Quote Originally Posted by russell View Post
considerable tilting force on the head. This must be countered by forces on the dovetails which will tend to lock things up.
That conclusion sounds plausible to me. A real life example is if I put my very heavy milling vice on the end of the table, so it's outside the area of contact of the dovetails then I can't get as high a feedrate as when it's in the middle.

However you've got the formula slightly wrong...

Quote Originally Posted by russell View Post
What torque is needed to raise the weight of the head? Torque is given by:

Torque = (W x g x P)/2 x π x η)
where W = weight of head
g = acceleration due to gravity
P = leadscrew pitch
η = leadscrew efficiency
Weight is measured in Newtons (W=mg if you like), so that should be mass in the formula, and you missed a bracket, i.e.:

Torque = (m x g x P)/(2 x π x η)
Where m=mass

So that implies, using your numbers T=(18*9.81*0.005)/(2*pi*0.5)=0.28Nm
(You seem to have used mass anyway in the formula but dropped one of the 2's -somewhere)

0.28Nm sounds good, and if you're using a ballscrew the efficiency is more like 90% ... but of course that completely ignores friction of the slides and acceleration of the head and screw. Acceleration of the head is easy, instead of lifting just m*g it's now F=ma+mg, so substitute that in and you get:

Torque = (m*P)(a+g)/(2 x π x η)

Not sure what's reasonable to expect for acelleration here, say 0.5ms^-2? That makes the torque 0.30Nm ... i.e. negligible difference. Can include the acceleration of the screw by knowing it's length and diameter, and thus it's moment of inertia ... but again that wont make much difference here.

To work out the friction on the dovetail slide we need the coefficient of friction of the slide ... about 0.2, and the contact force upon it which is not straightforward as we don't know where the centre of mass of the head is among other things. It will be proportional to that distance, so yes minimise it if you can. I'll think about it.
Ideally we'd work that out, then look up the speed/torque curve of a selection of motors and get one that has the calculated torque (plus a bit to make sure) at the required feedrate. If that doesn't get what you want then use pulleys to change the ratio.

I'd probably just try the motor you're going to use for X/Y (probably a 3Nm Nema 24?) and see what happens ... then you know which to by and only waste a small amount on postage.