Machine Design

Precise moves depend on ball-screw threads

The degree to which a ball screw performs well in the field is the direct result of the process used to make it.

George Jaffe
Executive Vice President
Steinmeyer Inc.
Bedford, Mass.

Alexander Beck
Steinmeyer GmbH

Rolling uses a pair of rolling dies to cold form continuous screw threads in soft-steel bars. Heat treatment (hardening) takes place after the rolling process.

Thread grinding is done on hardened bars, either in one step or as a followup to another process such as whirling. Center holes on the bar ends define the axis of symmetry and act as the main reference. This ensures optimal concentricity of all functional surfaces. Steady rests flanking the grinding wheel provide support during the grinding process.

Whirling cuts threads in hardened-steel bars with a hollow tool wheel (whirling ring). Bars are centered with steady rests. Chips coming off the bar are red hot.

Grinding, rolling, and whirling are the most common ways to thread ball screws. Some manufacturers use a combination of the three methods in an effort to lower cost, raise production, or boost device accuracy.

Often, both manufacturers and end users focus only on lead accuracy and compare it to permissible position error. This is a simple and, in most cases, insufficient approach.

How a screw is manufactured can significantly influence its precision and material properties, especially fatigue resistance. Lead deviation, bearing journal run-out (or concentricity with the ball tracks), thread roundness, lead wobble, profile tolerances, and surface roughness — all are affected by the manufacturing method. And these influence frictiontorque variation, audible noise level, nut stiffness, and other performance metrics.

Grinding consumes the most time of all three processes, not just because the metal-removal rate is relatively slow. It's also that grinding consistent, high-quality threads takes a number of additional steps. These include grinding the shaft OD so it can be supported in steady rests (steadies), and the forming or machining of threads to reduce grinding allowance — the amount of material grinding is to remove.

Longer cycle times aside, grinding has a decided edge in terms of accuracy and precision, especially for ball screws in grades better than Class 5 or 3. Grinding can make ball screws with oversize journals and dead-end threads without thread relief.

The process also allows individual control of various accuracy parameters. Screws may be measured throughout the grinding process and "touched up" as needed to reach extremely high precision levels. This holds true not only for overall lead accuracy, but for lead wobble, thread-diameter consistency, roundness, profile truth, and so on.

Grinding works best with hardened materials, so all heat treatment including hardening and tempering takes place beforehand to minimize thermal distortion. Removing the grinding allowance in several steps minimizes stress relief. Also, screws are cut to length before grinding so the center holes that serve for thread grinding also provide a reference for all subsequent operations such as cylindrical grinding of bearing journals. This ensures optimum concentricity of all functional surfaces.

Rolling, in contrast, cold forms soft steel. Hardening takes place after the rolling operation. Rolling introduces compressive stresses in the case material that considerably change bar length once the shaft undergoes hardening. Therefore, it is heat treatment of the rolled bars and not the thread forming process itself that controls lead accuracy. Moreover, overall lead accuracy may be held only over a limited distance. Wobble and other geometric thread tolerances are not directly controllable.

It is often said that the dense material and press polish from rolling improves ball-screw performance and durability. Unfortunately, this simply isn't true. Screws are heat treated after rolling, so both the dense microstructure and smooth surface go away. The high temperature rearranges the microstructure to that of a machined or ground screw. Also, the press-polished surface forms a dark, scaly oxidation layer that must be removed. Typically a polishing process cleans the surface, where surface roughness depends on the grit of the polishing wheel.

Because rolling is a continuous process, it is not possible to make deadend screws as with grinding. Long bars are rolled and all other machining takes place afterwards. The OD of the rolled bar serves as a reference for the machining of center holes. These center holes are then used for subsequent journal grinding, or for turning bearing surfaces directly. As such, certain screw concentricity errors are unavoidable with the method

Whirling is a newer, less-well-known process often used for removing metal on soft bars and reducing the allowance for grinding. The center groove in the ball track of ground screws typically comes from prewhirling. When whirling is the final thread-making process, bars are case hardened beforehand to minimize thread lead changes. These lead errors are smaller than when a rolled bar is hardened, however.

Because whirling takes place in one pass, there is no way of making corrections to the threads after the fact. The reference diameter is also the OD, like in a rolled screw. Center holes and journals are then machined using the OD as a reference. Thread-diameter control depends on the ability to keep a bar centered in the steadies under extremely high process loads, which is an inherent weakness of whirling. Concentricity of threads and journals are comparable to those of a rolled screw.

Some Fine Points
Many machine designers spec'ing ball screws consider only overall lead accuracy as it relates to maximum permissible position error, at any given point along the device's travel. But other factors should be considered as well.

First, when a nut is preloaded to make it backlash-free, preload and friction torque vary in step with diameter fluctuations of the ball threads. This can be problematic for the drive, either when torque exceeds certain limits, or because the servoloop is unable to compensate for the fluctuations.

Second, geometric tolerance variations of a shaft's ball track cause uneven compression of balls between a shaft and nut. Both stiffness and service life suffer because some portions of load-carrying surfaces receive peak loads, while others are mostly unloaded.

Third, side loads from poor concentricity of functional surfaces such as ball tracks and bearing journals, raise loads and shorten life, factors service-life calculations don't consider.

Finally, lead wobble can trigger excessive noise and vibrations. Lead wobble is a periodic lead error that happens once per revolution and is hard to detect in lead-accuracy measurements. Run-out, lack of roundness, or improper surface finish, are typical causes of lead wobble.

When comparing manufacturing processes consider their effect, good or bad, on material properties that are so important for long ball-screw life. Ball screws need good fatigue strength at and immediately beneath the surface of ball tracks. Simply put, the material must be sufficiently hard (58 HRC or better), be under compressive load from the heat treatment, and have a smooth surface yet retain sufficient lubricant. Conversely, cracks from the machining process that were not entirelyremoved with the chip, as well as annealing or even rehardening from thermal loads during machining, all are undesirable.

Gradual cooling — after a cutting edge passes by — can anneal a surface and compromise hardness. Annealing may penetrate material to sufficient depth such that a standard Vickers test is able to detect the loss. In most cases, however, the so-called soft skin from cutting heat is too thin to be caught without special microhardness testing. But the result is the same: lowered fatigue resistance of track surfaces.

On the other hand, rapid cooling after cutting can reharden the material, a phenomenon hardness testing fails to detect. The surface material tries to expand when hot but can't because it is still connected to cooler material below. So the hot skin compresses and in extreme cases may plastically deform. Once temperature normalizes the material tries to shrink but can't. This induces surface tensile stresses, which hurt fatigue resistance. This so-called grinder burn is sometimes visible as annealing discoloration, but not always. The detection of grinding burn requires chemical etching, or complicated eddycurrent measurements.

Metal removal through any sort of cutting involves mechanical stresses that exceed a material's strength so the chip can be removed. This thermally and mechanically "damages" the material along the cut. Should this damaged material not be removed with the chip, the resulting microcracks will stay in the surface. Ball-screw tracks are especially sensitive to crack propagation, so these microcracks can significantly shorten service life.

Surface finish is critical as well. Ideally, ball-screw surfaces should have a large, plateau-shaped, load-bearing surface with some indents to retain lubricant. In this context, a perfect, mirror-polished surface is as bad as one that's excessively rough.


Lead accuracy
Very good
Medium; controlled by subsequent heat treatment
Medium; depends on machine accuracy and stress relief
Diameter control
Very good; direct through steadies and infeed axis
Medium; indirect through bar diameter
Poor; direct, but depends on centering of bar under machining load
Lead wobble
Very good
Not controlled
Thread roundness
Very good
Poor; whirling cannot produce round threads
Surface finish
Good; depends on wheel used
Depends on polishing process
Good; depends on cutter quality


Steinmeyer Inc.
Bedford, Mass.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.