[Nth]

Thanks for the replies. The motor is connected to the ballscrew via belt. My thinking is by putting the encoder on the screw it'll keep track of the actual table position even if there is lost steps or belt slippage under load. I'm new so maybe I'm wrong. Anyone see a problem with this?

[hanermo]

Well, it depends.
The first answer is don´t do it.
The second answer is yes it can work.

For an encoder to work at the screw, you must use and have built an industrial level backlash free mechanism.
This means position control belts (HTD etc), not movement belts (XL etc).
High tension (but your motor axels probably wont support this).
Adjustable mount for the motor to tension the belts correctly.
Very good alignment, so back and forth movement wont change the initial movement characteristics of the motion control system.

Servos dont have lost steps. Is this for steppers ?
If you are using steppers, and you only want to see the encoder positions in mach3, then by all means put them at the screw.
If you want to try and control a servo system pid loop, the mechanics must be tightly coupled.
This means HTD belts, 15 mm wide, or better, and above mentioned mounts etc.
High res encoders are better, less chatter.

But you cant run hi-res encoders without a SS or other hardware motion control device.
If it´s for EMC, you cant run hi-res encoders as the max speed in sw is around 45 kHz. Again, you need a hw control board, and it gets more complicated or expensive.

I built a hi-res servo belt transport using 10k encoders, and 3:1 transmission to screw.
Works excellent, better than 1 micron resolution.
But it is neither the easiest nor the cheapest solution.

Power ? Motors ? Resolution (microns desired, screw used, etc.) Sw ? usage (mill, lathe, grinder ?)
Everything affects it.
Building a PID transport is not hard. Building a poor one is easier. Flexible mounts will spoil it.
A servo PID is a function of the (number of components x max error of worst) squared.

Yes, you can drill (tap) the ends if you want. Easy. Wont hurt anything.

As I said, belts are more accurate than you think. Better than 1 micron for sure.
Unless you have a Hardinge HLV or similar, all other mechacanical components will be weaker than HTD belt resolution.

Thanks for the replies. The motor is connected to the ballscrew via belt. My thinking is by putting the encoder on the screw it'll keep track of the actual table position even if there is lost steps or belt slippage under load. I'm new so maybe I'm wrong. Anyone see a problem with this?