On Thursday 30 April 2020 19:25:16 Leonardo Marsaglia wrote: > > However if that Yaskawa actually is counting steps, something I've > > never ever seen one of the clones I am using do, then you are > > probably just sending too many steps/second. I could be wrong, as > > its not impossible, but I'm thinking you may have miss-read the > > Chinglesh in the manual. But do check the polarity of the signal as > > it arrives at the vfd. In that event, switching to a pwmgen, > > running in pdm mode, may be the fix you need. > > Indeed it sounded strange to me too, and when we purchased the boards > to make the conversion we planned to drive the VFD with analog voltage > like usual but when I discovered that feature a few months ago I > decided to give it a try and save a 7i33 for later. It's working > pretty well now and I tuned it better today. It accepts frequencies > from 0.5 khz to 32 khz. We are peaking it at about 24000 khz but the > problem I had was the step lenght was to low for the driver to "see" > the pulses. Now I only need to filter the encoder from the spindle to > get rid of that noisy output wich is not hurting anything but it would > look better filtered.
I've had that noise come back thru the servo driver strong enough to hammer the backlash and make it sound like all the ball bearings in the mills head had suddenly worn square. A higher res encoder stopped all that noise. I can also run 10x the Pgain w/o any instability. Sweet! > On other thing, css is scaled from x=0, if its not 0 as touched off at > > > the exact center of the workpiece, then css is going to be > > miss-applied. It does have gotcha's in that regard. > > Yes, I double checked that today and it was ok. My problem was the > output source of the commanded velocity for the spindle. I spent > almost 3 hours trying to figure out what was the problem and all it > took to solve it was to change a signal name in the hal file :(. > Great, but from what to what, so the fix becomes part of the list archive? > El jue., 30 abr. 2020 a las 14:41, Gene Heskett > (<[email protected]>) > > escribió: > > On Thursday 30 April 2020 12:36:00 Leonardo Marsaglia wrote: > > > Just to be super clear. I'm sending the spindle speed cmd velocity > > > to a step generator in velocity mode. Could this be the source of > > > the problem? Or does LCNC internally set the speed of the spindle > > > based on the X diameter? > > > > I must say thats an unconventional mode to me. I know nothing about > > your Yaskawa, but as a CET, one thought keeps running thru my mind. > > Most vfd's will do a decent job of averaging a digital command > > signal, giving about 50% speed for a signal thats high 50% of the > > time. We often set steppers up for a true=logic low, and feed the + > > line of the driver at 5 volts, and the step signal to the > > corresponding - terminal at the driver input. > > > > primarily because our interface electronics can sink lots more > > current in the logic low state than they can source in the logic > > high state, its the nature of TTL circuitry and the capability ratio > > can be 10/1. So an output that can sink 24 mills is a much better > > opto driver than one that can only source 2.4 milliamps. If you > > feed that "inverted" signal into most vfd's, they would interpret > > that 98% of the time logic high which you are sending it for a low > > speed command, the vfd will treat it as if its a high 98% of the > > time resulting in a high speed command actually being done. You > > should be able to verify that with the halscope. Or most any scope > > with sufficient bandwidth to show the steps as a well squared wave. > > If you are seeing a triangular wave, the scope isn't fast enough. > > > > However if that Yaskawa actually is counting steps, something I've > > never ever seen one of the clones I am using do, then you are > > probably just sending too many steps/second. I could be wrong, as > > its not impossible, but I'm thinking you may have miss-read the > > Chinglesh in the manual. But do check the polarity of the signal as > > it arrives at the vfd. In that event, switching to a pwmgen, > > running in pdm mode, may be the fix you need. > > > > On other thing, css is scaled from x=0, if its not 0 as touched off > > at the exact center of the workpiece, then css is going to be > > miss-applied. It does have gotcha's in that regard. > > [...] > > > > Cheers, Gene Heskett > > -- > > "There are four boxes to be used in defense of liberty: > > soap, ballot, jury, and ammo. Please use in that order." > > -Ed Howdershelt (Author) > > If we desire respect for the law, we must first make the law > > respectable. - Louis D. Brandeis > > Genes Web page <http://geneslinuxbox.net:6309/gene> > > > > > > _______________________________________________ > > Emc-users mailing list > > [email protected] > > https://lists.sourceforge.net/lists/listinfo/emc-users > > _______________________________________________ > Emc-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/emc-users Cheers, Gene Heskett -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) If we desire respect for the law, we must first make the law respectable. - Louis D. Brandeis Genes Web page <http://geneslinuxbox.net:6309/gene> _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
