Linear-motion question

[grooveguy]

I have seen a few homebrew lathe videos utilizing a timing belt between two pulleys to effect lateral motion of the cutterhead, rather than the time-honored (and expensive!) feedscrew, its bearings, halfnuts, etc. Does anyone have practical experience using a belt or chain, especially with regard to the precision of groove spacing and freedom from patterning? Thanks.

[opcode66]

The pulleys are there for gear reduction. Servo motors are the only ones that are cheap and can turn very slowly with much torque. BLDC motors don't turn slowly unless they have a very high number of poles which makes them very expensive. So, most people use DC or BLDC motors with gear reductions. The reason servos are not used for linear traverse mechanisms is that they move in small steps. Smooth, fluid rotary motion is not possible with Servo. Any jerking whatsoever of the lateral traverse will result in noise on playback. A jerk is a dogleg in a groove which the stylus will trace and therefore will be audible.

Using a belt instead of a drive screw is the same reasoning. You can be much more precise with movements using a linear gear and screw combination. The belt itself will have some give and presumably whatever you attach to the belt to hold your head will have some weight. Weight and give combine to result in inaccuracies of travel that will not appear with a leadscrew carrying the same weight and movement.

Lets not forget the range that we need to be accurate within. That really determines the system. Groove to groove spacing can come down to 0.5 Mil (0.5 or 1/1000"). Inter groove spacing (space between groove floors) can come down to 3.5 Mils. Those are very tight tolerances to work with and be accurate and fluid within.

[grooveguy]

I think you're spot-on regarding most points, Opcode, although I have had luck 'sine-wave-microstepping' servo motors, albeit with a rubber belt between the motor and leadscrew pulley. Thanks for the input.

[opcode66]

I'm familiar with microstepping. Unless you actually physically plotted the movememt of the microstepped servo motor, and then analyzed the plot at 200x magnification, then you really can't make the claim that the motor moves in a completely fluid manor. I can tell you, they don't.

Sine wave or square wave. Doesn't matter. The motor controller is looking at rising edge until it meets a threshold.