To summarise - The MERG design we were following constrained the servo arm movement to just that required to move the point tie-bar - approximately 2mm. A bit deflated with the units performance, Geoff described how a 'mainline' actuator swung through an arc of about 160 degrees. The resulting horizontal arm movement was not uniform, but slower at both ends. Light bulbs flashed and we wondered if we couldn't do something similar. We could easily reposition the micro-switches and use a single-ended servo 'horn'/arm.
His new arrangement had no loop to take up the slack, just a curve in the connecting wire. The servo swung nicely and closed the switches positively, with no stress - it was wonderful!
Note how the connecting wire is positioned in the very bottom hole on the servo arm, creating maximum 'thrust'.
Sorted!
(Geoff apologises for not turning the radio off)