Hi Edward,

I promised some diagrams to show the difference in the two valve gears, though of course nothing is quantified. The first illustration is the Savage gear.

Top is full forward gear and bottom is mid gear. Extreme left shows the leads, which in Stephenson’s gear increase towards mid gear, and since lead becomes a greater part of the whole valve travel on notching up this also increases the preadmission.

The two pairs of curves falling down to 40% and 60% are the exhaust lines. That leaves two curves of cut offs starting upper right and they fall down to below 20%.

The mismatch between fore and back strokes can also be shown by the complete family of indicator diagrams. Following the red line, say, shows a large discrepancy between the work done on one side of the piston to the other.

My solution – quite dramatic equality on paper but still considered rather poor by railway locomotive standards, though fairly typical of locomotives designed with the one-sided suspension of the expansion link pre 1900.

One would hardly think that those ‘minimal mods’ I tweaked out of Little Samson could produce such an improvement, yet as you rightly point out the original appears to be satisfactory. Now let’s see what a proper central suspension can achieve.

This is a GWR locomotive, showing less than 1% inequality anywhere in the quadrant from full forward to full reverse gear. It is achieved by pitching the errors caused by conversion of rotary to linear motion on one stroke against those on the other stroke.

It also shows that of all the hundreds of patents taken out in the 19^th century for valve gears and ‘improvements’ all were vastly inferior to the Stephenson’s gear minor tweaks which the better designers knew but never published.

The major culprits of poor events lie in the suspension on two scores. Firstly, Stephenson’s gear can be regarded as two gears – one for forward and the other for reverse. Logical symmetry would obviously expect a central suspension pin to match. Secondly, traction engine draughtsmen seem always to have simply drawn symmetrical lines for the lifting arm and link, as though these did not constitute part of the valve gear. What a mistake!

I hope that you find this of interest without getting into deep technicality. It is likely that a career draughtsman might never be called upon to design a gear from scratch more than once or twice in his working life. There I have the advantage, purely as a hobbyist.