes

Read the table by row, not by column. The first row is the one most people are buying, and both designs deliver it. The remaining rows are where the two part company: they are all image-side effects, and a purely object-side lens cannot correct them because it has no control over where the chief rays land on the sensor. If your measurement loop never touches the camera after calibration, rows two to four cost you nothing. If it does — a filter wheel, a camera swap, a hot cabinet — they are the reason to pay for the second group.

What to specify at 1.0X in the current POMEAS range

Two readings of this table are worth stating explicitly. First, a 1.1″ sensor does not itself require bi-telecentricity: PMS-10VP110C111T covers the same 1.1″ format with object-side correction only (4.44 μm, NA 0.031, F# 7), so sensor format and telecentricity are separate decisions that are often confused in a quotation. Second, mount and correction are independent as well — the F-mount route is bi-telecentric at 1X, while the C-mount route offers both an object-side design with a built-in coaxial port and a bi-telecentric design without one.

For the three product routes, see the 1.0X bi-telecentric lens for 1.1-inch sensors, the F-mount 127 mm double telecentric lens, and the 1.0X coaxial object-side lens LTCM1-110C. Each page publishes its own specification table; the comparison above only lines up the fields that differ between the designs.

The Parameters That Actually Matter

Telecentric lens datasheets carry more numerically defined specifications than FA lens datasheets. These are the ones that decide whether a lens will measure accurately on your line:

• Magnification. Typically 0.2× to 2×. The field of view equals the sensor size divided by the magnification, so magnification and sensor format must be chosen together.
• Distortion. The geometric error that bends straight features. High-quality telecentric lenses hold distortion below 0.1%, and precision designs reach 0.05% or lower — one to two orders of magnitude better than a typical FA lens.
• Telecentricity. How close the chief rays are to truly parallel, usually specified as a maximum angle such as 0.1°. Low telecentricity is what keeps the image size stable when parts move in Z or sit at slightly different heights.
• Depth of field. Roughly 2–3× deeper than an FA lens of equal resolution. It is set by the F-number, the wavelength, and the allowable blur circle — stopping down trades resolution for depth.
• Working distance. The distance from the front of the lens to the object. Long working distances keep the optics away from moving machinery, at the cost of a larger front barrel.
• Sensor compatibility and mount. Telecentric lenses are designed for a maximum sensor format (for example 2/3″ or 1.1″) and usually ship in C-mount. Using a sensor larger than the design format puts the uncorrected corners of the sensor outside the optical field.

Telecentric Lens vs. Standard FA Lens

The physical reason behind the larger barrel is simple: to keep chief rays parallel over a given field of view, the front group must be at least as large as the field itself. A 40 mm field of view needs a front lens of roughly 40 mm — which is also why telecentric lenses become impractical for very large objects.

When to Use a Telecentric Lens — and When Not To

Where each application lands

• Telecentric Lens for Lead-Frame Coplanarity
• Telecentric Lens for Machined Parts
• Telecentric Lens for Spring Measurement
• Telecentric Lens for Shaft and Sleeve Outer Diameter

Choose telecentric when:

• you measure dimensions, positions, or angles and the result feeds an accept/reject decision;
• parts can sit at slightly different heights in the fixture, and the measurement must not change with them;
• you inspect deep features — bores, threads, connector pins — where side walls must not be visible;
• defect decisions depend on true defect size thresholds, not just defect presence;
• the same station must measure several part variants without recalibration.

Save the budget and use an FA lens when:

• the task is presence/absence detection, text or code reading, or robot guidance;
• the field of view is large relative to realistic lens diameters;
• the required accuracy is looser than the perspective error of a corrected FA lens plus software calibration;
• space or cost constraints dominate the project.

How to Select a Telecentric Lens in Five Steps

1. Fix the sensor and mount. Sensor format and C-mount come first, because every optical parameter is specified for a given format.
2. Compute magnification from the field of view. Magnification = sensor size ÷ field of view. Leave 10–20% margin for part placement tolerance.
3. Set the working distance. Include fixture clearance and any protective window between lens and part.
4. Match resolution to the sensor. The lens must resolve at least as many line pairs per millimeter as the sensor's Nyquist limit; otherwise the pixels, not the optics, will limit accuracy.
5. Buy accuracy where it counts. Convert the part tolerance into an allowable distortion and telecentricity budget, then pick the lens grade that satisfies it with margin — and pair the lens with a collimated (telecentric) backlight, because edge imaging quality depends on the illumination as much as on the lens.

For a worked example of this process, see the telecentric lens selection guide, or browse the full POMEAS telecentric lens family for magnifications from 0.2× to 2×.

Telecentric Lens Supplier Checklist: What to Verify Before You Shortlist

Most telecentric lens projects end with a shortlist of two or three suppliers and a request for quotation. The commercial side of that comparison is easy; the technical side is where projects fail, because two lenses can share a magnification and working distance and still differ by a factor of two or more in resolution and by an order of magnitude in distortion. Everything below can be checked from published data, before a sample is shipped.

1. One parameter row per SKU, not a family range

A catalogue that gives a single resolution range for an entire family cannot be used in a tolerance stack-up. Ask for the row that belongs to the part number you are buying, and check that the row states magnification, working distance, maximum sensor size, resolution, distortion and depth of field for that magnification. In the current POMEAS range each SKU publishes its own row; PMS-10110CVP111T, for example, publishes 5.37–13.42 µm resolution across F/8–20 rather than one number.

2. Resolution quoted at the aperture you will actually run

Diffraction widens the spot as the lens is stopped down, so a single resolution figure is only meaningful together with its F-number; the 1.0X lens above spans a 2.5× change across its published aperture range. Compare that resolution with the sensor pixel pitch: if the optics resolve 13.42 µm and the camera pitch is 3.45 µm, the lens — not the camera — sets the limit, and at 1.0X that 13.42 µm lands on the part unchanged.

3. Distortion and telecentricity stated with their conditions

Distortion claims are comparable only at a stated image height and sensor format: 0.01% at 1″ and 0.1% at 1.1″ describe different working points, not a better and a worse supplier. The same applies to telecentricity — the object-side figure in degrees is what governs how much magnification drifts when the part moves in depth. If a datasheet omits it, request it in writing; an unstated value is not a zero.

4. Depth of field, converted into fixture tolerance

Depth of field is the parameter that quietly rewrites the fixture design. Across the published 0.2X–2.0X range it falls from 12 mm (adjustable, 0.2X) to 0.27 mm (2.0X). Read as a total window, the 2.0X figure means the part surface must be presented within roughly ±0.135 mm of the nominal plane — arithmetic from the published value, not a separate specification — and that is a mechanical requirement on the stage, not an optical one.

5. Sensor coverage and mount, checked against your camera

A lens designed for a 1.1″ sensor on a 2/3″ camera pays for an image circle it never uses; the reverse combination vignettes and softens the corners. Match the published maximum sensor size to the camera that will actually be mounted, and confirm the mount — the range below is C-mount throughout, which also caps the practical sensor diagonal and therefore the field of view at low magnification.

6. Illumination path and mechanical envelope

Built-in coaxial illumination removes a beam-splitter from the optical train but adds body length, and body length is what collides with the machine frame. Ask for overall length and clearances with the optical data: PMS-10110CVP111T publishes a 132.88 mm body length, which is the number a fixture designer needs before the quotation is approved.

7. Documented test method and edition control

Two suppliers can both be accurate and still measure differently. Ask which test method backs the numbers — the POMEAS imaging-chain test methods are published in the technical reference — and which datasheet edition the quotation refers to, so that a later revision is a change you are notified about rather than one you discover on the line.

Published figures across the 0.2X–2.0X range

The table transcribes the published parameter row of each SKU (retrieved 2026-11-10). It is a starting point for a shortlist, not a substitute for the datasheet edition that ships with the lens.

Two patterns matter before the shortlist is fixed. Resolution tightens sharply at the top of the range — 20.14 µm at 0.2X against 4.53 µm at 2.0X — while the depth of field collapses from 12 mm to 0.27 mm, so magnification is chosen by the tolerance budget, not by field of view alone. And the 110 mm working distance is shared from 0.5X to 2.0X, which lets one fixture serve several part variants without re-machining the mount.

Where POMEAS sits in this comparison: the C-mount range shown here spans 0.2X to 2.0X with the parameter row published per SKU, and the full list is on the telecentric lens category page. If you are evaluating one magnification, the useful exchange with any supplier is a SKU plus its published row — sensor format, field of view, working distance, tolerance band — rather than a family brochure; the selection guide shows the arithmetic behind that row.

Frequently Asked Questions

1. What does “telecentric” actually mean in a lens?

It means the chief rays — the rays passing through the center of the aperture stop — are parallel to the optical axis on one or both sides of the lens. Parallel chief rays in object space make magnification independent of object distance, which is the property measurement systems need.

2. Which type should I choose: object-space, image-space, or bi-telecentric?

For dimensional measurement without coaxial light, object-space telecentric is the cost-effective choice. If the system uses coaxial illumination or a beam splitter, or if uniform brightness across a large sensor is critical, choose a bi-telecentric design. Pure image-space telecentricity mainly benefits color sensors and beam-splitter optics and is rarely bought as a standalone feature.

3. Why are telecentric lenses larger and more expensive than FA lenses?

The front lens group must be at least as large as the field of view, and the design needs multiple precision lens groups to hold distortion and telecentricity at the 0.1% / 0.1° level. More glass, larger elements, and tighter tolerances all raise cost.

4. What measurement accuracy can I expect?

Accuracy depends on the whole chain — lens distortion, telecentricity, sensor resolution, calibration quality, and illumination. As a rule of thumb, a well-specified telecentric lens with proper calibration supports measurements at the 1–10 μm level on suitable fields of view; the exact budget should be calculated from the datasheet values for your magnification.

5. Will a telecentric lens work with my camera?

Check three things: the lens is designed for your sensor format or smaller, the mount matches (C-mount in most cases), and the lens resolution meets your sensor's pixel pitch. A lens designed for 2/3″ should not be paired with a 1.1″ sensor even if it screws on.

6. When should I avoid a telecentric lens?

Avoid it for very large fields of view (the front barrel grows with the field), for simple presence/absence or code-reading tasks, and for cost-sensitive projects where a calibrated FA lens plus software meets the accuracy target. Telecentric optics solve measurement problems; they are not a general upgrade for every camera.

Does a 1X telecentric lens have to be bi-telecentric for accurate measurement?

No. At 1X the scale factor is already protected by object-side telecentricity, so a camera that is focused once and left alone is served by an object-side design such as PMS-LTCM1-110C. Bi-telecentricity earns its extra cost when the image plane can move after calibration - a filter changed in front of the camera, a camera swapped, or thermal focus drift on a long run - and when a large sensor at a short F-number needs brightness that holds to the corners. In the current range that is PMS-10110CVP111T at 1.1 inch and PMS-LTFM1-127 at F-mount, both published as double telecentric.

Key takeaway: a telecentric lens buys you one thing ordinary optics cannot give — an image whose size does not depend on where the part is. If your inspection decision depends on absolute dimensions, that single property usually pays for the lens many times over in avoided false rejects and recalibration downtime.

Further reading: the test methods behind this article are documented in POMEAS Technical Reference when a standard lens is sufficient and when a telecentric lens is required (§§10.3 选型判据).

Related reading

• FPC Ribbon 10 µm Defect Inspection: Single-Field Coverage at 1.0X