Belt drives were widely used in industrial machinery by the early nineteenth century and later came into use in agricultural machinery. Industrial use of belt drives continued well into the
twentieth century. Many of the belts were made of balata (a non-elastic rubber from a South American tree), although leather and rope were also used.
B. Smythe’s British Facilitator tricycle design of 1819 called for a two-speed belt drive.
(Whether such a machine was actually built isn’t known with certainty.) In 1869, Frederick Shearing of Norfolk designed a bicycle with belt drive to the rear wheel. He claimed to have built three machines (Ritchie 1975, 122–123). Also in 1869, Alphonse Barberon and Joseph Meunier of France patented a three-speed belt-drive system. Similar in principle to stepped pulley systems that had been used for decades in factories, it was a precursor of derailleur gearing. Whether Barberon and Meunier marketed their belt-drive system (for which they re-ceived British patent 2,626 of 1869) is not known.
A rare early instance of use of a flexible drive belt on a commercially made cycle was the BSA-manufactured Otto dicycle, discussed in the preceding chapter. Eduard Carl Friedrich Otto, a British subject living in Surrey, patented his belt-drive system in the UK and the US (British patent 1,673 of 1880). Most drive belts of that time were canvas or leather, but Otto’s was steel. It ran on pulleys covered with India rubber, vulcanite, or leather for better grip. As recently as 2011, David Gordon Wilson experimented with a modern stainless steel belt and found that the high tension it required caused the cycle’s frame to break (Wilson 2004, 326).
With belt drives, as Archibald Sharp pointed out (1896, 396), “the effort transmitted is the difference of the tensions of the tight and slack sides of the band; the maximum effort that can be transmitted is therefore dependent on the initial tightness.” “If the speed . . . be low,” Sharp added, “the tension necessary to transmit a certain amount of power is relatively high. In such cases the available friction of a belt on a smooth pulley is too low. . . .” Because the flat belts of the time depended on friction, belt drive made little progress in the early days of cycling.
In the 1970s, interest in belt drive revived. At the Massachusetts Institute of Technol-ogy, David Gordon Wilson converted a 1960s Moulton bicycle to belt drive, using an ordinary industrial toothed belt (Whitt 1978, 131). Since the 1960s, toothed belts had become more durable and more efficient, owing to the use of materials such as urethane for the belt and teeth and aramid fibers for tensile reinforcement. But because of the bending that occurred as the belt passed over its sprockets, toothed belts were generally not quite as efficient as well-lubricated roller chains. There was also the possibility of a slight loss of efficiency caused by stretching of the belt under extreme tension. Furthermore, a drive belt on a bicycle sometimes jumped a tooth, even when the tension was correctly adjusted—for example, when a heavy rider accelerated energetically after a stop. On the other hand, a toothed belt required no lu-brication and was more efficient than a poorly lubricated chain.
By 1985, Bridgestone had launched its Picnica, a folding model that Bridgestone claimed was the world’s first series-produced belt-drive bicycle. Bridgestone also produced belt drives for bicycles made by other companies, such as those used on the 2012 Corratec B-Drive city bikes.
From time to time, other belt-drive bicycles appeared on the market, mostly short-range folding or commuter models. One of the best-known and most enduring was the Strida, a fold-ing small-wheeler. Conceived by the British designer Mark Sanders, it has been in and out of production since 1987.
The Gates Corporation of Denver, founded in 1911, became a market leader in drive belts for many different applications. Gates designed and made the belt drives used on the Strida from the 1980s on. These belt drives were specifically refined for bicycle use. Gates had to overcome problems with thermal expansion of aramid tensile reinforcement, and also
Figure 4.3 A Gates Carbon Drive system. Note the provision for separating the seat stay from the wheel dropout for replacement of the belt (Gates Corporation, photo by Robert Gebler).
redesigned the belt teeth to minimize tooth jumping. Gates’s lightweight Carbon Drive belt-and-sprocket system, introduced in the early twenty-first century specifically for bicycle use, had strands of carbon fiber and teeth of nylon-jacketed polyurethane. The back of the belt was ribbed at an angle of 90º to the length of the chain to reduce bending resistance. The Carbon Drive achieved some success in racing and was used in 2009 by the British rider James Bowthorpe on what was then the fastest around-the-world bicycle ride. Moreover, in an independent third-party test conducted in September of 2007 the Carbon Drive was found to be as efficient as a chain drive (Microbac 2007).
The German Thun system of the late 1990s used a toothed V-section belt and an over-sized plastic rear sprocket that was also an internally toothed ring. The ring engaged with a plastic spur wheel on the rear hub. As the belt pulled the “floating” rear sprocket around, the engagement of its inner teeth with the spur wheel turned the rear hub. The Thun system, used on the special edition Alex Moulton Bentley bicycle, proved troublesome (typically the rear sprocket failed) and soon went out of production.
The Speed E flexible drive, created in the 1970s by Winfred Berg of East Rockaway, New York, had a belt that resembled a chain and a pair of stranded steel tension cables bridged by polyurethane buttons that engaged with the sprocket’s teeth. This lightweight system was extremely efficient when used on human-powered aircraft, such as the Gossamer Albatross, but it wasn’t sufficiently robust for use on bicycles (Wilson 2004, 324).
One problem with applying belt drive to most bicycle designs was the need to make a part of the frame removable. A drive belt couldn’t be split, as a chain could, so unless the bike had a cantilevered rear fork it was usually necessary to “break” the rear triangle to install or replace a belt, or even to change a tire or an inner tube. Typically a section of the seat stay had to be removed. Alternatively, a modified dropout, with a break between the chain stay and the seat stay, could be used. Such adaptations increased a bike’s cost and could result in weakness.
And despite many attempts by inventors (among them Barberon & Meunier) as early as 1869, no successful derailleur-style gear was marketed for a belt drive. Thus, the user was generally limited to a single-speed transmission, hub gears, bracket or chainwheel gears, or a combina-tion of these. In addicombina-tion, doubts remained about the durability of belt drives in stop-and-go city traffic—anecdotal evidence suggested that premature belt failure could occur. Thus, for the most part, belt drive remained a niche option offered by a few makers of portable and commuter bicycles.
In a recumbent with two-stage drive, a belt and a chain could be used together. About 40 years after he converted a Moulton to belt drive, David Gordon Wilson built a recumbent with belt drive from the cranks to the bottom bracket and a conventional final drive with a chain and a derailleur. There was no cycle-specific belt long enough, so he again used a standard industrial belt (a Browning Gearbelt 2400–8M.20).
Figure 4.4 Wilson’s recumbent, with Browning belt and conventional chain drive (David Gordon Wilson).