Re: Re: Two motors on one axis EMC

Hello Lars and Thomas (Comments mixed in)
<s>

> If you're doing servos, a Servo-to-Go card (which is EMC-compatible)
> can drive up to 8 servos. I do not think, however, that it would
> easily allow you to master-slave a pair of motors. Since I do not use
> servos, I cannot perfectly advise on this.

Syncing servo's will require a complete loop for each motor and a higher
level task that coordinates them. At the conceptual level, it would be
possible to ask the interpreter to do that gearing. Something like the
following code if x and y referred to a pair of motors driving the same
axis of a gantry. [g0 x21.5 y21.5] would do it for one move but each pair
would need to exactly match throughout the code.

I'm guessing here but I'd think that the task could be pushed off onto the
kinematics. Yes it's a trivial solution -- trivkins with a twist -- but
I think that by defining an X as an imaginary line half way between two
parallel actuators, it would produce the correct solutions without any
further complications. You'd need to post the question to the emc users
list and get help from the likes of Till, Sagar, Fred, and Will.

It is also possible now to push the task of electronic gearing deeper into
the code than the two examples above. A student at NIST recently
demonstrated this by grabbing the shaft on one motor and giving it a
twist -- the second motor exactly followed the motion. We would have to
help you find your way to that code and enable it for your machine. I've
asked for a copy of the student's paper and should have that available
soon in the integrators handbook.

> I cringe when I think about trying to sync two servomotors, because
> each motor had it's own feedback loop. For myself, I would use a
> long-distance mechanical coupling between two leadscrew, such as a
> precision timing belt, and have the motor turn both using the belt.

You're right Thomas, this does seem like the easier route but it is also
subject to mechanical degredation over time. Belts and mechanical
devices wear and this will cause great stress on a large system.

This is a philosophical aside for which I'll owe another burger at NIST,
but in my younger days I constructed the most perfectly designed,
alligned, supported and braced field wire fence for a small pasture. It
was a marvel to behold. In a matter of a few days, that %$##@* pony,
given to my kids, had that "marvel" looking just like every other fence
in the county. I should not need to extend the aside into the realm of
large machines in order for you to see the parallel.

> There are other issues, such as alignment/racking of the gantry when
> driven by two motors, synchonization of the motors and leadscrews, and
> the cost of each possible solution.

It may well turn out to be less costly to build electronic gearing code,
and multiple drives, than to maintain the mechanical system.

As ever -- hope this helps.

Ray