Roundness Tester

What is a Roundness Tester

A roundness tester, also called a roundness measuring machine, circularity tester, or form tester, is a metrology instrument that measures the deviation of a circular cross-section from a perfect circle. The part is held on a precision rotary spindle, an electronic gauge with a stylus contacts the surface, and as the spindle turns through one or more revolutions the gauge records the radial position of the surface at thousands of angular points. Software then fits a mathematical reference circle to that profile and reports the largest inward and outward departures from it. The result, the roundness deviation, is reported in micrometres, and on high-end machines it resolves down to single-digit nanometres of gauge resolution.

The distinction that matters most is between form and size. A micrometer or a caliper measures size, the diameter of a shaft, but tells you nothing about whether the cross-section is oval, three-lobed, or scalloped. Two diametrically opposed points can read identical on a three-lobed part that is grossly out of round, because an odd number of lobes preserves the apparent diameter. Roundness is therefore an independent geometric characteristic that requires a continuous trace around the full circumference and a reference circle derived from all the data, not from two points. This is the reason a roundness tester exists as a separate instrument class rather than being folded into a length-measuring device.

Historically, roundness assessment moved from comparison against a master ring and a two-point diameter check, which cannot detect odd-lobed error, to the precision spindle method introduced commercially in the 1950s. Taylor Hobson is widely credited with the first dedicated roundness instruments under the Talyrond name, establishing the rotating-reference principle that all modern machines still use. The decisive advance was the air-bearing spindle, which floats the rotating member on a thin pressurised air film and removes the mechanical contact that otherwise caps accuracy. From the 1980s onward, electronic gauging, digital harmonic analysis, and motorised columns turned the single-plane roundness gauge into a full form-measuring system capable of cylindricity, runout, and position.

In application scale, roundness tolerances span a wide range. A general turned shaft may carry a circularity tolerance of several micrometres, a precision pump plunger or a hydraulic spool around 1 micrometre, a rolling-element bearing race a few tenths of a micrometre, and a gyroscope or fuel-injector component below 0.1 micrometre. Each tightness band drives a different class of instrument: a benchtop gauge for the loose end, an air-bearing form tester for the bearing and aerospace end. There is no single universal roundness tester, and matching the spindle error motion to the part tolerance is the central engineering judgement of selection.

The economic stakes are high because roundness directly governs function. An out-of-round bearing race generates vibration and noise and shortens fatigue life; an out-of-round sealing surface leaks; an out-of-round bore wears unevenly. In high-volume production, a roundness tester is the gatekeeper that catches a drifting grinding wheel, a worn chuck, or a thermal distortion before it produces a batch of scrap, which is why the instrument is justified by avoided scrap and warranty cost rather than by the inspection itself.

Instrument Types and Architectures

Roundness testers divide first by which member rotates and second by how much of the form-metrology task they cover. In the rotating-table architecture the part sits on a turntable and the stylus is held stationary, which suits small and light components that can be centred and levelled on the table. In the rotating-spindle architecture the part stays fixed on a base and the stylus assembly rotates around it, which suits large or heavy parts that cannot be spun. The reference circle accuracy is governed by the spindle in both cases. Below this primary split, instruments range from single-plane benchtop gauges to full cylindricity form testers and, at the high end, multisensor systems that combine roundness with surface texture and contour. The table below compares the main classes.

Air-bearing spindle is the heart of any serious roundness tester. A diamond-turned shaft floats on a film of pressurised air a few micrometres thick, eliminating metal-to-metal contact, so there is no wear, no break-in, and radial error motion can reach below 25 nanometres. Because the spindle never touches its housing, accuracy holds for the life of the machine provided the air supply is clean and dry. Representative air-bearing instruments include the Taylor Hobson Talyrond 565, the Mitutoyo Roundtest RA-2200, and the Mahr MarForm MMQ 400. The trade-off is the need for a filtered, regulated compressed-air supply and a vibration-isolated location.

Mechanical-bearing spindle uses a precision rolling or plain bearing and is found on entry-level and portable gauges. It needs no air supply and is more tolerant of a workshop environment, but its radial error motion is typically limited to around 0.5 micrometre and it wears over time, so periodic verification is essential. This class is appropriate when the part tolerance is several micrometres and the cost of an air-bearing system is not justified.

Single-plane versus full form. An entry instrument measures roundness in one cross-section at a time. Adding a motorised, straightness-calibrated vertical column lets the stylus traverse along the axis, so the software can stack many roundness traces into a cylindricity assessment and also compute straightness, coaxiality, parallelism, perpendicularity, taper, and runout. A horizontal radial axis adds diameter and the ability to reach internal and external features. When a datasheet quotes only a spindle radial figure, it usually describes a single-plane gauge; cylindricity capability requires a separately specified column straightness, which is discussed in Chapter 5.

Centring and levelling stage. Every roundness tester needs the part axis aligned to the spindle axis. Eccentricity of the part introduces a first-harmonic (1 UPR) sinusoid that the software removes mathematically, but excessive eccentricity also introduces second-harmonic distortion that masquerades as ovality, so a fine centring and tilting table, manual on bench units and motorised on CNC systems, is part of the architecture rather than an accessory.

Reference Circles and Filtering

A raw roundness trace is just a closed profile; it becomes a number only after a reference circle is fitted and a filter is applied. Both choices change the reported value, sometimes by a large factor, so both must be specified on the drawing for any roundness figure to be comparable between two laboratories. ISO 12181-1 defines four association methods for the reference circle, and ISO 12181-2 defines the extraction and filtering conventions. The table below summarises the four reference circles and when each is used.

LSCI (least squares reference circle) is the everyday default because it uses all the data, is mathematically unique, and is insensitive to a single spurious point, giving stable and repeatable results across operators. The roundness deviation under LSCI is the sum of the maximum outward peak and the maximum inward valley measured from the least squares circle, reported as the total parameter. It is the right choice for process monitoring and statistical control.

MZCI (minimum zone reference circle) finds the pair of concentric circles with the smallest radial gap that still brackets the whole profile, and that gap is the roundness deviation. Because it is the smallest value that honestly contains the form error, ISO regards it as the reference definition, but it is more sensitive to outliers and can be slightly less repeatable. MICI and MCCI are functional circles: the maximum inscribed circle models the largest pin that a measured bore will pass, and the minimum circumscribed circle models the smallest ring that a measured shaft will enter, so they are chosen when the roundness value must predict an assembly fit rather than describe form in the abstract.

Once a reference circle is chosen, a filter selects which spatial frequencies of the profile are reported. The trace is decomposed by harmonic, that is Fourier, analysis into undulations per revolution (UPR), the number of full waves around one rotation. A low-pass Gaussian filter with a defined UPR cutoff then suppresses the high-frequency content so that surface texture does not inflate the form result. The 1 UPR term is eccentricity and is always removed by centring; 2 UPR is ovality; 3 UPR and higher odd lobes are the classic chuck and centreless-grinding signatures. The table below lists common cutoffs.

The practical consequence is that a lower UPR cutoff smooths the profile and reports a smaller roundness value, while a higher cutoff includes more waviness and reports a larger value. A roundness number quoted without its cutoff and reference circle is meaningless, and this is the single most common cause of disputes between a supplier and a customer measuring the same part on two machines.

Standards and Calibration

Roundness metrology rests on a small family of geometrical product specification (GPS) standards. ISO 12181-1 sets the terms, definitions, and parameters of roundness, including the four reference circles and the roundness parameters; ISO 12181-2 sets the specification operators, that is the extraction and filtering conventions. For three-dimensional form, ISO 12180-1 and ISO 12180-2 do the same for cylindricity, integrating the roundness assessment along the axial direction. The drawing-level tolerancing of circularity and cylindricity sits within the geometric dimensioning and tolerancing framework of ISO 1101, with the equivalent United States practice in ASME Y14.5. A complete roundness callout therefore names the tolerance value, the reference circle, and the filter cutoff, so that the same evaluation can be reproduced anywhere.

The roundness parameters mirror the surface-texture family. RONt is the total roundness deviation, the radial distance from the highest peak to the lowest valley relative to the reference circle, and it is the parameter that most drawings call out. RONp is the peak height, the maximum outward departure from the reference circle to the highest peak, and RONv is the valley depth, the maximum inward departure to the lowest valley, so that RONt equals RONp plus RONv. These follow the same logic as Rt, Rp, and Rv in roughness, which makes the parameter set easy to read for anyone already fluent in surface metrology.

Calibration uses two distinct artefacts because there are two distinct tasks. A glass or ceramic hemisphere, or a high-grade reference sphere, is a near-perfect zero reference: rotating it lets the spindle error motion be checked, and the residual roundness should read close to zero on a healthy machine. A flick standard, a precision cylinder with a single known flat ground into it, or a multi-wave standard, carries a certified roundness deviation and is used to set the gauge magnification (gain) so that a known departure reads the correct value. Multi-wave standards give a better signal-to-noise ratio than a flick across a range of filter cutoffs. Both artefacts are certified with traceability to a national metrology institute, and the instrument calibration procedure is normally repeated at least annually and after any relocation. The table below summarises the standards and artefacts.

A final standards point concerns the relationship between size and form tolerances. A common engineering rule is that the cylindricity or roundness tolerance of a bearing seat should be one to two IT grades tighter than its dimensional tolerance, so an IT6 diameter is typically paired with an IT4 to IT5 form control. This is why a roundness tester capable of an order of magnitude finer resolution than the part tolerance is needed: the form tolerance is deliberately tighter than the size tolerance, and the instrument must outresolve both.

Key Specification Parameters

Reading a roundness tester datasheet is a skill in itself, because a headline spindle number rarely tells the whole story. The parameters that actually drive a selection are spindle radial and axial error motion, vertical column straightness, horizontal axis straightness, gauge range and resolution, maximum part diameter, height and weight, and the software parameter and filter coverage. Each is explained below.

Spindle error motion (radial and axial). This is the deviation of the rotating axis from a perfect circle of rotation, and it adds directly to every measurement. Manufacturers express it as a fixed term plus a height-dependent term. The Mitutoyo Roundtest RA-1600 specifies radial accuracy of about (0.02 + 6H/10000) micrometre, where H is the probing height in millimetres, while the higher-class RA-2200 improves the coefficient to (0.02 + 3.5H/10000) micrometre. The Taylor Hobson Talyrond 131 air-bearing spindle is rated around plus or minus 0.02 micrometre, and the Talyrond 565 around plus or minus 0.015 micrometre with gauge resolution near 0.3 nanometre. The key reading skill is that the error grows with measuring height, so the figure quoted at the gauge plane is not what you get 200 millimetres up a tall part.

Column and radial straightness. For cylindricity, the vertical column that traverses the stylus along the axis must itself be straight, and its straightness error adds to the cylindricity result exactly as the spindle adds to roundness. A datasheet that quotes a superb spindle figure but is silent on column straightness is describing a roundness-only gauge. A horizontal radial axis straightness governs the accuracy of diameter and of internal-feature reach.

Gauge range and resolution. The electronic gauge has a finite measuring range, often around 1 to 2 millimetres, within which it is linear, and a resolution that can reach single-digit nanometres on high-end systems. A large range eases setup on roughly centred or stepped parts; a fine resolution is needed when the part tolerance is in the tenths of a micrometre. The two trade off, so high-end machines offer multiple gauge ranges.

Capacity envelope. Maximum part diameter, maximum measuring height, and maximum load on the table or spindle set the physical limits. A compact bench unit such as the Mahr MarForm MMQ 200 targets smaller parts in the production area, while universal systems such as the MMQ 400 series accept larger and heavier work with horizontal and vertical axes of various lengths. Verify both the diameter that fits over the spindle and the height the column can reach, since a part can fail either limit independently.

Software and parameter coverage. The instrument is only as useful as the analysis it can perform. Confirm support for all four reference circles, the full set of roundness parameters, the required UPR filter cutoffs, and, for form testers, cylindricity, straightness, coaxiality, concentricity, parallelism, perpendicularity, taper, and both circular and total runout. The relevant signals to verify on a datasheet are:

• Reference circles: LSCI, MZCI, MICI, and MCCI all selectable, not just least squares.
• Filtering: Gaussian filter with selectable UPR cutoffs across the 1 to 1500 range, and harmonic analysis output.
• Form and position: cylindricity per ISO 12180, plus straightness, coaxiality, runout, and taper for column-equipped systems.
• Reporting: polar plots, harmonic spectra, and tolerance pass or fail against ISO 1101 callouts.
• Automation: motorised centring and levelling and CNC part programs for repeat production inspection.

Stylus and probing force. The contact tip radius and the probing force interact with surface texture: too sharp a tip or too high a force can plough soft material and bias the reading, while too blunt a tip mechanically filters out fine lobes. Match the tip and force to the part material and to the UPR cutoff you intend to report, and record both in the measurement procedure so results are reproducible.

Selection Decision Factors

To turn the preceding five chapters into a specific model, follow the decision sequence below. Most selection mistakes come not from one wrong step but from deciding the instrument class before the part tolerance is known, so work strictly top to bottom. These steps double as a fixed RFQ template.

1. Tightest form tolerance: Find the smallest circularity or cylindricity tolerance you must verify across all parts, then require a spindle error motion no larger than one tenth of it. A 1 micrometre roundness tolerance points to an air-bearing system around 0.02 micrometre or better; a several-micrometre tolerance can be met by a mechanical-bearing bench gauge.
2. Form scope: Decide whether you need roundness in one plane only, or full cylindricity, runout, and position. Single-plane work allows a compact gauge; cylindricity requires a calibrated vertical column with its own straightness specification, which is a different and more expensive machine.
3. Part envelope and weight: Confirm the maximum part diameter clears the spindle, the maximum height fits under the column, and the part weight is within the table or spindle load limit. Heavy or large parts may force a rotating-stylus rather than rotating-table architecture.
4. Gauge range and resolution: Match gauge resolution to the part tolerance, with at least a factor-of-ten margin, and check that the gauge linear range covers your typical setup eccentricity and any stepped features.
5. Reference circle and filter coverage: Ensure the software offers the reference circle (LSCI, MZCI, MICI, or MCCI) and the UPR filter cutoffs your drawings call out, plus harmonic analysis to diagnose process faults by lobe order.
6. Environment and utilities: Air-bearing machines need clean dry regulated compressed air and a vibration-isolated, temperature-stable room near 20 degrees Celsius. Confirm the floor, air supply, and thermal environment before committing, or the rated accuracy will not be realised.
7. Automation and throughput: For production inspection, weigh motorised centring and levelling and CNC part programs against manual setup; in high volume the operator time saved often outweighs the capital premium.
8. Total cost of ownership (TCO): Add purchase price, installation and foundation, annual calibration with traceable artefacts, air supply and environment, and operator training. A bench gauge that saves capital but cannot resolve the tightest part tolerance leads to escaped defects whose cost dwarfs the saving.

One dimension that buyers regularly overlook is manufacturer serviceability: local availability of certified flick and sphere artefacts, accredited recalibration service, spindle rebuild or exchange capability, and long-term software support. A roundness tester is a multi-decade asset, and its accuracy depends on regular traceable verification, so a supplier with a national calibration laboratory and parts inventory matters more than a marginal headline-spec advantage. Taylor Hobson, Mahr, Mitutoyo, Zeiss, and Jenoptik Hommel have established form-metrology service networks, which makes them safe choices for parts whose roundness drives function and warranty exposure.