[CNC-Joe]

Joe, It's an older system but it's using the Ethernet Smoothstepper, and several PMDX boards.
PMDX-126 Breakout board
PMDX-107 Isolated Speed board for VFD
PMDX-170 Optical sensor switches
PMDX-179 Remote IO LED board.

Running Servos and encoders for the 3 main axis, then steppers for the Rotary table, and for the tool changer carousel.


I did low air pressure, draw bar position, extend and retract limits for the slide, but besides a home switch for the tool carousel, I didn't do a position check after each change. I mean after all, I'm controlling it with a stepper so I should know where it is unless something bad happens. It's been running like that for years and I have the accel and decel set where it hadn't ever lost any steps.

If I were to do it, just thinking out loud. I would use an optical sensor the then make a disc with a hole for each tool position and I'd mount it to the tool spindle somehow. It would be an inexpensive encoder sort of that would be a safety check. Then a single sensor could tell if it move into position. It wouldn't know which position because that would be up to the software to increment or decrement a counter.

Better yet, an encoder or a servo and encoder. You can make it as simple or complex as you want. You get the idea. The other issue is if you're converting something already there, then you have to figure out mounting and such.

I hate to say it, but I've seen several lathes that use a stepper, they don't worry about the actual step position, but they use a spring loaded pawl, then they advance XX steps so they know they are past the pawl, then back up so many steps. They actually back up more than required and the stepper looses steps, but it doesn't matter. Then next time, same thing, advance past the pawl, then back up against the pawl.
Pretty simple and in the fashion they're doing it, other than keep track of a counter for which tool position, there isn't anything else to do.

It's like the stepper guys that home to a stop, the gantry hits a stop and the stepper skips beats. Then they zero that axis. Seriously, many commercial products do the same exact process.

----------------------------------------------------------
Hi Richard,
- That is exactly how I am doing it currently- CW rotation to move to the tool position, then CCW rotation to backup against the pawl and to lock into position.
- I then set the OEMDRO to the position that it should be (i.e. Tool 1= 0 Degrees, 2=60, 3=120, 4=180, 5=240, 6=360) - to make up for the missed steps of the collision with the pawl.
- I want to enable the pawl retraction - then I could have dual-direction moves to take the shortest route to the tool - but with only 6 tool positions - you're really not talking about a big time save.
- I need to incorporate some type of failsafe sensorics to know that the turret actually got to its intended position - and I'm still trying to figure out the best way to do that (any input is greatly appreciated here). Originally DynaMechtronics incorporated 3 sensors on the shaft of the turret - with LED emitter/detectors mounted opposite one another.. and used the binary code to failsafe position.

I need to expand the amount of I/O that I have to add additional failsafe sensors - any input or guidance on this would be highly appreciated too.

Thanks for any help that you can provide.

Joe