Why Your Surface Roughness Readings Keep Varying (And How to Fix It)

You run the same part through your surface roughness tester twice and get two different Ra values. You check it again, a third number. Your customer wants documentation showing the part meets Ra 1.6 µm, but your readings are bouncing between 1.4 and 2.1. Sound familiar?

Inconsistent surface roughness measurements are one of the most common and frustrating problems in precision manufacturing. The good news: in most cases, the issue isn't your part. It's your measurement process. And it's fixable.

This post walks through the five most common causes of surface roughness measurement variation, how to diagnose each one, and what to do about it. If you're still seeing inconsistency after working through these, we'll cover when it's time to look at your equipment.

1. Stylus Condition

The stylus tip is doing the actual measuring by dragging across your part surface and translating micro-scale peaks and valleys into an electrical signal. If that tip is worn, chipped, contaminated, or the wrong radius, every reading it produces is suspect.

A stylus tip has a defined radius (typically 2 µm for most shop-floor applications). As it wears, that radius increases, and the tip can no longer reach into fine surface valleys so your Ra readings trend artificially low. On the other hand, a contaminated or damaged tip can snag on surfaces and produce erratically high readings.

How to check it: Inspect the stylus under magnification. Run a certified reference tile (most testers come with one) and compare your reading to the tile's known value. If you're more than 10% off, the stylus needs cleaning or replacement. As a rule, styli should be inspected every 500–1,000 measurement cycles in high-use environments.

The fix: Keep spare styli on hand. They're consumables, not lifetime components. If you're using a Starrett SR160 or similar portable tester, replacement styli are inexpensive relative to the cost of a failed inspection or rework.

2. Incorrect Cutoff Length (λc)

This is the most overlooked setting in surface roughness measurement and it causes massive variation when it's wrong.

The cutoff length (λc) defines how much of the surface profile the tester includes in each Ra calculation. ISO 4288 specifies which cutoff length to use based on the expected Ra range of your part. Using the wrong cutoff length is like measuring a road's smoothness with a ruler vs. a tape measure, you'll capture different features and get completely different numbers.

Standard cutoff length vs. Ra range (ISO 4288):

•      Ra 0.006–0.02 µm → λc = 0.08 mm

•      Ra 0.02–0.1 µm → λc = 0.25 mm

•      Ra 0.1–2.0 µm → λc = 0.8 mm (most common shop applications)

•      Ra 2.0–10.0 µm → λc = 2.5 mm

•      Ra 10–80 µm → λc = 8.0 mm

How to check it: Look at your tester's current cutoff setting. Compare it against the Ra range you're measuring. If your part is supposed to be Ra 1.6 µm but your tester is set to λc 0.25 mm, you're in the wrong range and your readings will be inconsistent and meaningless for documentation purposes.

The fix: Set the cutoff length based on your expected Ra range before measuring. If your drawing specifies a surface finish but not a cutoff length, default to λc 0.8 mm for most milled and ground surfaces in the Ra 0.4–3.2 µm range.

3. Measurement Direction Relative to Surface Lay

Surface roughness is not the same in all directions. A part that was ground longitudinally will show a very different Ra value when measured parallel vs. perpendicular to the grinding marks (the 'lay' of the surface).

Measuring across the lay, perpendicular to the machining direction gives the highest and most meaningful Ra reading. Measuring along the lay can give artificially low readings that look great on paper but don't reflect how the surface will perform in contact with a mating part.

How to check it: Rotate your measurement direction 90 degrees and compare readings. If there's a significant difference (more than 20–30%), you're seeing the effect of surface lay. Check your drawing, some specs call out lay direction explicitly with the standard ISO 1302 surface texture symbols.

The fix: Always measure perpendicular to the dominant lay direction unless the drawing specifies otherwise. Standardize this in your measurement SOP so every operator measures the same way.

4. Temperature and Vibration

Surface roughness measurement is sensitive to environmental conditions in ways that catch a lot of shops off guard especially when using a portable tester on the shop floor rather than in a controlled metrology lab.

Temperature affects both the part and the instrument. Steel expands at roughly 11.7 µm/m/°C at the micro-scale, even a few degrees of temperature difference between the part and the ambient environment can shift measurements. More commonly, vibration from nearby machinery is the culprit: the stylus is moving across the surface in microns, and any external vibration is read as surface roughness that isn't there.

How to check it: Take the same measurement with the machine running and with it off. If readings change, vibration is a factor. For temperature: let freshly machined parts cool to room temperature (typically 20°C / 68°F per ISO 1 standards) before measuring.

The fix: If vibration is unavoidable on the floor, use a vibration-damping pad under your tester or move measurement to a dedicated station. For critical applications, a benchtop profilometer in a controlled environment will always outperform a portable unit used next to a running CNC.

5. Inconsistent Measurement Location on the Part

Surface roughness naturally varies across a part, even on the same feature. A milled pocket will have different Ra values at the center vs. the edges, at the entry vs. exit of the cutter path, and on different passes.

If two operators are measuring the same part in different locations and comparing results, they will get different numbers and both of them will be correct. This is one of the most common sources of measurement disputes between a shop and its customer.

How to check it: Measure the same part five times in five different locations on the same feature. Record the variation. If it's large, your surface finish has natural variation and your SOP needs to define exactly where to measure.

The fix: Define measurement location in your control plan or inspection procedure. Use reference marks or fixtures to ensure every operator measures at the same spot. For PPAP submissions, this should already be documented if it isn't, add it.

When the Problem Is the Instrument Itself

If you've addressed all five causes above and still can't get consistent readings, it may be time to look at the instrument. Signs that your roughness tester is the problem:

•      Readings drift significantly throughout the day without any change in parts or process

•      Reference tile measurement no longer matches its certified value (even with a fresh stylus)

•      The tester has been dropped or subjected to impact — even once

•      You're using a tester with a resolution or range that doesn't match your application (e.g., trying to measure Ra 0.2 µm with an instrument rated for Ra 0.4 µm and above)

•      Calibration is overdue — NIST-traceable calibration is typically required annually for ISO 9001 and AS9100 compliant shops

For shops that need a reliable, portable roughness tester with solid performance for typical shop-floor Ra measurement (0.05–16 µm range), the Starrett SR160 is worth looking at. It covers the most common cutoff lengths (0.25, 0.8, and 2.5 mm), includes a reference tile, and produces traceable results suitable for customer documentation. We carry it in stock with fast shipping, see the link below.

Quick Diagnostic Checklist

Before assuming your parts are out of spec, run through this list:

•      Stylus tip inspected and verified against reference tile

•      Cutoff length (λc) matches expected Ra range per ISO 4288

•      Measuring perpendicular to surface lay

•      Part at room temperature, no nearby vibration sources

•      Measuring at a defined, consistent location on the part

•      Instrument calibration current

Fix these six things first. In the vast majority of cases, one of them is the root cause.

Need Help Choosing the Right Surface Roughness Tester?

At Precision Engineering Supply, we specialize in precision measurement equipment for manufacturing and quality professionals. If you're not sure whether your current tester is up to the job or you need to spec a new one, just call us at 617-398-7852 or use the chat on our site. We'll help you match the right instrument to your application, tolerance requirements, and budget.

Browse our surface roughness testers