Mark Pearson, of Belt Technologies Europe, explains how designers can make the most of metal belts in their designs

Metal belts offer design advantages beyond the scope of rubber belts, chains, and cables. Made of high-modulus spring steels, they are strong, light, and almost non-stretchable, making them ideal for precise positioning. Metal timing belts can be fabricated with extremely tight pitch tolerances, and the single-piece design is free from the pulsation of chordal action often seen in other belts.

A key step in optimising a metal-belt design is to determine loads and stresses induced by the system. In addition to steady-state operation, designers should consider unusual or intermittent conditions such as jam-up loading, high startup loads, or indexing. In most cases, loads should not exceed the belt’s ultimate strength.

Calculating belt stresses is a five step process. Designers first need to determine the working load and, secondly, peak load. Third, the working stress needs to be calculated by dividing the working load by the belt’s cross-sectional area. Fourthly, the bending stress should then be determined and finally, the total stress of the system needs to be calculated by adding the working and bending stress.

It is generally recommended that total stress does not exceed half the belt material yield strength. During the design process, it is advisable to review the various parameters and work back through the calculations to find the combination that satisfies application requirements. For example, a wider belt reduces working stress without changing bending stress. Larger pulley diameters reduce bending stress, or allow use of thicker belts that, in turn, reduce working stress.

One of the most important advantages a metal belt offers is overall accuracy. Perforated belts or belts with attachments can be fabricated with pitch accuracies of +0.013mm. However, engineers often mistakenly perceive absolute accuracy in timing belts as synonymous with nonaccumulation of pitch error. But to design an effective timing or indexing system requires only the ability to precisely repeat an operation, again and again.

In operation, the belt’s timing element (perforation or anchor) engages a complementary pulley timing element (tooth or pocket). The belt pitch tolerance (typically + 0.013mm) accumulates over the length of the belt causing two unfounded design concerns.

One says that the belt will not position correctly through its length. The second, that misregistration of the belt and pulley occurs as belt accumulation of error compounds relative to the pulley with each revolution of the belt.

Performing the calculations

In two-pulley timing/indexing system, work is usually performed at the centre-to-centre distance between pulleys, with pitches moving on and off this distance as the belt indexes. This pitch movement can be thought of as a moving average of pitches relative to the total number of pitches in the belt length. Therefore, what becomes important is how constant, or repeatable, the accumulation of error is over the pulley centre distance. A moving average also develops on the pulley.

Total accumulation of error in the centre-to-centre distance between pulleys determines how other system components, such as pick-and-place actuators, are set up in relation to the belt. After completing the setup, the belt must hit these targets within some tolerance: that is, the belt must be repeatable.

This concept is best illustrated by example. Consider a system with two 80.7mm diameter pulleys on centres 1828.8mm apart. Timing pitch = 25.4mm and belt length = ( 1828.8 3 2) + ( 3.178p ) = 154 pitch.

Assume a pitch error of 0.013mm, and total accumulative error in belt length = 154 3 0.013mm = 2.0mm. The total accumulative error in pulley centre distance = 72 pitches 3 0.013mm = 0.936mm. Further, assume that the belt must hit a target every sixth timing pitch. This master pitch Pm has an error of 6 (0.013mm) = 0.078mm. System components must be set up with P1 =152.4 1 0.078mm, P2 = 304.8 1 0.156mm, ….P12 = 1828 1 0.936mm.

Once the system is set up, the nonstretch metal belt prevents misregistration to individual system components. The constant, repeatable accumulative pitch tolerance between every P1 and P12 is 0.936mm.

This example makes the simplifying assumption that pitch tolerance accumulates positively by 0.013mm. With an actual pitch tolerance of 0.013mm, experience shows that a timed metal belt has total repeatability of 0.05mm to 0.125mm. For applications requiring repeatability better than 0.05mm, a metal belt should be used in conjunction with a final part-registration technique.

Misregistration of the belt and pulley does not occur because accumulative belt pitch error in 180i of wrap would have to exceed the clearance between belt and pulley timing element. In the example, the 80.7mm diameter pulley and 0.127mm-thick belt combine for a pitch diameter of 80.827mm, and belt circumference is 253.926mm. With 180iof wrap and a 25.4mm pitch, five timing locations engage the belt and pulley and total error = 5 (0.013) = 0.065mm. By design, the difference between belt and pulley timing-element diameters is up to 0.2mm, so the accumulated error of 0.065mm is insufficient to cause belt/pulley misregistration. Thus, while it may seem desirable to specify a nonaccumulative pitch tolerance, it is not typically feasible or necessary.

A number of other considerations come into play when designing metal-belt drive systems. Timing pulleys for metal belts are either toothed or pocketed, each engaging respective belt perforations or drive lugs. Timing pulleys should be designed to ensure that all timing elements have spherical or involute radii. This ensures smooth engagement and disengagement of the belt and pulley. To avoid problems due to accumulated tolerances, the diameter difference between driving and driven components should typically be at least 0.015mm to 0.2mm. Zero or near-zero backlash applications are a special case.

When manufacturing a toothed pulley, each timing tooth fits into a hole machined in the pulley body. Care must be given to the radial location of each tooth to ensure overall pitch accuracy. The pitch diameter of a timing pulley must be at the neutral axis of the belt (one-half the belt thickness for a thin flat belt), not at the base. Because metal belts are generally thin, there is a temptation to neglect their thickness in calculating the pulley tape support diameter. Failure to include the belt thickness in these calculations results in mismatching of timing elements.

{{Tape support diameter is determined by:

D = ( NPp/p ) – t

where N = number of pitch lengths or teeth on a pulley, Pp = perforation pitch and t = belt thickness.}}


Belt tension should generally be kept as low as possible to improve belt life and reduce wear on other components. Tension should not be increased to reduce sag between pulleys. To prevent sagging, the belt should be dragged across a stationary support surface such as one made of ultrahigh-molecular-weight plastic. Sliding the belt across a stationary surface has a negligible effect on tracking and belt life. Rotating surfaces, such as a series of tangential pulleys, will act as steering rolls and can cause tracking difficulties.

Overtensioned belts may develop a cross bow. In addition, overtensioning causes uneven motion and reduces repeatability. Tension should be determined by operating the system and selecting the lowest possible workable setting, and maintained with air cylinders, springs, or jack screws.

A stiff system frame allows fine adjustments for timing and belt tracking. If the frame allows uncontrolled flexing, the system will bow when the belt is tensioned. Offsetting one force (system flex) with another (axis adjustment) does not provide a controlled system and can result in tracking problems.

The best designs use only two pulleys. Multiple-pulley systems with reverse bends increase belt bending stress, compromising belt life. Because each pulley can have a steering influence, tracking problems can result as well.

Preferred designs have solid mounts at each end of the pulley shaft. Cantilevered shafts can create a pivot. Belt tension may deflect the shaft and cause tracking problems. If cantilevered shafts are necessary, designers should ensure the stiffness of the structure with proper frame design and shaft rigidity.

Certain system requirements may demand design trade-offs. For example, metal belts can operate with pulleys as small as 6.3mm, but a smaller diameter reduces belt life. Also, belts can operate in ovens at temperatures to 600iC but because much of the belt’s strength comes from cold working or specific heat treatments, high temperatures can reduce belt strength.

Tracking a metal belt is important because metal belts do not stretch to compensate for lack of system squareness or alignment, pulley/shaft deflection, differential loading, and belt camber. Three techniques are used to track belts on systems using friction pulleys, timing pulleys, or both.

Adjusting the pulley axis in a metal-belt system is the most effective way of tracking a metal belt. This changes belt-edge tensions in a controlled manner, steering the belt. Second, the designer can consider using crowning friction-drive-pulleys. When a system requires crowned friction-drive pulleys, they should be used in addition to axis adjustment. This is because crowned pulleys will not self-centre a metal belt. Crowned pulleys work best on thin belts because the belt web must conform to the crowned face of the pulley.

Where simple axis adjustment cannot eliminate improper tracking, forced-tracking methods such as flanged pulleys and cam followers may be necessary. But forced tracking techniques can decrease belt life. System design relationships may need to change, requiring, for instance, a thicker belt than might be otherwise recommended.

One forced-tracking technique of wider belts uses a V-belt bonded to the inner circumference of the belt. This two-element belt distributes tracking stresses in the V-belt, rather than on the metal belt, maximising the life of the belt in operation.

Figure 1: Metal belts are strong, light and do not stretch

Figure 2: Drive tapes are made from the same metal strip as belts but are not endless. Rather, they are fitted with special end attachments or perforations. They perform with near-zero backlash in applications that include carriage positioning, read/write head positioning, plotters and optical drives