Gene, Okuma does it based on... I think one sensor near the spindle nose. They are the only ones I know of building it in to a production machine, and the reviews are good. Detecting the emergence of chatter is wildly obvious when you look at the frequencies. I've even checked out audio of chatter and it's really easy to pick out in the frequency plots. I think it'd be doable in software.
Once it's detected there's a very simple algebraic expression from the current chattering spindle speed to get one that doesn't. My only concern is that you would also have to monitor load on the spindle in case dropping the spindle speed overloaded the spindle, requiring reducing feed until the spindle load dropped to a reasonable value. Jared On Thu, Apr 16, 2020, 5:42 PM Gene Heskett <[email protected]> wrote: > On Thursday 16 April 2020 16:46:59 Jared McLaughlin wrote: > > > Gene, > > > > Specifically I think it would be interesting to make things like > > spindle speed a "we will try to get there, unless ..." instead of a > > PID loop sort of behavior. Think of how you might implement something > > like Okuma's "Machining Navi" where it takes the spindle speed, > > monitors vibrations and adjusts the spindle speed to a more > > appropriate speed based on feedback. Basically, a lot of machining has > > zero feedback loops outside of "did we get to the desired spindle > > speed and x,y,z location"? Things like chatter are probably best > > controlled in the controller (as in the Okuma system), but tolerances > > are probably best controlled in the CAM system. > > > > But, wouldn't it be more interesting if a pocketing canned cycled > > could take the target Material Removal Rate, Surface Finish, and > > Spindle Speed ( aka SFM ), and juggle the radial and axial depths of > > cut, engagement angle, and spindle speed? Maybe one where you call out > > a max spindle speed and max feed rate based on tool life and tool > > deflection, and let the software do the rest. More of a concept than a > > fully fleshed out idea. > > > > Jared > > I am for it, but that would take more hardware monitoring than we are > currently building into or onto our machines. Since different parts of > our machines resonate at different frequencies, how many vibration > pickups would we need to get adequate coverage? I don't know without > actually building such a beast. But in terms of overall production > speed, I can well imagine that if properly applied it would pay for > itself by picking up the production time involved in putting that part > in the ship it bin faster. > > > On Thu, Apr 16, 2020 at 2:35 AM Gene Heskett <[email protected]> > wrote: > > > On Wednesday 15 April 2020 23:51:38 Jared McLaughlin wrote: > > > > To be fair, I haven't touched a Heidenhain controller in at least > > > > ten years. And I wasn't that great with them, anyway. > > > > > > > > I just feel like, outside of drilling routines, most canned cycles > > > > don't have very robust algorithms - it's not like they are > > > > optimizing for anything but ease. The seem mostly to have a > > > > constant step over that doesn't account for things like angle of > > > > engagement in corners, etc. Obviously, lathe turning cycles are a > > > > different beast. > > > > > > > > There was a point in time, when I wrote all my gcode by hand, that > > > > I wrote out some cycles for myself in sub programs - but even > > > > those weren't that great compared to what I know now. Given the > > > > right sort of feedback, it'd be super interesting to see canned > > > > cycles for linuxcnc that take in to account feedback and adapt on > > > > the fly - but I'm pretty sure that's out of scope for this > > > > discussion. > > > > > > Why? That is how we spread ideas is it not? With a full command of > > > the language, there very very little we cannot do. > > > > > > We may have to bounce data back and forth between gcode and hal, But > > > we can do it. You see me asking questions about the latter because > > > even after 15+ years, I do not consider that I have a full command > > > of the language. ATM I have a rigid tapping routine that if it knows > > > how deep the blind hole is, will not ever hit the bottom of the hole > > > and break the tap even if the overshoot at the bottom as it turns > > > around is 3 or more turns which it could easily be if the 40 lb 8" 4 > > > jaw chuck is mounted. To be really complete it will have to measure > > > the holes depth first, but that requires I modify a bar holder to > > > both measure the blind holes depth, then hold a tap hat to thread > > > the hole with, but my problem with the GO794 has distracted me. > > > That's now fixed, and the need to purchase a rider because my place > > > us turning into a jungle, plus the ever increasing needs of my bride > > > of 30+ years as she is close to dying from COPD. That and medical > > > emergencies since I am diabetic and now halfway around my 86th trip > > > around this aging yellow star we call the Sun. My goal is code that > > > will peck-tap without breaking a tap on any machine from TLM, thru > > > the G0704 mill to the 11x54 Sheldon. That will take hal mods on TLM > > > in addition to making tap hat holding tool holders. The overshoot > > > measuring stuff hasn't been hacked into its config yet. > > > > > > Stay healthy all. > > > > > > > > > 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-developers mailing list > > > [email protected] > > > https://lists.sourceforge.net/lists/listinfo/emc-developers > > > > _______________________________________________ > > Emc-developers mailing list > > [email protected] > > https://lists.sourceforge.net/lists/listinfo/emc-developers > > > 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-developers mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/emc-developers > _______________________________________________ Emc-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-developers
