Crankshaft Surface Inspection: Journals, Pins And 45-Second Cycles

Overview

A crankshaft is a series of polished cylinders arranged on two axes, with counterweights between them and oil holes drilled through the journals. Every functional surface is curved, ground to a fine finish and highly reflective, and the surfaces that matter most are often the ones that face inward towards the axis.

That combination makes it one of the harder parts to inspect optically. A polished cylinder returns a narrow band of specular reflection that moves as the viewing angle changes, and a defect inside that band is invisible because the surface is already at maximum brightness.

The method that works on this geometry is lighting rather than resolution. A tunnel light wraps the journal in a diffuse band so that the specular reflection is spread rather than concentrated, which lets a scratch or a pore appear as a local change in an otherwise even field. Big Bird Industrial (formerly Shimada Big Bird Industrial) has been building cleaning and deburring systems for automotive powertrain components since 2001, and the inspection figures here come from production crankshaft stations running at cycles of forty-five seconds or less.

1. Why A Crankshaft Is Hard To Light

The difficulty is geometric, and it appears at three different scales on the same part.

The Specular Band

A polished cylinder reflects a light source into a narrow angular range. From any fixed camera position, only a strip of the journal is correctly illuminated and the rest is either too dark or blown out. Rotating the part moves the band, which is why inspection of a journal usually involves either rotation or a light that wraps the full circumference.

Concave And Occluded Regions

The inner surface of a connecting rod journal faces the axis of the part, and the fillets between a journal and a counterweight are concave. Both present a surface that is partly hidden from every external position, and both are exactly where grinding damage tends to concentrate.

2. Tunnel Lighting

A tunnel light is a curved diffuse source arranged around the part so that light arrives from a continuous range of angles rather than from one direction.

Spreading The Reflection

With light arriving from every angle around the circumference, each point on the journal reflects some of it towards the camera, and the resulting image is an even field rather than a bright band. A defect then appears as a local deviation: a scratch scatters light differently, a pore absorbs it, a dent changes the local slope.

Where It Stops Working

A tunnel light cannot illuminate a surface that is geometrically hidden. A fillet under a counterweight, or a journal section inside a concave region, needs either a different view or a different light. On a crankshaft that means the inspection is a sequence of views rather than one image, and the sequence is what consumes the cycle.

3. What Is Inspected Where

The defect classes and the acceptance criteria differ by region, and the station has to treat them separately.

Journals And Pins

Main and connecting rod journals carry the bearing surfaces. The defects that matter are scratches, score marks, grinding burns, pores and inclusions, and the criterion is usually expressed as a maximum defect size rather than as a count. The oil holes drilled through the journal introduce an edge that must not be confused with a defect, which is one of the classification problems on this part.

Counterweights And Balance Holes

Counterweights are not bearing surfaces but they are visible, and their edges carry handling damage. The balance holes drilled through them are inspected for breakthrough on the side wall, which is a geometry question rather than an appearance one and needs a view that can see into the hole at an angle.

4. Handling And Rotation

A crankshaft is long relative to its diameter and is normally supported at the ends and rotated between centres. That arrangement gives access to the full circumference of each journal in a single rotation and keeps the part's own axis as the reference for inspection.

Rotation Speed And Image Blur

Rotating the part requires either a short exposure per image or a line-scan arrangement. A short exposure needs more light, which conflicts with the diffuse approach that makes the inspection work. The practical solution is to index the rotation and expose while stationary for the regions where resolution matters, and to scan continuously for the less critical surfaces.

Why The Part Axis Matters

Supporting on the centres means the measurement frame is the part's own axis, which is also how the drawing is dimensioned. That removes the datum reconstruction problem that fixtureless inspection of a pressed part has, and it is why crankshaft inspection leans on mechanics rather than on software.

5. The Forty-Five Second Budget

A production crankshaft station runs at forty-five seconds or less, with a miss rate around one part in a thousand and a false-call rate in the region of five to seven percent. Those two figures move together: tightening the miss rate raises the false calls on a surface this variable.

The cycle is dominated by the number of views rather than by the area scanned. A shaft with four main journals, four pins and a set of balance holes needs a view per functional region, and each view has to be positioned, exposed and processed. Where the cycle must be shorter, the answer is usually more cameras working in parallel rather than a faster processor.

6. Verification

Verification on a crankshaft uses parts carrying known defects of known size, presented at different rotational positions so that a defect that is only visible at one angle is caught by the test as well as by production.

The other half of verification is the false-call rate, measured on conforming parts from normal production rather than on the defect panels. A station that finds every reference defect and flags one part in five is not usable on a line, and that outcome is invisible if the test set contains only defective parts.

7. Resolution Is Not The Lever

A common instinct on a difficult part is to buy more pixels. On a crankshaft that instinct is usually wrong, because the limit is contrast rather than sampling.

What The Numbers Say

A journal of 60 mm diameter photographed across a 200 mm field of view with a sensor of a few thousand pixels resolves roughly a tenth of a millimetre per pixel. A scratch of 0.2 mm width therefore covers two pixels, which is detectable if the contrast is there and invisible if it is not. Doubling the resolution gives four pixels on the same scratch and does not create the contrast that the lighting failed to provide.

Where Extra Resolution Does Help

Resolution pays on the small features at the edges of the inspection: the oil hole breakout, the fillet radius, the balance hole side wall. Those are the regions where a defect is measured rather than judged, and where the camera is looking at a small area anyway.

The practical split is a wrapped diffuse light and moderate resolution for the journals, and a tighter field of view with higher resolution for the edges. That combination costs less than a uniformly high-resolution system and performs better on the part.

Frequently Asked Questions

Why Not Use A Standard Area Light?

Because a polished cylinder under a single source produces a bright band and dark regions, and a defect inside the band cannot be seen. A wrapping diffuse source removes the band at the cost of contrast, which is why the detection then depends on the algorithm.

Can A Crankshaft Be Inspected Without Rotating It?

Only the surfaces facing the cameras. Covering the full circumference of each journal without rotation would need a ring of cameras per journal, which is usually more expensive than rotating the part.

What False-Call Rate Is Acceptable?

Five to seven percent is typical in production on this surface. Lower is possible with a narrower defect definition, but the reduction usually comes from relaxing what counts as a defect rather than from better imaging.

How Are Oil Hole Edges Handled?

As a separate class with its own view. Treating an oil hole edge as part of the journal image produces persistent false calls, because the edge is a legitimate feature that looks like a defect.

Conclusion

A crankshaft is inspected by controlling light on a curved polished surface, not by adding pixels. Wrapped diffuse illumination makes the journal an even field in which a defect appears as a local change, and rotation provides the coverage that one camera cannot.

The forty-five second cycle is a count of views rather than a measure of speed. A specification that states the defect classes per region, the miss rate and the false-call rate describes a station that can be verified; one that states a resolution does not.