Welcome to Our Community

Some features disabled for guests. Register Today.

A-Axis Steps per Millimeter

Discussion in 'Control Software' started by Jeff Stuart, Dec 19, 2018.

  1. Jeff Stuart

    Jeff Stuart Well-Known
    Builder

    Joined:
    Apr 2, 2016
    Messages:
    4
    Likes Received:
    3
    Say you have a machine with a dual-driven axis (like the Y axis on my Lead), and the slave A motor needs a steps-per-mm different from the Y motor? I see no provision for that in software. Am I missing something?

    Thanks,
    Jeff
     
  2. David the swarfer

    David the swarfer OpenBuilds Team
    Staff Member Moderator Builder

    Joined:
    Aug 6, 2013
    Messages:
    1,090
    Likes Received:
    517
    this is impossible if you SLAVE A to Y, they are getting identical step signals.
    the 2 motors and drivers and leadscrews/belts must be identical in terms of steps/mm for this to work.

    The only way to make 2 different hardware systems work together like this is with a much more expensive controller that knows how to do this, or software like LinuxCNC which is free but a lot harder to set up.
     
  3. Jeff Stuart

    Jeff Stuart Well-Known
    Builder

    Joined:
    Apr 2, 2016
    Messages:
    4
    Likes Received:
    3
    Thank you David, that's what I suspected. As luck would have it, the motor I installed on A runs about 1% slower than the motor on Y. I swapped A with Z and that works pretty well. Given the xPro V3 runs the Y & A in parallel, that's as good as I will get with that setup.
     
  4. Spikeoz

    Spikeoz New
    Builder

    Joined:
    Jan 8, 2019
    Messages:
    1
    Likes Received:
    0
    Without knowing exactly how your machine is set up here a some observations from the big end of the world

    Master-slave relationships for motors often have speed imbalances , the master motor takes most of the load, the slave is their to provide support, when using 1000kw drives directly coupled to a conveying roller, the slave often falls behind the master slightly particularly as loads change, but at steady state the two are synced, but one draws more current that the other, it’s just the nature of the mechanics.

    Closer to the scale that we are dealing with, I’ve built a controlled lifting device, using 4 servo motors, 4 lead screws attached to 2m sq plate, leads in the middle of each plate (dumbest idea I’ve seen in awhile to be honest, I’d have placed the leads in the corners) the servos are all driven by the same signal, NOT a master slave relationship, it works fine except if the driven plate is not set up square to start with, if it’s squewed then one of the servos experinces higher load, and slows down, further tilting the plate to the point that it jams. It was a nightmare, ended up replacing the system with a geared system that is controlled from a single much larger servo, the leads are now all mechanically linked, I’m a mechanical guy I liked that set up, my off sider is an electrical guy and he hates it, he’s been playing around with the control software to solve the problem , he’s getting their.

    Take home, it gets complicated, but play to your stronger talents, and you can solve the problem maybe in a diffrent way
     
  5. Jeff Stuart

    Jeff Stuart Well-Known
    Builder

    Joined:
    Apr 2, 2016
    Messages:
    4
    Likes Received:
    3
    Spikeoz, thanks for the reply. 1000kW lordy no, I shouldn't be allowed anywhere near such a thing.

    In my case the Y and A motors have their own drivers, and when I built the machine I paid close attention to match the drag on each lead screw. Since I was testing the machine unloaded (without a spindle in fact) I doubt differential loading would account for lag on the A axis.

    I lucked out -- my Z motor matched the Y almost exactly so I swapped motors. All is well.

    I'm asking about software calibration because I'm surprised it's not a routine adjustment. Co-driven axes seem pretty common. If the "master" and "slave" don't move together you don't have a square machine, except for one value of Y.

    Since asking, I looked into the grbl code on Github, and yeah there's no provision to calibrate a "slave" axis. Software isn't the problem though, it's the timing on the microcontroller feeding the drivers. To the hardware it is a 4th axis.

    So, what David said, it can't be done with my setup. I'm not complaining, my machine works really well now that the motors match.
     
    David the swarfer likes this.

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice