I think using the index pulse will not detect "out of square". The reason is that you get out of square because of a "skipped step" on only one of the motors. So your idea (I hate to say) works great only in cases where it is not needed. If the router never skips a step only one switch is needed.
The reason to add the second switch is to detect the out of square condition and if you mask the error by using index pulses to defeat the purpose of using two switches. You will NEVER get two mechanical switches to 100% agree unless you use a special home algorithm that first finds approximately home by running fast into the switches then backs off and single steps slowly, slowing settling time after each step. Even then you need to have the software "remember" the difference in step counts, If that ever changes then you are "out of square" and you sound some alarm so the operator can fix it. But to find home you need one one switch and use the second switch only to check if something moved. Note the Prussia I3 printers and all the Prussia clones use a dual lead screw on the z-axis. This design might be the most common kind of drive. There are two parallel screws with 8mm pitch. The motors do 800 specs per rev and so have a resolution of 0.01 mm and it is very reliable and stays solid on square for months. As long as you don't try and step to fast they stay in sync "forever" BTW the method used to square the system is easy, turn off power and rotate one lead screw by hand and take measurements. On Wed, Oct 3, 2018 at 3:02 AM Leonardo Marsaglia <ldmarsag...@gmail.com> wrote: > Hello to all! > > I'm building a CNC router for wood machining for a friend of mine and the > best way I found to drive the Y joint for this particular design, given the > size of the machine, is the following: > > One rack and pinion on each side of the longitudinal axis of the machine > and each pinion conected to a servo motor using two steps of reduction with > synchronous belts to achieve a 10 to 1 ratio. > > I've found several topics on the forum talking about the homing of an axis > arranged like this. I guess to have screw regulated home switches for each > Y joint is almost a must (in case you don't use linear scales). But I was > wondering if it's possible to use ,in conjunction with that, the index > pulse of an encoder coupled to directly to the pinion. > > My idea is to offset the index pulse on each encoder via HAL to make both > sides of the Y joint trip the index pulse together and stay squared during > homing. Is this a good practice? > > About how to drive both Y joints as one axis: I've read that there's a way > of simply adding two Y joints for the Y axis in the 2.8 master branch but I > don't know if there's documentation available already. > > But I was thinking about slaving one of the motors (using them as pulse and > direction at the beginning to make things easier) and use the encoder on > the pinion only for following error and homing. I don't like the open loop > approach a lot, but I don't know if it's that easy to use them as servos in > position mode without having too much trouble. > > Any thoughts? > > Thank you as always! > > Leonardo > > _______________________________________________ > Emc-users mailing list > Emc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-users > -- Chris Albertson Redondo Beach, California _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
