Telecentric lens size: front optic diameter sets the field of view

You sized a telecentric by magnification and image circle, sent the requirement out, and the quote came back describing something the size of a soup can. That is not padding: in an object-side telecentric the front optic must be at least as wide as the field itself, and magnification decides how wide the field is. This guide gives you the arithmetic that predicts the size before you ask for a quote — sensor diagonal over magnification, the f-number multiplier at high magnification, and the vignetting margin that keeps the corners — plus the Fresnel-augmented pseudo-telecentric option for oversized fields and a pre-order checklist for mounts, brackets and weight.

Why does the front optic diameter set the field of view?

In an object-side telecentric lens the chief rays run parallel to the optical axis, so the front optic must accept the whole field side by side: a 100 mm field of view needs a clear front aperture of at least 100 mm, before any housing or mount hardware.

The stop sits at the front focal plane, and that placement is what makes every chief ray leave the lens parallel to the axis. Parallel rays at the edge of the field still have to pass through the front aperture, next to the rays from the center. The front of the lens behaves like a window exactly as wide as the field: narrow the window and the field goes with it.

The barrel adds housing wall, focus travel and mount threads around that glass, so the finished part is always bigger than the optical minimum. The telecentric field of view calculation from magnification works the same formula from the sensor side, and machine vision optics basics defines field of view, working distance and f-number if you want the ground floor first.

How large does the lens get at low magnification?

Field of view equals sensor size divided by magnification, so low magnification multiplies the front optic fast: at 0.166x a 1-inch sensor's 16 mm diagonal becomes a 96 mm field, and the front aperture has to match it.

Work the diagonal. A 1-inch sensor measures 12.8 × 9.6 mm, a 16.0 mm diagonal. At 0.166x that is 16.0 / 0.166 = 96 mm of field, so the front aperture must clear roughly 96 mm and the barrel runs past 100 mm. The LS-TCL-1422 is a 0.166x bi-telecentric on a 1-inch sensor with a 258 mm working distance; a stand-off that long is typical of the class, because the big front optic needs room to collimate the field.

Lower the magnification and the glass grows in a straight line. An APS-H sensor has a diagonal of about 34.5 mm; at 0.287x that is 34.5 / 0.287 = 120 mm of field. That 120 mm field is why low-magnification telecentrics in this class use an M58 thread rather than a C-mount. The LS-TCL-0885 is a 0.287x bi-telecentric on an APS-H sensor with an M58 mount. A C-mount thread is only 25.4 mm across, far too small for glass of this class.

The catalog shows the pattern at scale: 727 telecentrics sit in the 0.1-0.3x band and 282 more sit under 0.1x, so the telecentric lenses category is where the large glass lives. Under 0.1x on a 1-inch format the field passes 160 mm, and the front optic grows with it.

How does f-number change size at high magnification?

At high magnification the object-side numerical aperture is roughly the magnification times the image-side NA, so a 4x lens at f/6 collects an NA 0.33 cone at the part and needs front glass sized for it; slowing to f/16 cuts that cone to NA 0.125.

The relation is conservation of etendue, NA(object) ≈ NA(image) × magnification, and it is a paraxial estimate good to a few percent. At 4x, an image-side f/6 arrives as an f/1.5-class cone at the part, and the front optic has to pass that cone from every point across the field at once. Slow the same 4x design to f/16 and the cone narrows about 2.7 times, and the elements shrink with it. Two 4x telecentrics covering the same field can therefore have very different barrels: the speed sets the difference, not the field.

The cost of slowing is diffraction. The Airy disk diameter is 2.44 λ F/#, so at a 550 nm wavelength:

• f/6: 2.44 × 0.55 µm × 6 = 8.1 µm
• f/16: 2.44 × 0.55 µm × 16 = 21.5 µm

The slow lens fits in a smaller barrel but spreads each point 2.7 times wider. The Airy disk size and the diffraction limit article shows how to match that spot to the pixel, and aperture, f-number and depth of field covers the depth-of-field side of the same trade.

A slow cone is also one way a design stays compact at a long stand-off. The LS-FFL-1023 is specified at F13.4, holding 0.5x over a 150 mm working distance on a 1/2-inch format; the slow aperture is what keeps the barrel manageable there.

High magnification shrinks the field and the front optic together. The LS-TCL-2118 works at 1x on a 2/3-inch sensor with a 108 mm working distance. The LS-TCL-0980 runs 1.5x on a 1/3-inch sensor with a 65 mm working distance and a built-in coaxial port. At 1x the 2/3-inch field is 11 mm across; at 1.5x the 1/3-inch field is 4 mm. Front optics of that class fit in barrels a C-mount thread can carry.

Why is the usable field smaller than the geometry allows?

Nominal geometry is not the usable field: a barrel shoulder, filter holder or coaxial beamsplitter that clips the edge rays shrinks the image to a hard circle, and real designs have lost a nominal field down to an 11 mm circle this way.

The failure is called mechanical vignetting. Something inside the barrel — a shoulder, a retaining ring, a filter holder, the plate of a built-in coaxial port — sits inside the cone near the front, and rays from the field edge strike it first. Corners dim, then a hard circular boundary appears, and the field beyond it is simply gone. Unlike soft optical falloff it does not fade gradually, so it can pass a quick on-axis look and fail at the field edge.

The risk concentrates where margin is thinnest: large fields, where the front aperture already runs close to the largest housing a mount can carry. Size the clear aperture at least 10 percent above the field diagonal, and treat anything tighter as a clipping risk until a flat white target wider than the nominal field has imaged with its corners at full brightness.

When does a Fresnel-augmented pseudo-telecentric make sense?

For fields around 200 mm on a tight budget, a fixed focal length lens with a Fresnel lens in front behaves pseudo-telecentrically; the build gives up image quality, working-distance freedom and color correction to get the field without the glass.

The construction is simple in principle: mount a large Fresnel lens in front of a fixed focal length lens and space the pair so the chief rays reach the sensor roughly in parallel. A 25 mm fixed focal length lens with a Fresnel in front has reached an 8 inches (about 203 mm) telecentric field, at a size a designed telecentric of the same field would not match. The base lens is an ordinary fixed focal length model from the machine vision lenses range.

Three prices are not negotiable. Image quality lands below a designed telecentric of the same field, because one grooved element does no aberration correction. Chromatic aberration is substantial, so measured edge position shifts when the LED wavelength changes. And the working distance barely moves, because the spacing that straightens the chief rays is fixed. The telecentric distortion and telecentricity explained article shows the figures a designed lens holds; the Fresnel build will not approach them.

Decide by tolerance. Presence checks, coarse positioning and sorting across 150-250 mm fields on a constrained budget: the pseudo-telecentric is the only telecentric-class option that fits. Gauging with micron-class tolerances: it is not. Buy the designed low-magnification telecentric and accept the barrel, or take the calibrated-entocentric route below.

What should you check before ordering a large telecentric?

Check three numbers before ordering: field diagonal (sensor diagonal divided by magnification), front clear aperture at least 10 percent larger than that field, and barrel diameter the mount can actually carry — anything past a C-mount's 25.4 mm thread needs a bracket or an M58-class mount.

Then walk the rest of the list.

• Field diagonal. Sensor diagonal ÷ magnification, written down before the catalog opens; every later check hangs on it.
• Aperture margin. Clear aperture at least 1.1 times the field diagonal — the vignetting margin of the section above.
• Mount hardware. A C-mount thread is 25.4 mm across; barrels past roughly 40 mm need M58 or M72 hardware — 231 of the catalog's telecentrics use M58, 85 use M72 — carried on a bracket, not cantilevered off the camera.
• Image circle. The stated circle must cover the sensor diagonal; telecentric lens selection: match magnification and image circle walks that pairing.
• Working distance. Low-magnification designs pair big fields with long stand-offs — 1,142 of the catalog's telecentrics work at 200 mm or more — so give the station the clearance.
• Lighting path. A built-in coaxial port saves an external beamsplitter on specular parts; in-line illumination in machine vision covers when that matters, and the port's plates still count as vignetting suspects.
• Configurator pass. The lens configurator steps through sensor, field and mount, and the lens finder filters by magnification band and working distance.

When should you step away from telecentric optics?

When the field is large and the tolerance is loose, an entocentric lens plus calibration beats a telecentric: a 100 mm-class telecentric field costs front glass, weight and budget that a perspective-corrected 16 mm-class lens does not.

The focal length comes from the standard relation:

A 100 mm field on a 2/3-inch sensor at 200 mm working distance needs f = 200 mm × (8.8 mm / 100 mm) = 17.6 mm, a standard 16-25 mm lens. Calibration absorbs the distortion and the perspective, but only over the height range you calibrated: parts that move in Z bring the scale error back. Entocentric vs telecentric lenses: when to use each puts numbers on that trade, and how to choose a machine vision lens focal length walks the pick.

So the decision, case by case. Fields of a few tens of millimetres: buy the telecentric; it is compact and the measurement is trustworthy. Fields of 30-100 mm with real tolerances: a designed telecentric in the 0.1-0.5x range, sized with the diagonal rule above. Fields past 100 mm on a budget: the Fresnel pseudo-telecentric for presence and sorting, the calibrated entocentric where perspective error is acceptable. A designed telecentric beyond 100 mm of field is for jobs that genuinely need micron-class, perspective-free measurement across the whole field — order it with the bracket it deserves.

If the barrel is about to become a mechanical problem, send the station layout through the contact page before ordering; mounts and brackets are cheaper to solve on paper. The lenses topic page groups the rest of the lens guides.