Z axis backlash adjustment

After one of those days where you seem to have been busy but don't seem to actually achieved much, followed by a couple of hours A Level maths tuition, I thought I'd better get something done before the day is done.

Tried in vain to understand how the tool (length) table works by vainly stabbing at the keypads. I may have to reread that section in Smid to understand how it should work, then try to figure out what the manual might be trying to say. Sod it.

I also have a 3m length of 1.4m wide, 3mm thick black neoprene on order from an ebay trader. That will look quite exciting, covered in lubricant, fnnnaaaaarrr! Actually it will replace the missing lubricant / coolant / swarf curtain that used to live at the back of the machine until it perished and was chucked out by the previous owner according to the note he left on the framework. That will need to go on soon, along with the giant chip tray. Ideally I would take delivery of it first, mind.

Thought I'd get on with a little(?) job in the meantime. I was a bit pissed off to find that there was measurable backlash on the spindle movement (Z axis) when I was tuning the servos. Backlash creates a non-linear response in the load characteristic, so I suspect that is what results in the grumbling from the Z axis servo when at rest. The system gain will be significantly higher when the only load on the motor is the tiny inertia of the motor and Z axis ballscrew. Without backlash, any movement of the motor would result in the whole quill moving and it's got a lot more inertia. That's the cunning stunt anyway.

Looking at the section drawing of the head, there appears to be a double ballnut. If so, this arrangement allows backlash adjustment by adjusting the thickness of a shim between the 2 nuts. The shim forces the nuts apart, removing any backlash. In a simple world I'd have a fairly good idea how much thicker the shim needs to be, if there is one - not far off the measured backlash figure of 50um (0.002" in old money). Bang an extra shim in of that thickness and Bob's your auntie - bishy bashy bollocks. May not be as simple as that of course....

Disconnecting the ballnut(s) from the yoke is a matter of undoing four M5 machine screws. The ballnut assembly is by NSK (most of the ballscrews and bearings on this machine are) and truly is beautifully made. 

Sure enough, there is a (beautiful) ground 2-piece shim held in place by 2 keepers that fit into the locking slots on each side very precisely. There is even a tapped hole for an extraction screw, as it is a pretty tight fit....

With a pitch of 5mm, the 50um backlash would equate to about 3.6 degrees relative movement of the ballnuts. I'm pretty certain I am not seeing anything like as big an angle as that when I tighten the ballnuts up against each other. This suggests that either my original backlash figure is wrong or (more likely) there are other elements contributing to the total figure. The prime suspect would be the thrust bearing on the top of the ballscrew but there could be others. I need to do some proper, careful measurements with life-sized loadings before getting busy with shims....