2013/5/24 Jon Elson <el...@pico-systems.com> > Dave wrote: > > I did a gantry system with LinuxCNC a few years back. The gantry was > > very stiff and self squaring when powered off. The gantry was 10-12 > > feet across and very heavy. It used used 2 - 1 KW servo drives driving > > two ball screws on each end of the gantry for the Y axis. In order to > > keep things simple I used step and direction on the servo drives and fed > > the same signals to both of the Y axis servo drives. > Thanks for the info. This machine has servo motors on each end of the > gantry, > so it really needs LinuxCNC in the loop for BOTH motor/encoder units. > What you did works great for step/dir drives, but won't work for servos. > > When LinuxCNC starts, the encoder counters are zeroed. When it is moving > toward the home switch, it is going to keep them synched the same way > as at startup. But, when it arrives at home, the first axis' counter is > going to get zeroed. At this point, something special has to happen to > prevent > the two motors from diverging. Assuming it has moved 10000 counts > from startup to the home switch, the first motor to find the index mark > on its encoder will suddenly have the encoder count jump from 10000 > to zero. I'm not sure what the other motor will be following at that > instant. >
I have been following this thread and it seems to me that I have missed something. Last week I finally made some chips on my self-built router with servos. The construction is not most rigid, but it squares itself within 10 mm or so, the gantry is 3 m long. I use gantrykins, homing simple - each joint homes to its own switch (I have small inductive proximity switches). The main thing is to set up homing sequence and set all the search and latch and final velocities to be the same. The way it works - both joints start their homing moves simultaneously, move at the same speed, so gantry is not racked and, as much as I can see on screen, it seems that both joints meet their homeswitches pretty much at the same time. I have pretty slow home search velocity (and latch velocity even slower), so that is why I already have G0 G53 X10 Y10 in my MDI history to bring machine back to "almost home". I recall that in first posts there was mentioned something about homing to index, but I really do not see a point for that, so I somehow do not understand, where is the problem. Andy Pugh last autumn shared a way to add a new HAL pin to Axis GUI to switch machine to world mode, so I have it connected to axis.n.is-homed pins through and2 components and it works pretty nicely for me. You could use that to make sure that operator will not start jogging the machine in joint mode, once it has been homed. Viesturs If you can't fix it, you don't own it. http://www.ifixit.com/Manifesto ------------------------------------------------------------------------------ Try New Relic Now & We'll Send You this Cool Shirt New Relic is the only SaaS-based application performance monitoring service that delivers powerful full stack analytics. Optimize and monitor your browser, app, & servers with just a few lines of code. Try New Relic and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
