Re: Optical CNC

Hi Tim.

I hate posting these little nit-picks, but in Windows and other OSs
the cursor is programmed to move farther when you move the mouse
quickly. The usability engineers have figured that this is more
natural than making each mouse tick (called a mickey) worth a pixel.
If you slide it with a slowly increasing speed, you should be able to
see the pointer speed jump a bit at certain thresholds. Ever fire up
an old program have the mouse feel "slow"? That's the benefit.

But otherwise, you're right in that optical mice aren't really
calibrated for true distance. They can't be by their construction.

Just imagine floating in space with one eye covered, with points of
light floating by in the distance. Are the points close by and
moving slowly, or far off and moving fast? The only meaningful
observations are that they aren't moving apparently as fast as they
were a moment ago, which is all a mouse really needs to know in order
to work. But for CNC, we need to know how far.

Now for some actual content:

I've though about this a bit, and it would be *really* something
slick to program a small camera, like the sensing element in an
optical mouse, to read a variety of plain scales like the 6" steel
ones we all have.

I imagine that it wouldn't be hard to pick out a location accurate
to a thou when two or three of the 1/64 lines are the whole FOV.
Heck, some machinists can eyeball .005 with a scale on a good day.

Further, if the program can distinguish long lines from short ones
from shorter ones, and even recognize a "3", it wouldn't even need to
be told where it is either.

But the real advantage would be that the DRO strips would then be,
er, ubiquitous and non-proprietary.

Anybody up for something like that?

Dave Kowalczyk
Everett WA

--- In CAD_CAM_EDM_DRO@y..., "Tim Goldstein" <timg@k...> wrote:
> Jim,
>
> I haven't noticed a reply to your mouse question so here goes.
>
> Using a mouse as the encoder for position feedback has a problem in
that
> the update rate is too slow. The fps figure you give for an optical
> mouse is misleading as it is not the rate at which the encoding
chip can
> provide position updates out the serial line. To prove that it loses
> track of the encoder counts all you have to do is position your
cursor
> to one side of the screen and note the position of the physical
mouse.
> Not rapidly move it to the other side of the screen (not so far your
> cursor goes off screen) and then slowly move it back to put your
cursor
> where you began. You will most likely not end up at the starting
> position as you will have lost counts on the fast move. So you can
see
> that using the mechanism as is will not work.
>
> Now you can use the encoder out of a mechanical mouse as it is just
a
> low res quadrature encoder. There has been discussion on using the
> optical sensor on the newer mice, but the problem that came up is
the
> driver circuits drop into a low res mode when they sit still and
you are
> likely to drop counts when it first moves as it jacks up the power
to
> the led and the resolution.
>
> Tim
> [Denver, CO]
>
>
> > -----Original Message-----
> > Has
> > anyone gave some thought to using an optical sensor from a mouse
as
> > guidance for the router? I see some mice (mouses) have up to 6000
> > frames per second and even wireless. Most computers already use
the
> > mouse as an input, but could it be used with a router?
> >
> > Standing by,
> >
> > Jim