Both the L & Y and the LNWR used Joy gear quite extensively for a period and the L & Y in particular sought diligently to remove the inherent distortions in the driveline. No current simulator can deal with these peculiarities adequately: curved guides did not reflect the radius rod (technically not so), guides did not sit square to the driveline and distortions were created through the anchor link – all in order to appease imperfections. In addition, the weighshaft position was so critical to the production of good events that the L & Y used temporary erection on each locomotive in order to assess the best permanent location, before finally bolting into position. This sensitivity, even with the short valve travels of the day, is clearly suggestive of the battle with distortions unhelpful in procuring the excellent distribution of the primary gears.
Both Gooch and Allan straight link gears are logical descendants of Stephenson’s gear. The former obviates the necessity to lift the heavy gear to reverse motion but is otherwise flawed. The principal culprit lies in the lifting arrangement, as both lifting link and lifting arm find difficulty in properly serving both forward and reverse gears.
Allan gear employs a straight link slot and can reduce any tendency to excessive lead variation but otherwise introduces more complication. No lever system to create straight line motion will bear close mathematical scrutiny and Allan gear is no exception. The short lifting arms attempting to produce this action by ratio are ill-fitted to suspend the valve gear, through two conflicting lifting links, in any decent manner. The common result of manipulation to satisfy a reasonable lead pattern hampers the ability to secure good cut off equality and a decent fore gear often leaves back gear wanting.
Baker valve gear, originally designed for traction engines, enjoyed some success in its home environment, America. It introduces angular elements of Marshall gear into Walschaerts’ in a complex yoke drive with a bell crank. Apart from the support given by the Pilliod Company any advantages are more apparent than real, though doubtless the Company would not agree. Simulation confirms its complexity in arriving at valve events predictably similar to Walschaerts’ yet more costly to produce.
Countless gear inventions merely sought to bypass patents and can be discounted in practical terms unless the student has a particular interest. Unprotected by patents, Stephenson’s gear, properly designed, can be made to produce equality second to none and Walschaerts’ can approach this performance. Their supremacy was never destined to be endangered. A most interesting learned Paper read before the Institute by T.H.Shields in 1943 has been reprinted by TEE Publishing and appears in the DOWNLOADS page. It is highly recommended to the historical student, yet is not technically a treatise in depth.
It seems odd that elements like the Marshall lever arms, Baker yokes and bell cranks, Greenly offsets and many other attempts to allow correction have not been generally perceived as what they really are. The timing distortions of rotary to linear motion and vice versa are similarly present in all swinging arms and rockers - an arc is simply part of a circle and the same principle applies to both. It is therefore not surprising to find complex mechanisms little better than Hackworth gear in its simplest form. It is wise to include the members of suspension and reversal as legitimate constituents of the mechanism with the same ability to distort distribution.