Inline Measurement of Transparent and Reflective Parts: What Actually Works

Transmitted-light gauging is elegant and accurate when the product blocks light cleanly: the shadow has a hard edge and diameter falls out of geometry. Clear tube, transparent film, polished metal and mirror-finish parts break that assumption in two different ways, and the failure looks like a measurement problem when it is really an optics problem. This page sets out what physically goes wrong, and which principles handle each case.

Why Transmitted Light Needs an Opaque Object

The method projects collimated light past the product and reads the shadow on a receiver. Every micrometre of the measurement depends on where the shadow edge lands. That edge is crisp when the product absorbs or blocks the light. In a transmitted-light gauge such as the LM-1030, or in the projection heads of the HM series, the instrument is built around that assumption.

Case 1: Transparent and Translucent Material

Clear material does two things that corrupt the shadow. It transmits most of the light rather than blocking it, so the shadow is faint; and it refracts at the walls, so the light that does reach the receiver has been bent, and the apparent edge is displaced inward or outward depending on wall thickness and refractive index. The result is a low-contrast edge in the wrong place. Thin-walled transparent tube and clear conduit are the worst cases. Two practical points:

• Some translucent materials that look clear to the eye are measurably opaque at certain wavelengths. Where a material transmits visible light strongly, a source at a different wavelength can restore contrast — this is a material-dependent question that has to be tested on the actual product, not assumed.
• Pigmented, filled or frosted versions of the same part often measure acceptably where the clear version does not. If the product has a colourant option, test both.

Case 2: Highly Reflective and Polished Surfaces

A mirror finish does not transmit, so the shadow exists — but the geometry fights you. Specular reflection off a curved polished surface throws light into the receiver from directions that have nothing to do with the silhouette, which lifts the apparent edge and can make a part measure undersize or oversize depending on where the glare lands. Grinding marks, oil film and coolant on a shiny part produce the same effect patchily along the length. The mitigation is lighting geometry: keep the source collimated and the receiver looking only at light that came straight down the axis, which is what the telecentric arrangement on both sides of the HM heads is for.

Which Principles Handle Which Case

The three principles side by side, including what each one cannot do, are set out in online gauge, confocal and vision compared.

What This Means for a Procurement Spec

Do not write "measurement accuracy ±2 µm" into a specification without stating the material. The LM-1030's published ±2 µm accuracy and ±0.15 µm repeatability are instrument figures; they assume a shadow the instrument can find. On clear or mirror-finish product, achievable accuracy is set by the optics of your material, and no instrument rating overrides that. Test on the real product, at the real line speed, before the number goes into a contract. How to assemble the full budget is in building an inline diameter measurement error budget.

Limits We Are Not Going to Paper Over

• We are not going to claim that a projection gauge measures clear tube to datasheet accuracy. It does not, and the material physics is the reason.
• Whether a specific translucent polymer can be measured by wavelength selection is material-specific and must be tested — it is not a specification anyone can quote in advance.
• Where a job genuinely needs confocal or interferometric measurement, the right move is to say so rather than to stretch a projection gauge into an application it will fail.

FAQ

Can a transmitted-light gauge measure clear plastic tube?

Not reliably. Transparent material transmits rather than blocks the light, and refraction at the wall displaces the apparent edge. The shadow is both faint and in the wrong place. Test your specific material, and consider confocal displacement sensing instead.

Does a mirror-finish metal part measure correctly?

The shadow exists, but specular glare off the curved surface can lift the apparent edge and bias the reading either way. Keep the illumination collimated and use telecentric optics on both sides to reject off-axis light.

Would a different light colour help?

Possibly. Some materials that transmit visible light strongly absorb at other wavelengths, which restores shadow contrast. It is material-dependent and has to be tested on the actual product — it cannot be quoted as a specification.

Does the datasheet accuracy apply to any material?

No. Published figures such as the LM-1030's ±2 µm accuracy assume a shadow the instrument can locate. Achievable accuracy on transparent or highly reflective product is governed by the material's optical behaviour first.

What should I send a supplier to get a straight answer?

A sample of the actual material in the actual finish, the diameter range, the tolerance, and the line speed. Material behaviour cannot be assessed from a drawing.

Is confocal measurement always the answer for clear parts?

It handles height, thickness and displacement on transparent material far better than projection does, but it is a different instrument with its own range and standoff constraints. Compare them in the principle-level guide linked above.

Next Step

Send us the material, finish, diameter range and tolerance — ideally with a short sample of the real product. We can tell you whether transmitted-light gauging will hold, whether it needs a lighting change, or whether the job belongs on a confocal route. Start with the method selection guide.