One of my theories has been that Percy’s loose connecting rods were more forgiving of minor perturbations in quartering or crankpin eccentricity. So, I printed the middle experiments (8 degrees out of quarter and .022″ off eccentricity) with the thicker crankpins that allow only minimal clearance. These are about 20% bigger than the original crankpins.
I was surprised to find that while there was a slight hesitation, the bigger crankpins still did not bind. They were not much different than the standard crankpins.
Well, when the facts don’t conform with the theory, you need a new theory! My latest conjecture is that the relative amount of slop is important. Perhaps 10 degrees out of quarter does not make such a big difference when the crankpins are large. Both #622 and #10 have almost scale-size crankpins, and so, this theory was worth exploring too.
Unfortunately, this experiment required new connecting rods. I followed the time-honoured tradition of soldering two layers of nickel silver together. I then glued one of Percy’s rods to the sandwich with CA. A little custom-turned bushing helped to guide the drill bit into the centres of the two original holes, and then I used CA debonder to remove the original rod. I filed the blank (very) roughly to shape before separating the two rods, and cleaning up all the burs.
I then printed a set of good wheels with 0.9mm crankpins to fit inside the 0.97 mm holes in the new connecting rods. These plopped right in, and Percy had no difficulty trundling down the test track, confirming that the new rods are a close-enough match to the old rods. We’ll have to wait until the printer finishes some new bad drivers to find out if my hypothesis is correct, however.