Without a doubt the newer technology is nothing short of amazing.? I based my ELS on the paper:

Electronic Lead Screw Drive

Fritz.Linck@... (Germany)

?

Implemented in 1999 and 2000

Documentation completed March 1, 2001, Revised Dec. 5, 2005

?

His was all done in hardware with digital counters and multiply/divide logic.? Fascinating approach.? I started doing it that way with an FPGA so all the logic would be in one chip.? But eventually I went with the PIC18F family after an initial false start with a TI Digital Signal processor that required too high a development tool cost.

?

The approach used by Mr. Linck was to read a pair of rotary switches that created a multiplier/divisor combination.? So he could multiply the number of encoder lines from the spindle by X and divide that by the other switch value Y.? The Leadscew motor was a DC motor with an encoder and he used the difference between the leadscrew motor encoder counter and the X/Y spindle counter as an error term.? The larger the error the faster he turned the leadscrew.? As the leadscrew caught up to the spindle speed ratio the error dropped to zero and the motor speed stayed stable.

?

He could now treat the leadscrew as coupled through a gear chain and engage the half nut using the threading dial indicator to choose when to engage.? ?I'm not sure how well it worked for stopping the spindle, reversing it and running carriage backwards before restarting in forward. ?For imperial threads on a metric lathe.? I don't think it could do it.?

?

And certainly not bore to depth repeatedly.

?

But with a faster processor or a dual processor core, one could create that ratio closed loop motion that also tracked position. ?I suspect that is what LinuxCNC does since it needs multi-line quadrature encoder on the spindle and can run Servo motors on the lead screw. ??It doesn't suffer from the RELS limitations.

?

John Dammeyer

?

?

?

?