Re: [CAD_CAM_EDM_DRO] Re: Rotary to linear translation - an interesting idea.

>
>
>
> "D.F.S." wrote:
>
> > > the tape must travel around
> > > the radius faster in order to maintain it's length since the tape is
> >
> > > not stretchy (to that degree anyway).
> > Nope. Not True.
> >
> > The length of the Tape/Wire/Cord/Shim stock is already of a length
> > to allow for the extra bends it has in it.
> >
> > Please Excuse the ugly ascii art.
> >
> > C
> > /0\
> > 0/ \0
> > J ------------------ --------------------------- K
> > A B
> >
> > The presumption above is:
> > The "Wire" must move from A to C to B in the same 'time' it would
> > have taken it at a constant speed to travel from A to B.
> >
> > This in simply not true.
>
> Yes, you are absolutely right. No way the tape moves from a to b in the
>
> same time as the linear move. BUT, there is a slight change in the
> effective
> diameter of the ribbon passing around the wheel. This is just like the
> pitch diameter of a belt, due to the inner surface of the ribbon being
> in
> compression. With a very thin ribbon, the effect would likely be
> smaller than
> good measuring instruments could measure

Snip...

This is absolutely correct as well.
I agree so wholeheartedly because I was the first one to bring it up,
and couldn't agree more ;-).

There are several issues being discussed here.
The reply above was relative to the incorrect speculation that the
counter would run "Fast" or "Long" depending on your criteria,
due to the fact it was longer because of going around the pulleys.

That was why I excluded it as a whole, I forget if I said assume a
zero thickness of the ribbon or the length at the neutral axis.


In any event, there is more to it than the flexing of the ribbon or
cable showing a larger effective diam.
That flexing will be taking place on the surface at the points of
initial contach with the ribbon.
Presuming you have perfectly equal reactions they will counteract
and there will be no creep relative to the the surface of the ribbon
and the wheel. We all know the act of cutting a pulley will not leave
a perfectly smooth or symetrical surface.

That surface will teeth, and they will point in the direction it was
cut. It may take a microscope to see them, but they ARE there.

Example.
Lay a file on a surface.
Without holding it down slide a Block back and forth along the surface.
That file will creep until it moves out from under the block even with
tiny side to side movement.

Call the block the ribbon/cable and the file the encoder shaft.

The effective diameter change will be linear presuming a constant tension.

This is not an issue, unless someone is dead set on making an encoder
where each count on the wheel is EXACTLY .0001000000".

That would take a lot of trial and error.

In reality if one count is .0001234567, and we figured it would be
.0001234424, so what?
It's just a number in a floating point constant in the software anyway.

like I said before, Either analyze it to death with a dozen unknowns,
or simply know at the outset the distance of one count WON'T be
Diam x Pi/Counts per Rev. of the encoder.

It'll need to be calibrated after assembly and the number set for
the actual values. End of problem.

The creep issue mentioned above is another issue.
IT may or may not be important, the answer is the same
as before though, build it and test it.

Maybe the answer it a re-zero counter that will prompt the
user to home the system every thousand inches of movement or
whatever to keep the errors to an acceptable level.


Marc