What a surface roughness tester is
A roughness tester measures the texture a machining operation leaves behind. The drawing asks for it with a Ra figure beside a surface symbol; the instrument is what turns that callout into a number that can be recorded and argued about. The traverse is short - millimetres rather than metres - and it is normally taken on the finished part rather than on a sample, because the finish is a property of that particular part and not of the process in general.
The figure depends on the settings as much as on the surface. A stylus instrument separates roughness from the waviness beneath it with a filter whose cutoff is the sampling length, and Ra computed at one cutoff is not the same number as Ra computed at another on the same surface. The parameter set is not universal either: the same standard family defines Rz differently between editions, so a drawing that names a standard is asking for the parameters of that edition and not of some other one.
Choosing one is therefore mostly a question about the feature rather than about the instrument. A plane that can be laid on a bench is straightforward; a bore, a groove, a conical seat or a curved surface of small radius narrows the choice of pickup and sometimes rules out a skidded instrument altogether. The sections below work through the pickup, the parameters and the standards that have to be declared before a reading means anything.
How a surface roughness tester measures
The instrument draws a diamond stylus across the surface at a constant speed and records the vertical displacement, either against a skid that rides the surface or against an internal reference that does not. A skidded pickup is quicker to place and is the usual field instrument; a skidless one follows the nominal geometry of the part, which is what a curved or interrupted surface needs, because a skid would otherwise carry the curvature of the part into the result.
The recorded profile is filtered so that long-wavelength waviness is separated from short-wavelength roughness, and the cutoff that performs the separation is the sampling length. From the filtered profile the instrument computes the parameters: Ra as the arithmetic mean deviation, Rz as a maximum height over the sampling length, Rq as the root mean square, and a long tail beyond them covering spacing, slope and material-ratio parameters. Instruments differ in how much of that tail they offer, which matters as soon as a drawing names an unusual one.
The limits are geometric rather than electronic. A stylus cannot enter a groove narrower than its tip and cannot trace a peak sharper than its tip; the very tip radius that reads a fine finish faithfully is also the one most easily damaged, and it degrades quietly rather than failing. A skidded instrument reads along an arc when the surface is curved, and soft or gummy materials may be marked by the traverse itself. Where any of these governs the answer, the instrument has to be chosen around the feature instead of the feature being made to suit the instrument.
Configuration choices
Most of the specification is fixed by these decisions rather than by the accuracy class. Each one constrains the next, which is why they are settled in order.
| Configuration | Options to fix | What it actually decides |
| Pickup datum | A skid that rides the surface, or a skidless pickup referenced to the instrument body | Whether a curved, convex or interrupted surface can be traced at all, and how the instrument has to be placed on the part. |
| Parameter set and standard edition | Which standard family and edition the instrument computes, and how many parameters beyond Ra it offers | Whether the callout on the drawing can be measured as written. Rz in particular is defined differently between editions. |
| Sampling length and evaluation length | The cutoffs the instrument supports, and how many sampling lengths are averaged | The number that comes out. A reading is only comparable with another taken at the same cutoff. |
| Stylus tip radius and cone angle | The tip radius fitted, and the angle of the cone behind it | The finest detail the instrument can resolve, and the narrowest groove or slot the stylus can enter. |
| Traverse stroke | How far the pickup travels, and whether a single stroke covers the evaluation length required | Whether a coarse or a long surface can be measured without relocating the instrument part way through. |
| Orientation and fixturing | Hand-held placement, or a stand and measuring platform | Whether small or awkward parts can be held still enough for a valid traverse, and whether a curved part needs a platform. |
| Output and reporting | Direct display, internal storage, and export by cable, card or wireless link | Whether a batch record can be produced without transcribing readings by hand, which is where most transcription errors appear. |
| Verification | The reference specimen supplied with the instrument, and the interval at which it is used | Whether the instrument can be shown to be reading correctly, which most quality systems require at a defined interval. |
How to choose a surface roughness tester
| Factor | Why it decides the answer | How to settle it before enquiring |
|---|
| The feature to be measured | A plane, a bore, a groove, a step and a curved surface each need a different pickup geometry, and some need a different datum | Describe the feature with a sketch, and give the smallest radius, the narrowest width and the deepest reach involved. |
| The parameter and its cutoff | The answer depends on both, so an instrument that cannot compute the named parameter at the named cutoff cannot do the job | Quote the callout exactly as it appears on the drawing, with the standard and the edition it names. |
| How fine the finish is expected to be | A figure below the resolution of the instrument, or a tip too large for the peaks, produces a reading of something other than the surface | Give the target figure together with the process that produces it, rather than the figure on its own. |
| Size and accessibility of the part | The stroke has to fit the feature, and the instrument has to reach it without being supported by a surface that is not part of the measurement | Give the part size, the reach required, and whether the measurement is made on the machine or on a bench. |
| Whether the part may be marked | The traverse leaves a track, which is acceptable on a sample and not always acceptable on a finished or coated surface | State the material and whether a visible trace is tolerable on the delivered part. |
| Whether readings must be recorded | A batch record changes the output requirement, and sometimes rules out the simplest instrument in the range | State whether a report has to be produced, in what form, and at what interval. |
If any of these are still open, send what is known. A quotation can be issued against a description of the feature and refined once the gaps are filled.
The nearest alternatives
These are the types an enquiry for a surface roughness tester usually turns out to be, and the point at which the choice is genuinely decided.
| Nearest alternative | What separates it from a roughness tester |
| Hardness testers | Measures how resistant the material is to being pressed into, not how its surface is shaped. Both are taken on the finished surface and both are called up on the same drawings, but a rough surface and a soft one are independent faults and neither instrument sees the other. |
| Coating thickness gauges | Measures the depth of a layer sitting above the substrate rather than the texture of the outermost surface. A painted or plated part usually has both requirements, and the two instruments are used together rather than in place of each other. |
| Ultrasonic thickness gauges | Measures how much material there is, through the wall, rather than how the exposed face is finished. It answers a wall-loss or remaining-life question; a roughness instrument answers a fit, sealing or appearance question. |
| Optical and confocal profilers | Read an area rather than a line, so they capture the three-dimensional texture and are not limited by a stylus tip radius. They cost more and their result is not identical to a stylus result, which matters where the drawing names a stylus method. |
| Roundness and form testers | Measure departure from a nominal geometry - out of round, taper, squareness - which is form rather than texture. The two requirements sit on the same drawing and are measured on different instruments for different reasons. |
| Visual comparators and specimens | Give a pass or fail judgement against a reference surface without producing a number. Fast and adequate for shop-floor screening, but they cannot be recorded as a measured value or arbitrated in a dispute. |
Roughness, hardness and coating thickness are the three surface requirements that a drawing most often names together; the boundary between them is what is being measured, not where it is measured.
What to include in the enquiry
The first reply can be a quotation rather than a set of questions if the enquiry carries these. Where a figure is not known, send the duty description and we will confirm what it implies.
- The feature to be measured, with a sketch of the geometry and its smallest radius, width and reach
- The surface callout exactly as written on the drawing, including the standard and its edition
- The target figure, and the process expected to produce it
- The material, and whether the traverse may leave a visible trace
- The size of the part, and whether measurement is on the machine or on a bench
- Whether the pickup has to be skidless, and whether a curved or interrupted surface is involved
- Whether readings must be stored and exported, and in what form
- The reference specimen or calibration arrangement available locally
- Ambient conditions at the point of use, including temperature and vibration
- Quantity, destination port and required delivery date
