Cylinder Head Bore Inspection: Valve Seats, Oil Galleries And The 0.01 mm Resolution Problem

Overview

A cylinder head is inspected from the outside every day and still fails from the inside. The surfaces that decide whether an engine runs clean are the ones nobody sees: valve seat bores, guide bores, oil gallery drillings and the transition where a cross-drilling breaks into a main gallery.

Big Bird Industrial (formerly Shimada Big Bird Industrial) has built cleaning and inspection systems for cylinder heads since the early 2000s, and the recurring lesson is the same on every project: if the bore is not inspected, the cleaning process has no feedback. This article covers what has to be inspected inside a head, why a borescope cannot carry production volume, which optical configuration fits which bore, and how to write an acceptance criterion that means something.

1. What Has To Be Inspected Inside A Cylinder Head

The internal inspection scope is narrower than the part and much harder than the outside. Four features carry most of the risk.

Valve seat bores and the throat below them collect chips and lapping compound during seat machining. A chip pressed into the seat face becomes a compression leak that only shows up at hot test. Guide bores are deep, small and usually inspected by feel rather than by measurement. Oil gallery drillings carry debris from machining and from the deburring operation, and a single blocked drilling starves a cam bearing. Cross-drill intersections are the worst case: burrs form on the far wall where no tool reaches, and the geometry hides them from any straight-line view.

These four features have different diameters, different depths and different defect priorities, which is why a single camera on a single arm rarely covers a whole head.

2. Why A Borescope Cannot Carry Production Volume

A borescope is an excellent engineering tool and a poor production instrument. Three reasons.

The first is coverage. An operator decides where to look. Two operators on the same head will not inspect the same set of surfaces, and neither can prove afterwards what was covered.

The second is judgement. A scratch, a sand hole and a smear of residue look similar on a small screen. Without a reference and a trained eye, the call is subjective, and the recorded result is a word rather than an image.

The third is throughput. Manual insertion, focus and rotation take tens of seconds per bore, and a head has many bores. Sampling becomes the only economically possible strategy, and sampling cannot detect a process drift that affects one head in fifty.

3. Three Optical Configurations And Where Each Applies

Inline bore inspection is a family of optics, not one product. The choice follows from bore diameter, depth and the defect size that has to be detected.

Depth-to-diameter ratio is the number that decides feasibility. A probe that must reach a depth equal to its diameter is a different instrument from one that must reach three times its diameter. In practice the ratio, not the diameter, is what limits how deep an oil gallery can be inspected.

Lighting matters as much as the lens. A machined bore wall is specular; a cast gallery wall is diffuse. The first needs controlled illumination that does not blow out, the second needs enough light to reveal texture. A laser probe is attractive on the first because it is largely immune to ambient light and to LED interference, which is why it appears on gallery inspection where a strong reflection would otherwise hide a defect.

4. Defect Classes And What They Cost Downstream

The defects that matter inside a head are few and specific, and they do not share a single acceptance number because their consequences differ. Treating them as one class is the most common reason an inspection specification ends up either too strict to run or too loose to catch anything.

Chips and foreign objects

The most frequent finding and the easiest to understand. A chip on a valve seat face becomes a compression leak that only appears at hot test; a chip in a gallery blocks flow and starves a bearing. Detection threshold is usually a contour size in the 0.25 to 0.5 millimetre range, and the material is often the same colour as the wall, which is why lighting design matters more than magnification.

Burrs at cross-drill intersections

Burrs form on the far wall of an intersection, where no tool reaches and no straight-line view sees. They are a manufacturing defect rather than a cleaning defect, so improving the washer will not remove them. If the deburring process is not under control, the inspection system will simply report the same finding on every head until the upstream process changes.

Scratches and score marks on bore walls

These indicate handling damage or a tool problem, and their significance depends on position. A scratch that crosses a bearing surface is a functional defect; the same scratch on a non-functional gallery wall may be cosmetic. Position-aware classification is therefore as important as detection.

Sand holes and porosity

A casting problem exposed by machining. No cleaning or deburring change will fix it, and the scrap decision depends on wall thickness and on whether the defect intersects a pressure or oil path. Reporting these separately from loose debris lets the foundry conversation start with data instead of samples.

This is the argument for classifying defects rather than merely detecting them. A system that reports foreign object, burr, scratch, dent and porosity separately gives the process engineer something to act on. A system that reports pass or fail gives them a rejection rate and an argument.

5. Inspection Only Works When Cleaning Is Controlled

Bore inspection finds what cleaning leaves behind, so the two processes have to be designed together. A head that is cleaned to a residue limit of a few milligrams and then inspected will show a stable defect pattern. A head that is cleaned without oil gallery attention will produce a burst of findings at the first inspection and then look acceptable, because the gallery has already dumped its debris into the bath.

The cleaning side of the same problem is well understood: plug-sealing cleaning forces flow through the galleries instead of around them, and high-pressure targeted cleaning reaches the far wall of a cross-drilling. When cleaning and inspection sit on the same line, the inspection result becomes the cleaning process record, and a drift in residue shows up as a change in findings rather than as a customer complaint.

6. Writing The Acceptance Criteria

An inspection specification for a cylinder head has to answer four questions: what is inspected, at what resolution, with what miss rate, and in what time. Anything less and the system cannot be accepted or rejected objectively.

A realistic resolution target for gallery walls is a defect contour in the 0.25 to 0.5 millimetre range, with optical resolution an order of magnitude finer so that classification has something to work with. Miss rates below a tenth of a percent are achievable on a defined defect set, but the set has to be agreed before the acceptance test, not after it.

Frequently Asked Questions

Can one probe inspect both valve seats and oil galleries?

Rarely. Seat bores are short and wide, galleries are long and narrow, and the required depth-to-diameter ratio differs. Most heads are served by two configurations on one platform, with a tool change between them.

Is X-ray or CT a substitute?

They answer a different question. CT finds internal porosity and voids in the material; optical bore inspection finds surface conditions and loose debris inside passages. For assembly risk, the surface view is usually the direct one.

How does the system handle a blocked gallery?

A blocked passage has no visible wall to inspect, so it is normally detected by flow or pressure testing rather than by optics. Combining a flow check with optical inspection covers both the blockage and the residue case.

What happens to the images?

They are stored against the part serial so a finding can be reviewed later. That is the practical difference between inline vision and a borescope: the evidence exists after the shift ends.

Conclusion

Cylinder head quality is decided in the places that are hardest to see. Valve seat bores, guide bores and oil galleries hold the debris that causes compression leaks, oil starvation and cam bearing wear, and they are exactly the surfaces that manual inspection covers least reliably.

The practical sequence is to define the bores, choose the optics that reach them, agree the defect set and the miss rate before installation, and place the system downstream of a cleaning process that is itself under control. Done in that order, bore inspection stops being a quality gate and becomes process feedback.