Probe_basic Toolchanger

Tool changer, it is 3D printed in your pictures.

yes, it´s my prototype and all parts are 3D printed...if this toolchanger works fine i will mill all parts in alloy...

but now it´s time to configure this thing...

but i don´t now how. so i need much help to realize this.

i will make a deal with you... help me to get this to work fine and i will give you the complete projekt as an .step

Ok, thanks for the background.  As I mentioned there a quite a few ways to control a toolchanger, but let's settle on the hardware config before working on the control scheme.

The fundamental objective is to rotate the ATC platter to a specific location.  That must be done using a mechanical stop/index (of some sort) or by using a positioning system.  All the other functions of the ATC system are secondary - or at least there is some flexibility in how/what is implemented.

• Geneva wheel mechanism
  • Acts as a mechanical indexer
  • Can use a 'dumb' DC/AC motor
  • Can also use a stepper or servo in velocity or position mode
  • Likely requires more fabrication and mechanical fiddling than a planetary/belt drive
  • Requires much less programming modifications to Probe Basic
    • The "OEM" version of Probe Basic (PB from now on) is designed to control an ATC disc/platter/carousel using a fixed-speed reversable motor (DC or 3phase AC) and a geneva wheel mechanism.
• Planetary reducer (or belt drive, or similar)
  • Does not mechanically align the pocket with the spindle
    • Thus, requires a motor to be controlled in position mode, OR
    • Requires some other mechanism to ensure the pocket is aligned (like a pneumatic shot pin or detent at each pocket)
  • Must use stepper or servo motor (if no additional locking alignment mechanism present)
  • Likely requires less fabrication and mechanical fiddling
  • Requires more programming modifications to PB
    • My modified version of PB uses a stepper motor and a 5:1 planetary reducer to index the carousel.
• Other drive mechanism
  • If you want, you could have a 'one-way' carousel/platter like some lathe turrets.  These have a ratchet & pawl mechanism and you drive the turret in one direction past a stop and then the motor reverses against the stop for pocket alignment.
    • This could use a dumb motor or a step/servo
  • I'm sure there are other hardware configurations you can use, but the further you get from the "OEM" version of PB or my stepper w/planetary config, the less help I can be.

You mentioned that you are out of stepper motor outputs on your Mesa config.  As the motor system will likely drive (or be driven by) the above hardware choices...you have a few options:

• Geneva wheel or ratchet/pawl only:
  • Buy a DC or reversable AC motor and buy or fabricate a geneva wheel mechanism
• Planetary OR geneva wheel:
  • Buy another Mesa card to support an additional step/direction output
  • Ethercat step/servo drive and control it directly (not wired to Mesa)
  • Step/servo drive that has internal indexing functions and uses a digital input (or an array of inputs) to move a specific distance
  • Depending on your computer, you might have a parallel port and can use that to control an additional step/servo drive
  • Use an arduino (or similar) as an 'indexer'
    • LCNC outputs a digital on/off signal to the arduino
    • Arduino sends a pre-programmed number of pulses to the step/servo drive

That's a lot to chew on.  Once you've decided on a motor/drive & hardware scheme - or you have more questions about what to do - we can move on to the control side of things.
 

What Mesa board?
And thank you for the offer/deal, i would like to have the time to build a tool changer to experiment with and to test the carousel component with real hardware, and i have plenty of hardware so i can do stepper/position driven, DC motor driven, or whatever might be a good idea to try, also have a big box of sensors in all types and shapes.
Also, i would like to build a decent mill from all the cr@p... stuff i have in the shop, the HiWin and THK rails are getting rusty without ever being used!

Hello now i have a mesa 7i76e...

i see 2 options for me....

7i76e + 7i76 or 7i76e with Arduino

the second option is low cost. the first option will open more doors

i´ve seen the 7i76e is available now.

but for this prototype we can try it to do with arduino.

Ok, arduino it is.

I have no idea how to program or control an arduino, so you're on your own with that.

The stock PB control scheme is a little complicated, but the basic platter/carousel motion works like this:

• Toolchange.ngc
  • Do a bunch of stuff we don't need to worry about now, then
  • call M10
• M10
  • Determines how many pockets (steps) are needed to move from the current pocket to the destination
  • Determines which direction to turn the carousel to get to the desired pocket
  • Write the number of steps to move down in a parameter
    • If CCW, call M11 and pass the number of steps as an argument (#P)
    • If CW, call M12 with same P argument
• M11 (or M12) - Turn on a 'dumb' motor FWD (or REV)
  • While motor is turning, a sensor is triggered by a flag at each pocket
    • This sensor is connected to a digital input in LCNC
  • The M11 subroutine keeps track of how many pulses it sees from the pocket sensor
  • When the correct number of pocket pulses has been reached, turn off the motor
  • Write the new (destination) pocket number to a 'current pocket' parameter

This works because a geneva wheel permits the motor to overrun or underrun quite a bit while the pocket is locked in place by the mechanism.  Each time it's locked in an index position LCNC has time to read the pocket pulse from the sensor, M11 (or M12) has time to do the math and stop the motor.  If it rotates a little more than necessary that's fine - the pocket is still in position.

If you want to use an arduino and still use the stock PB control scheme then you may need to duplicate the motion of a geneva wheel.  That is, the arduino receives a 'ON/FWD' command and then:

• Rotates the stepper the number of pulses between pocket centers
• Stops the motor for a half second (or some time) while the pocket pulse is read by LCNC and M11 does the math
• If the ON/FWD command is not turned off by the end of dwell period, rotate another pocket and dwell
• Repeat until the ON/FWD command is turned off during the pause.

If you can program an arduino to do this you should be in good shape to move to some bench testing.

I suggest you get your prototype set up:

• a microswitch or proximity sensor and install a flag (screw, bent sheetmetal, whatever) on each pocket.
• A momentary switch for ON/FWD and another for ON/REV
• Arduino
  • Programmed to move the correct number of steps between pockets
    • example: 8 pockets, 1.8 degree stepper, no gearbox (for testing), 10x microsteps
      • 200 x 10 = 2000 steps/rev
      • 2000/8 pockets = 250 steps/pocket
  • Programmed to dwell a second after each pocket even if the ON switch is held down
• Stepper drive

Once you get that working, the next step will be to install a home sensor and a second flag (or a longer flag) on whatever pocket you want to be considered "home" or pocket #1.

Make sense?