Re: [Machinekit] Re: Machinekit and closed loop

['schoone...@btinternet.com' via Machinekit]

Now we know what you are doing, it is easier to assist. You really need to ask a question about your problem, not how to effect the solution you have thought of. The usual way to test backlash with a mill / lathe is to use a dial gauge and move the bed a distance within the gauges range. No need for a closed loop stepper. May take several bites, but backlash should be taken up by the initial movement in either direction and be pretty constant. If you are getting a loss of position at a particular point in the bed travel aside from the initial backlash, that could be caused by binding, flexing or similar. Insufficiently robust, loose or out of square / out of parallel, rails, screws or many other mechanical components could be to blame. A very common one is where the bed runs on 2 rails that are not perfectly parallel and there is a pinch point at one or even both ends. Another is with Chinese castings and gib strips that have not been scraped and mated properly ( I bought one once where the gib strips had obviously been 'finished' with an angle grinder!) The ini file has facility to read a compensation file for backlash on a per axis basis. http://www.machinekit.io/docs/config/ini_config/#axis_-num-section-a-id-sub-axis-section-a That should take care of backlash, which can differ in opposing directions, especially on the Z axis where head mass comes into play and may pre-load some moves and play catch-up on others. Have a good search on their forum first, you won't be the first or last person with a particular problem, it is defining what the problem is that is key. Good luck. On 16/02/18 15:23, pasteur wrote: Ok thank you guys, it was very useful all the data that you provided to me.  Maybe I'm gonna to search or post in LinuxCNC forums...  I read very interested the PID algorithm , but I'm using steppers motors and it isnt too eficient I guess, for my problem. My problem is simple I need to create a close loop because the backlash on my machine it isn't equal in all points and I need to measure and correct the errors... I'm gonna to search in LinuxCNC forums and  post a question if it is necessary  Thank you a lot. -- website: http://www.machinekit.io blog: http://blog.machinekit.io github: https://github.com/machinekit --- You received this message because you are subscribed to the Google Groups "Machinekit" group. To unsubscribe from this group and stop receiving emails from it, send an email to machinekit+unsubscr...@googlegroups.com. Visit this group at https://groups.google.com/group/machinekit. For more options, visit https://groups.google.com/d/optout. -- website: http://www.machinekit.io blog: http://blog.machinekit.io github: https://github.com/machinekit --- You received this message because you are subscribed to the Google Groups "Machinekit" group. To unsubscribe from this group and stop receiving emails from it, send an email to machinekit+unsubscr...@googlegroups.com. Visit this group at https://groups.google.com/group/machinekit. For more options, visit https://groups.google.com/d/optout.