On Thursday 18 August 2016 23:57:23 Todd Zuercher wrote: > But this is a BLDC servo and free wheels quite easily when powered > down. I put the .2" lead screws back on today and re-tuned with a > couple of new tricks and got much better results than I've had in the > past. > > The drives use torque commands, and had been difficult for me to tune > well and get stiff and quiet. They need a lot of I, but using a lot > of I caused them to be unstable. The new trick I've tried was to use > the pid.N.maxerrorI. I turned up the I-term till it started to get > unstable (large overshoots not oscillating), then I set an error limit > on I, and made it smaller and smaller until the I induced overshooting > was minimized. > pid.N.MaxerrorI? I don't believe linuxcnc's pid's have such a knob. We've a maxerror, but applying that has so far, not had a desirable result, making the instability worse in my explorations. And it has not been mentioned in any of the reading I have done on P.I.D. controls.
Interesting, link? > ----- Original Message ----- > From: "Gene Heskett" <[email protected]> > To: [email protected] > Sent: Thursday, August 18, 2016 5:04:09 PM > Subject: Re: [Emc-users] Leadscrew Efficiency? > > On Thursday 18 August 2016 11:16:02 Todd Zuercher wrote: > > They are both almost identical screws and nuts except for their > > leads. They are both Kirk Motion VHD series LEAD screws. > > http://www.haydonkerk.com/LinearActuatorProducts/LeadScrewsAndNuts/N > >ut s/VHDSeriesNuts/tabid/148/Default.aspx > > > > So you are saying that all things being equal the turning force > > should be about the same for the 2 screws. Which must leave me with > > the conclusion that the older lower lead screw I pulled off the > > shelf is probably mostly worn out. I'll go disassemble it and > > inspect the threads in the nut. > > > > I'm considering swapping to ball screws, but I'm a little concerned > > about it back-driving when off, and I don't think I can find a 0.1" > > lead ball screw (screws are currently direct coupled to the motor). > > So far, I have not found that to be a problem as the steppers I have > used have a sufficient cogging action when powered off that the only > time I have had one unwind and drop the head was when I took the > 1600oz/in off the G0704 to replace it with the 940oz/in. > > It unwound and stopped on some relatively 'cushionable' wood stuffs > laying on the mills table, no damage to anything other than my > surprise when I realized and understood why it was happening. I won't > say its not happened, like the knee on a bigger bridgeport, which > weighs about 15x what the head on a G0704 weighs, but I've not seen it > happen, just read about it. So I'd go a decent way out on a persimmon > limb and say thats a servo problem 99% of the time. If anyone here > has had a problem like that while using steppers, please speak up, we > would like to hear about it. > > > ----- Original Message ----- > > From: "John Kasunich" <[email protected]> > > To: [email protected] > > Sent: Thursday, August 18, 2016 10:50:59 AM > > Subject: Re: [Emc-users] Leadscrew Efficiency? > > > > Are they LEADscrews or BALLscrews? > > > > With ballscrews, the finer pitch screw should indeed provide a much > > better mechanical advantage. > > > > With leadscrews, the mechanical advantage almost doesn't matter, > > because friction is by far the dominant force. Somewhere between > > 60% and 95% of the torque required to turn a leadscrew under load is > > due to the friction between the screw and nut rather than the force > > actually needed to raise the load. > > > > I'm not sure why the 10tpi screw would actually be worse - I would > > expect it to be roughly the same. However, many factors come into > > play: > > > > 1) materials: are both screws steel? both nuts bronze? (pr > > plastic, or whatever) The materials and lubrication can make a huge > > difference in the amount of friction. > > > > 2) surface quality: a rough screw or nut will have more friction > > than a polished one > > > > 3) thread form: 60 degree threads have a lot more friction than > > Acme due to the wedging action of the 60 degree flanks. Square > > threads are best because the flanks are perpendicular to the load, > > but Acme is almost as good. > > > > 4) diameter: if all of the above are equal, the friction FORCE will > > be the same. But the TORQUE required depends on the radius of the > > screw, so a larger diameter screw will require more torque. > > > > On Thu, Aug 18, 2016, at 10:14 AM, Todd > > Zuercher > > wrote: > > > I have a machine that I converted from step-motors to servos, and > > > I'm having a little trouble with the Z axis. It has an anti > > > backlash lead screw with a 5tpi screw. This sort of worked, but > > > the servo was working hard to move the head, and I wanted a little > > > higher encoder resolution for better tuning. So I swapped in a > > > nearly identical 10tpi lead screw set I happened to have on hand > > > (removed from another stepper machine to get better speed and > > > perfomance). I thought that the 10tpi screw should be easier for > > > the servo to turn, but I'm finding that the opposite is true and > > > the servo can't raise the motor without counter balance assistance > > > when it could with the 5tpi. Does that make sense, or is the > > > problem more likely that the 10tpi screw and nut are worn out and > > > binding? > > > > > > -- > > > > > > ======================================== > > > > > > Todd Zuercher > > > mailto:[email protected] > > > > > > ======================================== > > > ------------------------------------------------------------------ > > >-- ---------- _______________________________________________ > > > Emc-users mailing list > > > [email protected] > > > https://lists.sourceforge.net/lists/listinfo/emc-users > > Cheers, Gene Heskett 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) Genes Web page <http://geneslinuxbox.net:6309/gene> ------------------------------------------------------------------------------ _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
