开云体育

Bill,

I'm not even using my dro any more. It did for a year or so and then the
kids wanted the extra computer for a "gaming" computer. I haven't used
my dro since. When I was using it...it worked most all of the time but I
feel it
was a little mickymouse with the wire and wheel. Yes...I had to be very
carful not to get oil on the wire. There must be a better way.
If US Digital has encoders to read from a linear strip this
would cure the problem of slippage.

Steve

On Sat, 8 May 1999 20:10:15 -0500, "Steve Lindsay"
<slindsay@...> wrote:

If US Digital has encoders to read from a linear strip this
would cure the problem of slippage.

Well, they DO, but doggonit they (or you should really read "HP") don't
have a linear resolution that in quadrature comes out to some nice
English (or even Metric) quantum.

I even had a chat with one HP 'marketing droid' type about this, and
couldn't seem to get across to him the value of a 250 (X 4 = 1000 in
quadrature) cpi sensor. Since they make a 200 and a 360 cpi model
presently, there's no reason they couldn't build a 250! And NO,
unhappily, their engineering guy told me (hope springs eternal) you
couldn't USE a 250 cpi optical strip with a 200 cpi sensor to get 0.001"
native resolution. The sensor arrays are specific to almost exactly the
optical resolution, so they need to MAKE the bloody thing!! 8(

Their rotaries, however, have lovely and suitable resolutions, so I
guess if you're using zero-backlash ball-screws, and just wanna count
rotational steps to determine position, it's a simple slam-dunk
done-deal already.

I trust most everyone realizes that pretty much all US Digital does is
make the mylar discs/strips that are compatible with the HP sensors (as
well as a line of 'encoder' interfaces for these sensors, to be fair
about it), and then resell the HP sensors? HP has a VERY nice
book/catalog on these "HEDS" sensors of theirs, and I find it much
better a reference than the US Digital "re-makes" of HP's data sheets.
HP's book also shows that building the 'encoder' interfaces is merely a
matter of collecting a few off-the-shelf parts, just incase you're
interested. B)

[not to diminish USD's contribution with the optics ; just that them
'shielding' their clients from HP leaves me a tad cold]

Gar

Gar.... 360 will work! In the setup screen in the dro program
you can enter a number to count by.... and it supports a number
up to 14 places right of the decimal. The 360 cpi would
quaduture to 1440 and 1440 divided by 1" is 0.00069444444444
Enter this number into the dro axis setup. To get to the setup
screen after the program starts....hit R "Read Encoders" and then
X or Y or Z to get to that axis setup. In this setup you can also
tell it how many numbers to display right of the decimal on the readout
screen so all 14 places don't clutter your screen.

Back when I was writing the program the most HP had for linear encoders
was 200 and then HP told me they don't make the strips and I would have
to find another company that made them. I do have a HP big encoder book
and called lots of those numbers in the back finding a company
that would sell what encoders I was after (since HP won't sell direct).
Later
I found US digital. They are just another reseller of HP stuff but I
thought
their web pages layout describing what HP had was helpful... At
least compared to the luck I was having calling all the suppliers/resellers
in the back of HP encoder book. Also the way I understood it they were
making the linear mylar strips. I haven't looked at US Digitals web pages
for a long time but if they are coming across like they are big stuff
without
HP I would find it cold too.

Steve
Digital Read Out 4.0 (encoder reader)

On Sun, 9 May 1999 10:00:50 -0500, "Steve Lindsay"
<slindsay@...> wrote:

Gar.... 360 will work!

Gotcha; yeah, I know, anything really will work, as long as you have
enough precision in the numbers so you don't bump a count due to
roundoff, but it's sure messy to be working with (and displaying the
results of) native resolutions of 0.00069444444444", doncha tink? B)
Specially when, if HP would just listen to us home machinists (hee hee),
they could just MAKE one perfectly suited for us! I just don't
unnerstand why they don't consult with us hobbyists when doing their
marketing surveys. B)

Back when I was writing the program the most HP had for linear encoders
was 200 and then HP told me they don't make the strips and I would have
to find another company that made them.
...
Also the way I understood it they were
making the linear mylar strips. I haven't looked at US Digitals web pages
for a long time but if they are coming across like they are big stuff
without HP I would find it cold too.

Yeah, agreed; like I said, I really DO appreciate their being both a
supplier and specially a reasonably priced supplier of the optic media,
and on their web pages they're not REALLY tryna "hide" the fact that
HP's sensors are key to their products, it's just that they coulda put
links there to HP's website where the HEDS sensors and app notes for
them are available, so as to help out their clients. But I also
understand the business concept of presenting yourself as a sole
supplier or rather "one-stop shopping", so it's not really any serious
complaint I was alludin to, ifyagetmedrift. No biggie.

BTW, the HP link to their motion stuff is:


Punch in the part number "heds-9200 opt 360" to go directly to:
"HEDS-9000/9100/9200 Extended Resolution Series"

I noticed when checking this link to make sure it was still alive, that
them bozos have ADDED a 300 cpi sensor to their repertoire, but NO 250
cpi one!! Da noive a dem guys. Sheesh. Time for a write-in campaign, eh?
:)

Gar

Gar Willis wrote:

I even had a chat with one HP 'marketing droid' type about this, and
couldn't seem to get across to him the value of a 250 (X 4 = 1000 in
quadrature) cpi sensor. Since they make a 200 and a 360 cpi model
presently, there's no reason they couldn't build a 250! And NO,
unhappily, their engineering guy told me (hope springs eternal) you
couldn't USE a 250 cpi optical strip with a 200 cpi sensor to get 0.001"
native resolution. The sensor arrays are specific to almost exactly the
optical resolution, so they need to MAKE the bloody thing!! 8(

actually, you can make the whole works yourself, if you can get the
grating. What you do is use another piece of the grating as an analyzer.
To obtain the 90 degree quadrature shift, you TILT the analyzer grating
such that the spots you pick for your photocells are skewed 1/4 of a
grating pitch with respect to itself. The only limitation on this is that
the grating needs to be wide enough so the 2 photocells can be placed
a short distance apart (1/4 to 1/2 inch) and when the analyzer grating is
tilted, both photocells can see through both the analyzer and the
measuring grating. A 1/4" wide grating would be about the minimum
width. If this isn't the case with whatever strip you plan to use, then
you need to make 2 analyzer pieces, and give one of them a position
adjustment, so you can set up the quadrature phasing. It really isn't
that hard. I was working with a .000005" version of a dial test
indicator that used the tilted analyzer grating trick. Even setting
that one up wasn't very hard, and that was a VERY precise set
of gratings.

Jon

On Sun, 09 May 1999 23:56:22 -0500, Jon Elson <jmelson@...>
wrote:

actually, you can make the whole works yourself, if you can get the
grating.

Yeah, but honestly, I don't have time nor patience to be making sensors
and adapting optics. Fine if it's my avocation, but it's not. I DO
appreciate the rundown on the basic simplicity of 'quad' via optics, but
I wanna make parts, not develop sensors and so-forth. It's just sad that
HP doesn't see the value of a 250 X 4 sensor array. If they did, then US
Digital would make the optics post-haste, and it'd be a done deal.

Me, I got other fish to fry. Sad (maybe) but true, life is short. B)

Gar

Gar Willis wrote:

From: garfield@... (Gar Willis)

On Sun, 09 May 1999 23:56:22 -0500, Jon Elson <jmelson@...>
wrote:

actually, you can make the whole works yourself, if you can get the
grating.

Yeah, but honestly, I don't have time nor patience to be making sensors
and adapting optics. Fine if it's my avocation, but it's not. I DO
appreciate the rundown on the basic simplicity of 'quad' via optics, but
I wanna make parts, not develop sensors and so-forth. It's just sad that
HP doesn't see the value of a 250 X 4 sensor array. If they did, then US
Digital would make the optics post-haste, and it'd be a done deal.

Well, I really don't think it would be that big of a project. But, knowing that,
I went with precision ballscrews and relatively cheap shaft encoders. You
can get a much greater resolution that way. A leadscrew with .200" lead
(5 TPI) and a 1000 line encoder (counting all transitions for 4000 counts/rev)
gives resolution of .00005" per count. If you can get a .100" lead screw
(I have one for the Z (quill) axis) you get .000025" per count - that's 25
micro inches. Great for EDM feed.

Jon

On Mon, 10 May 1999 13:55:16 -0500, Jon Elson <jmelson@...>
wrote:

Well, I really don't think it would be that big of a project. But, knowing that,
I went with precision ballscrews and relatively cheap shaft encoders. You
can get a much greater resolution that way. A leadscrew with .200" lead
(5 TPI) and a 1000 line encoder (counting all transitions for 4000 counts/rev)
gives resolution of .00005" per count. If you can get a .100" lead screw
(I have one for the Z (quill) axis) you get .000025" per count - that's 25
micro inches. Great for EDM feed.

Good point. Lack of lash in the screws sure DOES help out in other ways,
don't it! And if you used steppers on those screws, at those
resolutions, you would even get to watch the motors twitch! B)

Gar