In case of a locomotive with short valve rod, the traditional practice is to place a knuckle joint in the rod to allow for additional flexibility. The diagram below shows that the valve rod must be able to move in vertical direction, because it’s connected to the rocker arm which moves in an arc.
So indeed the original CK Holliday drawing shows a knuckle joint just aft of the cylinder.
But in looking at the pictures of the “contemporary” CK Holliday, we see a different connection:
It looks like the flexible knuckle joint has been replaced with a rigid, male-female thread type connection; the male valve stem is screwed into the female valve rod, with a hex nut to tighten the connection. And it appears to be adjustable. Could this be used to allow the engineers to finely tune the valve slide’s travel, by shortening or lengthening the valve rod?
And is it even possible to have a rigid connection here?
Below is a graph of the rocker’s motion through 45 degrees of travel derived through geometry. Red graph is horizontal projection of the rocker’s arm, and the blue is the vertical.
From the valve slide motion study, the maximum amount that the valve needs to travel, in one direction, to allow full exposure of the steam port is 1.13in.
Using the graph, or directly solving, the rocker will need to rotate, in one direction, 7.84 degrees to allow the valve slide to travel the distance it needs. This also results in a vertical deflection of the rod in amount of 0.077in. That’s about 6.88% of the rod’s diameter, or L/401.
So, it certainly works, as the induced stress on the rod and the rocker is very minimal. But, still, it seems like an unnecessary stress.
The computer model needs that rotation degree of freedom to function properly. The above shows that the model is rigid.
Besides, I wanted to do the CK Holliday as it was intended anyway. This will bring the model much closer to the 1955 build. So, back to the knuckle joint!