On 12 Jun 2014, at 02:34, Todd Zuercher <zuerc...@embarqmail.com> wrote:
> The machine is one that has had a problematic (under engineered drive train) > from the get-go. It is an 8 spindle gang router for wood carving, with a > 12ft long gantry that has about 5ft of travel (Y-axis). Even though the X is > on a 12ft gantry it only has a little over 3ft of usable travel. X and Z > both currently use 0 backlash lead screws (these are showing some age) and I > want to replace them with ball screws, with a shorter pitch and instal better > fixing blocks (what is there now is pretty el cheapo and is adding a lot of > backlash). When reducing the pitch, you can have a higher thrust. Mainiainting the same speed means you’ll have higher RPM. When designing the fixing bearing construction there’s deciding which side of the spindle to fix, an which to slide. The spindle is not only about the ratio, the geometry (diameter) of the spindle and it’s lengths have a great influence I’d advice doing some calculations on buckling and vibrations for the spindles max RPM and thrust. most vendors have example calculations. > My hope is to reduce backlash, reduce friction, and increase mechanical > advantage, in the hope of improving acceleration and stability. The changes > on the Z will change the encoder resolution from about 8K counts/inch to > about 10K, X will go from about 4.5K to 10K. this all sounds relatively > simple so far. > > That leaves the problem child Y axis. It has a helical rack and pinion (~2 > inch diameter pinion) on each end of the gantry connected together by 2.5 > inch diameter torsion tube. The drive is on the outside and consists of a > large fine toothed timing belt pulley (about 8 inch diameter on the pinion > shaft) and a very very small one on the servo motor (less than .75"). Tooth > count on the big one is 212, and I can't see the little one to count it right > now it might be as small as 14 teeth. Regardless what the actual ratios are, > a 2000 count (8K quad) encoder on the motor gives about 7.5K count/inch. > Other than the step motor esque gearing, I think my biggest problem is belt > squirm on the ultra small motor pulley. > > I think a would be happy with +-0.001 inches accuracy and satisfied with > +-.002 The machine is all aluminum construction so is far from heavy duty, > but it isn't some light hobby job either. > > All the servos are 850oz/in. (I think I'll have to check the numbers to be > sure). > > I was considering just putting a ball screw on just one side of the gantry > and leaving the racks and pinions alone (still connected). I know it kind of > still leaves the far side to flop around, but it's doing that already > (torsional flex is a bitch ain't it). yes it is :) Especially because you want the nut of the spindle to be parallel with the axis. If you have forces other than pure couple, then the life of the ball nut is reduced. > > I had thought about using 2 90 deg gear boxes, but their expense and fear of > backlash has me shying away. > > I may be best just biting the bullet and going the 2 motor route. > > > > ----- Original Message ----- > From: "Bas de Bruijn" <bdebru...@luminize.nl> > To: "Enhanced Machine Controller (EMC)" <emc-users@lists.sourceforge.net> > Sent: Wednesday, June 11, 2014 1:52:55 PM > Subject: Re: [Emc-users] Ball Screw Drive? > > > On 11 Jun 2014, at 19:00, Todd Zuercher <zuerc...@embarqmail.com> wrote: > >> Any one have any ingenious ideas for a good way to drive two parallel ball >> screws that are 12 feet appart (gantry router), with only 1 motor (instead >> of 2) ? > > Hi Todd, > > could you elaborate a little bit on the actual setup? Do you want to rotate > the nut on the gantry, or do you want to rotate the spindles on the frame > beneath? > > Either way, if you want to use 1 motor, then position it in the middle so > elasticity (I roughly calculate 3 feet in 1 meter) over the 4 meters is equal > on both sides. Otherwise you will have one of the nuts/spindles lagging > w.r.t. the other one. > > What is the torque you expect? and the tolerance you expect to mill? > What kind of finish (part/component quality) are you after? > > - 2 90 degree gearboxes with a shrink collar could be used. with a gearbox in > the middle. > - Another way could be to put bevel gears on the end of the spindle and drive > that with 1 axis (again, motor in the middle). > - or you could use an eccentric plate construction. A motor with an eccenter > who drives 1 plate. That plate is held in place by 3 executers. Then you put > an executer on both spindels/nuts and when the eccentric plate starts to move > the spindle/nut will be rotated. You’ll have a bad time fighting vibrations > :) you’ll need to counterbalance it so I guess it gets a little bulky. To get > the idea: something like this, http://www.google.com/patents/US2342251 or > http://www.google.pl/patents/US4260301 > > >> >> -- >> >> ======================================== >> >> Todd Zuercher >> mailto:zuerc...@embarqmail.com >> >> ======================================== >> ------------------------------------------------------------------------------ >> HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions >> Find What Matters Most in Your Big Data with HPCC Systems >> Open Source. Fast. Scalable. Simple. Ideal for Dirty Data. >> Leverages Graph Analysis for Fast Processing & Easy Data Exploration >> http://p.sf.net/sfu/hpccsystems >> _______________________________________________ >> Emc-users mailing list >> Emc-users@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/emc-users > > > ------------------------------------------------------------------------------ > HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions > Find What Matters Most in Your Big Data with HPCC Systems > Open Source. Fast. Scalable. Simple. Ideal for Dirty Data. > Leverages Graph Analysis for Fast Processing & Easy Data Exploration > http://p.sf.net/sfu/hpccsystems > _______________________________________________ > Emc-users mailing list > Emc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-users > > ------------------------------------------------------------------------------ > HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions > Find What Matters Most in Your Big Data with HPCC Systems > Open Source. Fast. Scalable. Simple. Ideal for Dirty Data. > Leverages Graph Analysis for Fast Processing & Easy Data Exploration > http://p.sf.net/sfu/hpccsystems > _______________________________________________ > Emc-users mailing list > Emc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-users ------------------------------------------------------------------------------ HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions Find What Matters Most in Your Big Data with HPCC Systems Open Source. Fast. Scalable. Simple. Ideal for Dirty Data. Leverages Graph Analysis for Fast Processing & Easy Data Exploration http://p.sf.net/sfu/hpccsystems _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
