Surface Testing

Part of Precision Measurement

Methods for verifying the flatness, straightness, and texture of precision surfaces without specialized optical equipment.

Why This Matters

Making a flat surface is one thing. Knowing that you have made it flat — and how flat — is another. Every precision operation depends on accurate assessment of surface quality. Machine ways that are thought to be flat but are actually bowed by 0.1 mm will cause measurement errors, poor fits, and accelerated wear in every machine built on them.

Surface testing is also about knowing when a surface is good enough. Spending three extra days scraping a surface plate from 0.005 mm flat to 0.002 mm flat may be worthwhile for a national standards laboratory and wasteful for a workshop that needs 0.05 mm accuracy. Understanding the tests lets you judge “good enough” rationally rather than guessing.

In a rebuilding workshop, surface testing methods are divided into what you can do with simple tools and what requires optical instruments. Both have value.

Contact Methods

Bluing and Transfer

Engineer’s blue (marking blue) is the universal method for testing surface contact:

Setup:

1. Apply a very thin, even film of blue to the reference surface (surface plate or known flat)
2. Place the test surface on the blue, apply gentle pressure, lift
3. The blue transfers to all points where the test surface made contact
4. High spots show dense blue; low spots receive no transfer; intermediate spots show thin transfer

Reading the pattern:

Pattern	Interpretation
Blue across entire surface	Surface is flat and in good contact
Blue only in center	Surface is convex — high in middle
Blue only at edges	Surface is concave — low in middle
Blue in diagonal band	Surface is twisted
Blue at one end only	Surface is tilted
Irregular spots	Local high spots, general low areas

For good workshop flatness: You want blue showing contact at 60–80% of the surface, distributed evenly.

For precision work: 85–95% contact, with no single gap larger than about 5 mm.

The Feeler Gauge Test

A feeler gauge (thickness gauge) slid between a straightedge and a surface reveals the size of any gap:

1. Lay a precision straightedge across the surface
2. Try to slide a 0.05 mm feeler under the straightedge at multiple points
3. If it passes freely, the surface deviates by more than 0.05 mm at that point
4. Work down through 0.025, 0.013, 0.008 mm to quantify the error

This is less sensitive than optical methods but requires no special equipment. Useful for checking machine ways, large surfaces, and rough plates.

Light Gap Methods

Direct Observation Against Light

Two surfaces held close together against a bright background reveal gaps as hairlines of light:

1. Clean both surfaces thoroughly
2. Hold one on the other against the light source
3. Look obliquely along the contact line
4. Light leaks through gaps of 0.002 mm or more

This method detects errors too small for feeler gauges. It requires good lighting (bright daylight or a lamp behind a gap) and a practiced eye.

Limit of detection: A gap of 0.003 mm (3 micrometers) is easily visible. A gap of 0.001 mm may be detectable by an experienced eye. Gaps smaller than this cannot be detected by light.

Interference Bands (Optical Flat Method)

When you have an optical flat (a piece of polished plate glass ground perfectly flat), even more precision is possible:

1. Clean both the flat and the work surface
2. Lay the optical flat on the work surface without pressing — it rests on the high spots
3. In monochromatic light, alternating dark and bright bands (interference fringes) appear
4. Each band represents a height difference of half the light wavelength (~0.3 micrometers)
5. Straight, parallel bands = flat surface. Curved bands = curvature. Wavy bands = irregular

This method detects errors at the submicrometer level but requires monochromatic light (a sodium lamp, or sunlight through a yellow filter). It is practical even without electronic equipment.

Making an Optical Flat

Plate glass polished on both sides to optical quality is sufficient for many surface testing tasks. The glass must be thick enough not to flex. 20–25 mm thick borosilicate glass works. Polish using progressively finer abrasive on a flat cast iron lap, ending with cerium oxide or jeweler’s rouge.

Testing for Straightness

For edges and ways, not full surfaces:

The stretched wire test:

1. Stretch a piano wire or fine steel wire tightly between two end supports of equal height
2. Use a microscope eyepiece or magnifier to compare the wire position to the surface being tested
3. The wire defines a straight line; deviations of the surface from the wire are the errors

Accuracy: Detects errors down to about 0.01 mm over 1 meter.

The auto-collimator substitute: A trough of water provides a perfect horizontal reference for testing whether a surface is level. Lay the surface near the water and compare with a height gauge measuring from both water surface and test surface.

Surface Roughness Assessment

Precision surfaces should be smooth as well as flat. Surface roughness is measured in Ra (average roughness, in micrometers):

Ra Value	Description	Method to Achieve
25 µm	Very rough	Rough machining
3.2 µm	Machined	Turning or milling
0.8 µm	Finished	Fine turning, grinding
0.2 µm	Ground	Surface grinding
0.025 µm	Lapped	Lapping
0.006 µm	Superfinished	Precision lapping

Comparison method without instruments:

1. Make a set of surface roughness comparison specimens from your own machining
2. Measure their roughness by tactile comparison (fingernail dragged across)
3. Calibrate your specimens against a known standard if one is available
4. Use these specimens to assess unknown surfaces by touch comparison

A skilled machinist can estimate Ra to within a factor of 2 by touch alone, which is sufficient for most decisions about whether more finishing is needed.

Documenting Surface Quality

Always record:

• Method used to test
• Reference used (which surface plate, which straightedge)
• Result (qualitative: good contact across full surface; quantitative: max gap 0.025 mm at X location)
• Date

This documentation allows surfaces to be compared over time and problems to be caught before they become critical. A surface plate that is developing a low spot in the center will show progressive worsening in records, warning you to recalibrate before the error becomes significant.