Two steps forward, one step back:

After backing off the pawl screw, there wasn't enough tension to hold it in place, so the clack of the turret cam moving round unscrewed it.

Removing some of the springs and tightening it up fully solved the unscrewing problem, but then I found the hole for the screw breaks into the operating area of the cam, so when the screw was fully tightened, the cam caught on the end of the screw.

By adding a brass washer, and adjusting the layout of the disc springs, I was able to get the tension right, and the screw can't unscrew any more as it's now flush with the back face of the turret block, so held in position by the X axis block.

Next, I found that, even after adjusting the tension of the spring, the turret was still stalling at various points. Having tried all sorts of cures, I've had to back off the pre-load on the main cam shaft bearings. Originally, I was too tight to turn by hand. Having removed the shaft, I can't feel any roughness in the bearings themselves, which are BR-7005-2RS angular contact bearings, so I reassembled it with just a smidgen less pre-load than before. It's still tight to turn, and there's no appreicable play or end float, but it now turns on command without stalling.

A little work on the gcode and it looks like an 8.75 CCW turn followed by a 1.75 CW turn is all that is needed to rotate the turret one tool position and lock it back into place.

Hopefully, the 5mm dowels should arrive tomorrow, so I can get the turret aligned and re-installed.

I've already written some simple gcode scripts to make various brass axles. Now the auto-changer is working, I should be able to combine some of them.

There's no homing facility on the turret, so I'll need to make sure that lathe is set to tool 1 at the start of each program.

Adrian.