Telecentric Lens for Machined Parts: Contour, Length and Spacing

Turned parts, sleeves, spacers and small milled brackets are the highest-volume work for telecentric measurement, and also the most forgiving — until a feature has depth. This page covers what a single silhouette genuinely resolves on a machined part, where the magnification should be set from, and which features quietly leave the measurable set.

The three measurements that work well

Contour (outer profile), overall length or width, and hole centre-to-centre spacing are all lateral measurements on a part that can be presented flat to the sensor. They are what a telecentric silhouette does best, because each one is a distance between two edges in the same image plane.

The reason telecentricity matters here is not magnification stability for its own sake — it is that machined parts arrive with real thickness variation. A shaft turned to ±0.02 mm on diameter presents its top surface at a different object distance than the nominal. With a standard lens that changes the magnification and therefore the measured diameter. With object-side telecentricity it does not.

Setting magnification from the envelope, not the feature

The field of view has to contain the whole part plus fixturing margin. A 40 mm bracket on a 2/3″ sensor needs about 11 / (40 × 1.25) ≈ 0.22X. If you then need to resolve a 0.05 mm chamfer on that bracket, the pixel-limited resolution at 0.22X with a 3.45 μm pixel is roughly 15.7 μm per pixel — the chamfer spans about three pixels. That is measurable with sub-pixel edge extraction, but it is not comfortable, and it tells you the envelope and the tolerance are in conflict.

When they conflict, the answer is usually two stations or a move, not a better lens. Use the FOV and magnification calculation to make that conflict explicit before specifying hardware.

Telecentric models covering the envelope of typical machined parts

Magnification from 0.3X to 1.0X covers parts roughly 11 mm to 130 mm across on 2/3″ sensors, which is the bulk of turned and small milled work. Ordered here by working distance, because on machined parts the fixture and the part height usually constrain WD before resolution does.

Hole spacing and the datum problem

Hole centre-to-centre is the most robust measurement on a machined part, because a circle centre is far less sensitive to edge noise than a single edge position. Hole diameter measured in silhouette is less robust than it looks: it measures the smallest opening through the full thickness, so a tapered or belled hole reads smaller than a contact gauge at the surface.

If the drawing gives the diameter at a specific depth, a silhouette cannot reach it. That is a drawing-interpretation question, not an optics question, and it should be settled before the station is built.

What leaves the measurable set

• Counterbores and stepped bores. A silhouette shows the narrowest section. Depth-specific diameters are out.
• Thread geometry. Major diameter can be estimated from a silhouette; pitch diameter and flank angle cannot. Those need a different instrument.
• Burrs at the edge. A silhouette shows a burr as extra width, not as height. If the spec is burr height, the measurement is the wrong one.
• Surface roughness and flatness. Out of scope entirely for silhouette measurement.
• Deep features occluded by the part body. Common on milled brackets with pockets.

See when a telecentric lens is the wrong choice for the broader version of this list.

Fixturing is where the accuracy actually comes from

On machined parts the fixturing error usually exceeds the optical error. A part resting on a datum face that has a chip under it is tilted, and tilt inside the depth of field is invisible in the image but real in the measurement. Two practical rules: present the measurement plane closest to the lens, and make the fixture locate on the same datums the drawing uses.

Lighting for contour work

Backlight gives the silhouette. For parts with reflective edges, a collimated backlight reduces the halo that otherwise makes the extracted edge sit inside the true edge. Where the outer contour and a surface feature both matter, a coaxial front light can be added — but not simultaneously with backlight, because the silhouette contrast collapses. The lighting guide covers the selection in detail.

Questions engineers ask about this setup

What size of machined part can one telecentric lens cover?

One lens covers one field of view, so a single station handles a limited size band. In practice 0.3X to 1.0X covers roughly 11 mm to 130 mm across on a 2/3″ sensor. Parts spanning a wider range than that need either a lens change or two stations.

Can I measure hole diameter in a silhouette?

You can measure the smallest opening through the full thickness. That equals the surface diameter only if the hole is not tapered or belled. If the drawing specifies the diameter at a particular depth, a silhouette cannot reach it.

Why does my diameter reading differ from a micrometer?

Usually because the silhouette measures the narrowest section through the whole part while a micrometer contacts two points on the surface — taper, bellmouth or a burr makes these differ. Check whether the discrepancy is constant across parts; a constant offset is geometry, a varying one is fixturing.

Does telecentricity help with reflective metal edges?

Not directly. Reflection at the edge is a lighting problem: a collimated backlight reduces the halo that pulls the extracted edge inward. Telecentricity handles magnification stability, not reflection.

How do I know if my tolerance is achievable?

Divide the sensor pixel size by the magnification to get the pixel-limited resolution, then check how many pixels the tolerance spans. Under about three pixels the measurement depends heavily on edge-extraction quality and lighting stability, and should be validated on real parts rather than assumed.

Can thread be inspected with a telecentric lens?

Major diameter can be estimated from a silhouette. Pitch diameter, flank angle and thread depth cannot — they are not visible in a single projection. Those features need a different instrument.

Related reading

• Telecentric lens selection guide
• Magnification and FOV calculation
• Lighting for telecentric measurement
• What a telecentric lens is and when to use one
• When a telecentric lens is the wrong choice
• Battery electrode and tab
• Spring: wire, pitch and free length
• Gear pitch and bearing dimensions

Send one part and we will tell you what the measurement can hold

Specification arithmetic gets you to a shortlist, not to a verdict. Send one part with the drawing and the tolerance you have to hold, and we will run it and report what the system returns — including the features that come back as unmeasurable. A written feasibility note is worth more than another datasheet.