Re: [CAD_CAM_EDM_DRO] DRO stuff

"Arne Chr. Jorgensen" wrote:

> From: "Arne Chr. Jorgensen" <instel@...>
>
> Hi,
>
> First, - thanks Jon for the link to the datasheets. I have not looked
> at it
> yet, and it may take some time.
>
> { the ISP problem - after a short time - I am down to 6 bps, or the
> line
> comes to a full halt. (They are working on it) }
>
> Why do you insist on an optical system ? Do you now something that I
> don't ?

The advantage is that optical measuring systems can be made with fairly
high
resolution with no need for interpolation. Gratings with over 1000
lines/inch
(roughly 40 lines/mm) are no big problem for the industries that make
them.
Not so easy for the home shop machinist, though.

The inductosyn needs interpolation, which is a substantial complication.

There are chips to handle the entire interpolation job, but they are
expensive.
Making an inductosyn by printed circuit board techniques might work
well.
Cutting grooves in some insulating material and laying wires into the
grooves
sounds like it would be hard to get sufficient accuracy to beat even a
low
resolution optical scale, and it also sounds like a LOT of work.

> Don't missunderstand me - but I would like to dwell a little more on
> these
> other systems.
>
> Sure, the price on these chips may be high, - but you could make a
> cheaper
> circuit, - I think. I fully agree about the PCB layout, - I just
> mentioned it
> could be done in some other way.

Oh, don't be so sure, until you've read the data sheets. they have all
SORTS of exotic
sophistication in the AD 2S80 series to knock out sources of error.
Also, if you want
to make a CNC system, not a DRO, you need a position update rate of at
least 1 KHz,
maybe more, to keep a system like EMC happy.

> Jan wrote: ( He is to blame, if the Mitutoyo stocks went up :)
> -------
> > My <$100 Mitutoyo is "perfect" and just keeps on working and working
> and
> > working... And, it is SUPER-accurate! I can't believe it
>
> Well, this is the stuff that make me want to investigate into other
> than just
> optical systems.

Yes, but a caliper gets maybe 5 updates a second, and only the
super-expensive
Mitutoyo calipers do better than .0005" resolution. And, the accuracy
must be
less than the resolution.

> Jon wrote:
> -------
> >The Chinese scale has little squares about 2 mm square of copper
> >on a PC board material, with the squares ocurring about every 2 cm.
> The
> >head has a bunch of thin lines all connected to the control chip.
> <snip>
> > .. strips are only about 3 or 4 per mm, at the most.
>
> How large is the area that these thin lines enclose ?
> How many of them ?

I thought there were about 25 strips, and they must cover a length of
about 2 cm, to bridge
the space between squares on the spar. I think the strips are about 1
cm long, long enough
to cover the squares.

Jon