JLC 960

[Jack75]

Splines are one way but  simple interference between the two parts is sufficient. There are tables of interference fits for all sizes of shaft and holes. The overlap is measures in hundreds if a mm (can be measured with the JFK)

Here is an example that I had to deal with recently - the drive wheel of the JLC Cal 960. The little gear wheel at the bottom drives directly the ratchet wheel and is an interferece fit (press fit if you like) onto the arbor. The pivot is 0.3mm the arbor 0.53mm and the gear wheel outer diameter 1.34mm. The gear wheel takes all of the force needed to fully wind the mainspring. An astonishing design for a top end movement that is used in several fancy watches such as the JLC Reverso. Sometimes the gear splits and as it also holds the mainspring back through the automatic system to the only clic in the movement, when it fails even manual winding is out. But straight interference works most of the time for all of those JLC and IWC watches using  the Cal 960.

[Jack75]

Here is the gear train of the JLC Cal 960 automatic system, The clic that I mentioned is arrowed and the drive wheel is that with the three-prong spring that engages in the holes of the drive wheel for automatic winding