Re: [CAD_CAM_EDM_DRO] Re: Synchronizing Two Servo Motors

As far as I know, you can get belts almost as long as you like. The belts
you are referring to are "endless belts", and are cut off a tube. These are
available in limited lengths

What you need to ask for (Fenner, Trumotion etc) is an "open ended belt".
Bascially this is a reel of timing belt. You will affix the two ends at the
carrige.

Look here;

https://sdp-si.com/eStore/Direct.asp?GroupID=343

Also select a rounded, quieter profile, like this;
http://news.thomasnet.com/fullstory/540514


And some good general info here;
http://motion-controls.globalspec.com/Specifications/Motion_Controls/Power_Transmission_Mechanical/Timing_Belts_Metric_Pitch


Regards
Roland



2009/3/11 Yahoo <yahoo@...>

> Jon,
> Thanks, to you and a couple of other guys, for the ideas on big machines,
> but this is a small one, only 24" travel X & Y. I've thought of many
> different mechanical solutions, but every one ends up being too
> complicated.
> The only practical one would be to slave the two screws by a chain or belt,
> but it seems like this could cause the motors to fight, leading to
> potential
> oscillation problems or excess current when idle. Setting the controller to
> stop everything when any channel faults, and slaving two channels for the
> dual lead probably would work, except in case one of the motor couplers
> failed. This failure also would not be covered by any electrical solution
> that relies on the motor encoder signals, as they are mounted on the motor,
> not the lead screw. The only really fail safe solution would be a
> differential switch controlled by a timing belt at the far end of the
> screws. If only I could find a belt long enough. It would need to be
> roughly
> 6' and anything too thick would eat up a lot of torque just in friction.
> Longest I've seen in the 3/8" width range is about 4'.
> --
> Phil Mattison
>
>
> ----- Original Message -----
> From: Jon Elson <elson@... <elson%40pico-systems.com>>
> To: <CAD_CAM_EDM_DRO@yahoogroups.com <CAD_CAM_EDM_DRO%40yahoogroups.com>>
> Sent: Tuesday, March 10, 2009 7:50 PM
> Subject: Re: [CAD_CAM_EDM_DRO] Re: Synchronizing Two Servo Motors
>
> > Yahoo wrote:
> > > I thought about sharing a belt but haven't been able to find a belt
> long
> > > enough. Seems like it would be sort of a kludge anyway.
> > >
> > > I guess if this has been discussed ad naseum and nobody has a ready
> > > solution, then it probably has never gotten beyond the talking phase.
> > > Apparently I'll have to build something from scratch (sigh).
> > >
> > No, I think we DO have a solution. There have been several schemes,
> > depending on whether the gantry is rigid or flexible.
> > In big commercial machines, the beam across the gantry sides has a pivot
> > on one side and a slotted bearing on the other.
> > This allows the gantry sides to get a certain amount out of sync without
> > any binding. They probably put an angle limit switch on the pivot so if
> > the angle exceeds a certain amount, it E-stops the machine. Then, they
> > have a scheme for homing the two sides of the gantry simultaneously.
> > The two sides run at the same (servo controlled) rate. Each one hits
> > its own home switch and completes the precision homing sequence by
> > itself. After that, the tandem axis stays in sync as well as the servo
> > can, which will usually be quite close. So, that is how the "pros" do
> > it on $100K plus machines like gantry bed mills and vertical boring
> mills.
> >
> > I think you could do the same on a smaller machine. It would also allow
> > you to adjust the orthogonality of the machine by just offsetting the
> > home position in software until a square came out square. But, this
> > would require buidling in the joints in the gantry.
> >
> > Another scheme I've heard, but not sure it has been used, is to home one
> > side while allowing the other side to just coast with the servo disabled.
> > When the "master" side is homed, then the slave side servo is enabled.
> >
> > I know EMC could easily be set up to independently home the two sides of
> > a tandem axis, as described in my first paragraph. It would take very
> > little special HAL coding to do this.
> >
> > Jon
> >
> >
>
>
>

[Non-text portions of this message have been removed]