[WoodSnarfer]

I bought brand new servos off eBay, and they came with encoders already mounted. Encoders are 1000 CPR, and my design calls for using 5 TPI screws. I want to jog at around 175 IPM.

Am I figuring this right:

- One revolution of the screw = 1000 * 4 = 4000 pulses
- Five revolutions of screw, or 1 inch of travel = 20000 pulses
- 175 IPM jog speed means 58,333 pulses per second out of the parallel port (i.e. 175 / 60 * 20000)

Now, after I did that math I figure, okay, the parallel port needs to support a 60Khz clock rate. Even older PCs have p-ports sitting on an 8Mhz bus -- so there should be plenty of headroom there, right?

I understand that data throughput might never reach the actual bus speed, but even it needs 4 cycles per bit, plus a couple of wait states, the speed out of the p-port ought to be way higher than required...right?

Or is there some other limiting factor (maybe in software?) that will cause my computer to not to be able to drive the motors with these 1000 CPR encoders at the speed I desire?

Thanks,
Chris

(EDIT: I forgot about hardware IRQs, and the fact that a machine running Windows is going to have a *lot* of other housekeeping to do - but a newer machine with an ECP p-port ought to be plenty fast...I hope! Anyone with any real-world data or experience about this, I'd love to hear from you.)

[ger21]

Mach2 will give you 45,000 steps/second from the parallel port, with a ~ 1Ghz computer. There is nothing faster, using a parallel port.

I just remembered, there is something faster. http://www.abilitysystems.com/indexer.html
But it's $700 vs $150 for Mach2.

You could use DeskCNC, which uses the serial port and an external pulse generator, which can output 125,000 steps/sec.