The Walschaerts & The Radials

Content Tools

35640 Avenue F Yucaipa, California 92399

Upon several occasions in the past, the Walshaerts valve gear has been referred to as a radial type gear. This misnomer must be put to bed here and now. Let casual observers not be misled by the combination lever as required on this gear, and which same has also been applied to other types of valve gear. The Walschaerts is NOT a radial type gear, and was never claimed to be. This gear was designed primarily to afford accessibility and to avoid crawling down and under for servicing and lubrication. But it is regarded as a 'sloppy' gear, unlike all-wrist motion of a true radial gear. While I do not wish to castigate this design, well do I recall in the late 'Teens when I worked in the Case factory and the old railway roundhouse, there was carried in 'Railway Age' an advertisement by the Pilliod Company of New York, distributors for the Baker radial gear for locomotives, which depicted the outlay of a Walschaerts gear with the comment, in effect, 'This compromising mechanical monstrosity was forced upon the American locomotive industry!' And final attempts at uniform steam distribution in locomotives in this country included those offered by Franklin and Caprotti, which utilized gear-driven cams, operating poppet valves.

A true radial gear, by definition, is one in which a single eccentric (or crank) is strap-connected to a radius arm, the opposite or fulcrum end of which may be adjusted about a central pivotal point in order to affect the point of stroke steam cutoff and also control engine reversal, where the latter feature is incorporated. The principle valve motion is taken off at a median point of this linkage, or a point so actuated by it. Now, let us see why the combination lever on the Walschaerts modified eccentric type of gear is similar to that employed with radial gears, (this is composed of a conjugate arm from the combination link off the cross-head to the valve actuating rod).

In a conventional simple eccentric gear, wherein the valve is designed with lap (and if it does not have lap it will of necessity be very inefficient because of using steam full stroke, as on simple throttle-reversing hoisting engines commonly used on shipboard) the valve leads the crank position by approximately 120 degrees in the case of outside admission, common with D-valves, or trails it by the same amount with inside admission as common practice with piston valves.

Thus, if an engine is designed for reversible operation from a single fixed eccentric, this eccentric must of necessity be set effectively midway twixt both modes of operation; that is, 180 (or zero) degrees from crank position. Actual geometrical differences are accommodated through bell-crank linkages, etc. But a 180-degree setting, operating in a geometrical configuration 90 degrees above (or below) the crank-crosshead line of action would normally result in a relative eccentric angle of 90 (or 270) degrees with the crank. So a combination must be picked up through linkage with the crosshead, wherein the resultant of two angles, zero and 90 degrees for instance, is resolved as 45 degrees, the opposite end of a fulcrumed lever from which would result in a 135-degree displacement. This is satisfactory for the purpose when the valve lap and lead are designed accordingly. And from this portion of the discussion, it is apparent that this conjugate action applies to both radial and Walschaerts motions.

Now, since we have divorced the Walschaerts out of the radial family, it is of interest to consider the Joy radial valve gear as applied in some locomotive practice, the closest relative to traction engine design, since the primary gear takeoff power was from a pivotal point in the main connecting rod between the crosshead and the crankpin. This compares favorably with the Wolff and similar radial gears. However, in many traction versions the second radial (swing-able) arm was replaced by a sliding block. While most of these blocks were straight, Case did, upon later occasions, resort to a curved block to simulate radius action. In these designs, the combination or conjugate action (radial and crank) was taken off at a point between the eccentric and the block. Another refinement to such radial gears was incorporated by Reeves, probably the ultimate in this respect, by adding a compensation to equalize one end of the cylinder against the other (due to the angularity of the connecting rod, as also to that of the valve motion as it is practically impossible to balance one motion against the other in an out-of-phase displacement). In this case (not Case) Reeves took advantage of the property of a discordant bell crank linkage, wherein the takeoff speed of travel of one leg of the bell is not the same as in the other leg.