Why not use three cylinders to improve balancing?

The following answer is taken from Wardale’s response to an unpublished letter from Bryan Attewell written in response to Wardale’s two-part article “Steaming into the Future” published in Steam Railway magazines #272 and #273:

“Mr Attewell criticises the choice of 2 cylinders for high speed, but the 2-cylinder locomotive cannot be simply condemned without fully exploring its possibilities. The acceptability of a 2-cylinder engine hinges on the issue of balancing, as I have acknowledged in my article. Although Mr. Attewell invokes Newton’s laws to support his case, it is in fact these same laws which show us how a 2-cylinder engine can be made acceptable. Newton tells us that force = mass x acceleration, so (i) the reciprocating masses cause forces and (ii) these forces act on a mass (the locomotive) to produce accelerations (i.e. vibrations). Considering item (i), force is proportional to the mass of the reciprocating parts, which in the BR 5MT are far from being as light as they should be (notwithstanding the LNER type crosshead, a design which is not amenable to lighter construction by using aluminium). A reasonable yardstick is the mass of the reciprocating parts per ton of piston thrust, so let us consider the following:

This illustrates the extent to which the reciprocating parts can be lightened, correspondingly reducing the forces they exert when in motion. But this is only half the story, the other half being the engine-tender drawgear. Using an American-type coupling incorporating a friction-damped radial buffer allows the tender mass to be added to that of the engine in absorbing fore-and-aft forces, which from Newton’s equation proportionally reduces the resultant accelerations (vibrations) of the engine, which is the important parameter. Put simply, if we can roughly halve the reciprocating masses and double the mass of the locomotive resisting fore-and-aft forces (by incorporating the mass of the tender) then the balancing problem is solved, eliminating the objections to a 2-cylinder machine for high speed. The key is optimum design of the reciprocating parts, and it can be revealed that the pistons, piston rings, piston rods and piston tail rods for the 5AT are already fully specified, the total mass of this assembly being only 200 lb., with a stress-based fatigue life equal to the full expected life of the locomotive. It would therefore appear that the target of 550 lb. for the total mass of the reciprocating parts per cylinder is going to be bettered.

Mr. Attewell’s other arguments in support of 3 cylinders are rather unconvincing. Adhesion is a complex subject – suffice it to say that there is little practical evidence (especially in Britain!) to suggest that 3-cylinder engines have superior adhesion. And at the kind of running speeds the 5AT would normally be used at, the damping of combustion gas flow through the boiler tubes produces fairly uniform draught on the fire whatever the number of cylinders, as I can attest from having taken thousands of draught readings. Of much greater importance is to dampen the draught peaks due to over-rapid release of steam from the cylinders, and I would remind Mr. Attewell that this problem destroyed the performance of the 3-cylinder “Duke of Cloucester”, but not that of the 2-cylinder Caprotti Class 5’s!