Selsyn Mechanical Interface Selection: Shaft Diameter, Flange And Mounting Dimensions—Precision Can Collapse with The Smallest Deviation

When the electrical parameters of a resolver are selected correctly, the system may still exhibit angle fluctuation and zero-position drift. The problem is often not in the decoding circuit but in a more hidden place: the mechanical interface. A resolver essentially measures angle through electromagnetic coupling between the stator and rotor. Any mounting eccentricity, axial misalignment, or loose fit will directly distort the symmetry of the air-gap magnetic field and write errors into the output signal. The core of mechanical interface selection lies precisely in three dimensions that seem to have “no technical content”: how to match the shaft diameter, how to define the flange, and how much margin to leave for the mounting dimensions.

Shaft Diameter: “It Fits Over” Is Not Enough

The connection between the resolver rotor and the motor shaft is the physical starting point of the entire signal chain. The first step in selection is to determine the resolver rotor’s inner diameter based on the actual diameter of the motor shaft. Manufacturers’ selection logic is usually as follows: determine the resolver rotor inner diameter according to the motor rotor shaft diameter, and customization according to shaft diameter is possible. Common through-hollow-shaft specifications cover sizes such as 6 mm, 8 mm, 9.52 mm, and 12.7 mm, while tapered shafts include 9 mm or 12 mm options with a 1:10 taper.

However, “fitting over” is only the bottom line. The real selection judgment lies in the nature of the fit. Between the rotor of a frameless resolver and the shaft, a slight clearance fit is recommended rather than an interference or tight fit. Dynapar’s mounting guide gives a reference range of 0.0002 to 0.001 in. of clearance (approximately 0.005 mm to 0.025 mm) between the shaft and the rotor inner diameter, to avoid line-to-line fits or interference fits. An interference fit may appear “secure,” but in reality it can introduce shaft bending or rotor core deformation during press-fitting, which instead damages air-gap uniformity. Although splined or keyed fits require lower press-in force, their machining and assembly clearances may introduce additional eccentricity components.

There is another detail that is easily overlooked: the coefficients of thermal expansion of the shaft and the resolver rotor materials should be as close as possible. If the motor shaft is steel while the resolver rotor sleeve is made of a material with a significantly different thermal expansion behavior, the fit clearance will change after temperature rise, and the air gap that was originally adjusted may shift in the hot state.

Flange: The Structural Form Determines “Who Takes Responsibility for Alignment”

The flange structure of a resolver essentially answers one question: who guarantees the positional accuracy of the stator. Different flange forms correspond to different ways of allocating alignment responsibility.

Frameless direct-mount type is the most “bare” solution: the stator is fixed directly to the equipment housing, the rotor is mounted on the drive shaft, and the resolver itself has no bearing. The coaxiality between stator and rotor depends entirely on the machining and assembly accuracy of the equipment. This type is suitable for scenarios with limited space and high lightweight requirements, but during installation the coaxiality between stator and rotor must be controlled within ≤0.05 mm. It places the highest demands on the equipment designer’s mechanical expertise and also offers the greatest freedom.

Flange-mount type has a mounting spigot and bolt holes on the resolver housing, and uses the fit between the spigot and the equipment end face for “pre-positioning.” After the bearing is integrated, the relative position of the stator and rotor has already been calibrated at the factory. During installation, it is only necessary to align the flange spigot with the corresponding recess on the equipment end face and tighten the bolts. Alignment accuracy is ensured by the flange spigot and end-face perpendicularity, making this type suitable for heavy equipment or working conditions with large vibration. The cost is greater axial length and higher cost, and the flange spigot itself becomes a new error link—the fit clearance between the spigot and the equipment recess, and the perpendicularity of the end face to the axis, still need to be included in the tolerance chain.

Bearing-integrated type further incorporates the bearing internally, with the rotor integrated with the bearing, so it can be mounted directly on the drive shaft without the equipment needing to provide an additional support structure. It absorbs “alignment” into the resolver itself to the greatest extent, but during selection it is necessary to confirm whether the bearing’s axial and radial load capacities match the working conditions, and whether the resolver’s own maximum speed covers the application requirements.

The essence of flange selection is a trade-off among equipment accuracy capability, axial space, cost, and ease of installation. If the equipment end face has high machining accuracy and space is tight, frameless direct mounting is a reasonable choice; if on-site assembly conditions are limited and rapid positioning is required, the flange-mount type uses the spigot to “outsource” alignment to the resolver housing and is more reliable.

Mounting Dimension Tolerance Chain: Eccentricity and Axial Misalignment Are Direct Killers of Accuracy

Resolver output accuracy is extremely sensitive to radial eccentricity and axial misalignment between the stator and rotor. A variable reluctance resolver modulates the signal through periodic changes in the air gap between stator teeth and rotor saliencies. Eccentricity changes the spatial distribution of air-gap permeance and introduces position-dependent error components into the output.

The general mounting tolerance framework given by Dynapar can serve as a design baseline: axial misalignment (axial deviation or dimensional variation between the rotor and stator mounting faces) should not exceed 0.015 in. (approximately 0.38 mm); radial eccentricity (eccentricity between the rotor and stator mounting faces) should not exceed 0.003 in. (approximately 0.076 mm); and the shoulder face of the mounting spigot should be perpendicular to the bore and shaft, with perpendicularity controlled within 0.0005 in. (approximately 0.013 mm). The engineering meaning of these numbers is clear: the eccentricity tolerance is about five times tighter than the axial tolerance, indicating that radial alignment is the main aspect of the accuracy conflict.

GB/T 10241—2020 also incorporates “coaxiality of mounting fitting surfaces” and “perpendicularity of mounting fitting end faces” into technical requirements and test methods at the standard level, meaning that these two indicators have risen from “manufacturer recommendations” to inspectable specification items.

When actually designing the tolerance chain, one should not focus only on the resolver’s own mounting face. The fit clearance between the rotor and the motor shaft, the fit clearance between the stator flange spigot and the equipment end face, and the perpendicularity of the end face relative to the axis—these three error segments accumulate in series onto the final air-gap uniformity. A practical principle is to treat the resolver mounting tolerance as the end link of the system tolerance chain and allocate it accordingly, leaving reasonable margin for upstream machining and assembly rather than concentrating all margin on the resolver mounting step.

Selection Checklist: From Shaft to Flange, Then to Verification

Connecting the above logic, resolver mechanical interface selection can proceed in three steps.

Step 1: Lock the shaft diameter and define the fit. Measure or confirm the actual diameter and tolerance of the motor shaft, and select a rotor with the corresponding inner diameter specification. Prefer a slight clearance fit and avoid interference; confirm that the thermal expansion behavior of the shaft and rotor materials roughly matches. If the shaft diameter is non-standard, customizing the rotor inner diameter is a common practice, but the customization lead time and cost need to be evaluated in advance.

Step 2: Allocate the flange according to the equipment’s accuracy responsibility. Can the equipment end face provide a high-precision spigot and perpendicular end face? If so, frameless direct mounting can provide a more compact structure. If on-site assembly conditions are limited or quick replacement is needed, the spigot positioning of the flange-mount type is more convenient. If the equipment itself cannot provide bearing support for the resolver, the bearing-integrated type is the natural choice. The flange’s bolt-hole pitch circle, spigot diameter, and axial mounting depth need to be matched item by item with the equipment end face; one cannot simply check that “the outer diameter is about right.”

Step 3: Zero-position verification after installation cannot be omitted. Even if the mechanical interface is aligned, the electrical zero position may still deviate from the motor magnetic pole position. For resolvers used for commutation, the standard practice is: first temporarily fix the stator, then use DC energization to “lock the shaft” or back-drive the motor and observe the back-EMF zero crossing, rotate the stator until the resolver zero-position signal aligns with the motor magnetic poles, and then finally tighten. No matter how well the mechanical tolerances are controlled, if the zero position deviates by several electrical degrees, commutation performance will still deteriorate significantly.

Resolver mechanical interface selection is essentially an error allocation problem: the shaft-diameter fit determines how “tightly” the rotor follows the shaft, the flange structure determines how “definitively” the stator is anchored to the housing, and the mounting tolerance determines the “fidelity” of air-gap symmetry. If any of these three links is loose, no matter how high the nominal electrical specifications are, the angle reading at the shaft end will be compromised. Clarify the mechanical path of the shaft and flange first, and then review the electrical parameters—this often leads to fewer detours than doing it the other way around.