Re: rotary hand encoder wheels with detent

Hi All,

I've recently been working on several projects which have the need
for a tactile feedback control knob so I've been evaluating options.
Since one of the projects I'm working on is for commercial release,
I've been looking into available commercial solutions so that I can
reliably get parts. Therefore, what follows may or may not be
relevant the discussion.

First, I'm confused by the need for a 1:1 correspondance between knob
detents and the typical 100thou per rotation. While certainly
convenient, it seems part of the utility of a computer controlled
system is the ability to scale controls on the fly.

For the system I'm working on, I'm using two controls. One needs high
precision, and for this I'm using an expensive optical rotary
encoder. The second requires known steps, so for this, I'm using an
inexpensive (~$5) mechanical encoder, available through the standard
electronic suppliers. The particular device has 36 positions per
rotation. This will be running a stepper based drive system .
Assuming a 1:1 coupling between the stepper and the drive axis, with
a 100thou advance/rotation, as you note, you get 1/2 thou per full
step.
I will be mapping individual detent positions to:
a) two microsteps = 1/200/4 = 1/800 motor rotation = 1/8 thou
b) one full step = 1/200 rotation = 1/2 thou
c) several full steps. i.e. 10/200 rotation = 5 thou
d) many full steps. i.e. 100/200 rotation = 50 thou

In this way, one can easily dial in any desired position. Granted,
with a 36 detent encoder, one doesn't get a nice correspondance
between knob position and feed screw rotation, but in practice it
would seem that it's more important to have the detents correspond to
a nice neat drive amount.

I'll soon find out anyway, as I'll be doing exactly this after I
finish these other projects. I've got a neat little DRO and fly-by-
wire control panel project in the design phase, and once that's done,
I'll put all the info out there in the public domain for others to
play with.

Have fun,
Ted

--- In CAD_CAM_EDM_DRO@y..., "Ian W. Wright" <Ian@i...> wrote:
> Hi Alan,
>
> I did make the suggestion of using a mouse partly tongue in cheek
as there
> have been so many discussions of their use as encoders. However, I
have
> taken this a little further and done a dental check up on a couple
of mice
> which has revealed that, as with anything in nature, everything has
> variations and, of three mice I inspected, two had 36 teeth, one 24
teeth
> and one 30 teeth so it may be that somewhere out there, there is a
mouse
> with a useful number of teeth!
> Having said that, why do we need 100 clicks per rev of the
handwheel? I
> thought the whole idea was that, rather than working to any precise
portion
> of a revolution, you actually counted the clicks by touch. This
being the
> case, it wouldn't really matter how many 'teeth' are on the wheel
as long as
> you could depend on one click moving the machine table one
increment - so,
> give the handwheel 72 detent positions and you have a nice ratio
between the
> wheel and the mouse encoder disk. (and, of course, 72 would be a
nice
> relevant number for a rotary table..)
> In actual practice I suspect that the number of teeth on the
encoder disk
> will not matter a lot and that all you need is a precise ratio
drive from
> the handwheel to the encoder disk which can be achieved by
carefully turned
> shafts and decent rubber band drives (like in good tape recorders),
then, so
> long as the encoder spins so as to produce a number of pulses for
each
> detent click, you can divide these down in electronics to give one
pulse per
> position and even out any errors in the encoder teeth (not that
this is
> particularly relevant in this application either!)
> The point really is that this would give a more positive pulse than
a
> stepper motor, would have a more positive 'click' (spring working
on the
> teeth of an old clock wheel or other gear fixed to the handwheel)
and the
> hardware would be cheap and easily fabricated by anyone with even
> rudimentary equipment. The electronics to interface it with EMC may
be more
> problematic but I'm sure there are people on this list who would
scoff at
> the challenge//;O)
>
> Best wishes,
> Ian
> --
>
> Ian W. Wright
> Sheffield, UK
> ----- Original Message -----
> From: <flyspeck1@a...>
> To: <CAD_CAM_EDM_DRO@y...>
> Sent: Friday, February 15, 2002 12:07 AM
> Subject: [CAD_CAM_EDM_DRO] Re: rotary hand encoder wheels with
detent
>
>
> >
> > It doesn't
> >
> > I am not too certain that using a mouse would work very
> > well. I just counted the number of serrations that you
> > get on a mouse and it is 36; based on previous
> > conversations on this group, that cannot be increased.
> > What we really want, however, is a something that
> > divides 100 cleanly..... If we were to use a 1.8 degree
> >
> > Or.... I am missing something! Sure, I guess you could
> > multiply it up pretty far, but it still seems like a
> > less than ideal setup. Too bad... I liked the idea too.
> >
> > --Alan
> >
> > <<<<Original Post>>>>
> > Message: 11
> > Date: Thu, 14 Feb 2002 19:28:02 -0000
> > From: "Ian W. Wright" <Ian@i...>
> > Subject: Re: Re: rotary hand encoder wheels with detent
> >
> > Perhaps, at last, this is a use for the encoder wheel
> > out of a mouse!!! ;o)
> > Hand wheel with clicks, and a rubber band drive from the
> > hand wheel shaft to
> > the encoder spindle to gear it up in a suitable
> > ratio.....
> >
> > Best wishes,
> > Ian
> > --
> >
> > Addresses:
> > FAQ: http://www.ktmarketing.com/faq.html
> > FILES: http://groups.yahoo.com/group/CAD_CAM_EDM_DRO/files/
> >
> > Post messages: CAD_CAM_EDM_DRO@y...
> > Subscribe: CAD_CAM_EDM_DRO-subscribe@y...
> > Unsubscribe: CAD_CAM_EDM_DRO-unsubscribe@y...
> > List owner: CAD_CAM_EDM_DRO-owner@y..., wanliker@a...
> > Moderator: jmelson@a... timg@k... [Moderator]
> > URL to this page: http://groups.yahoo.com/group/CAD_CAM_EDM_DRO
> > bill,
> > List Manager
> >
> >
> >
> > Your use of Yahoo! Groups is subject to
http://docs.yahoo.com/info/terms/
> >
> >
> >