Re: [CAD_CAM_EDM_DRO] encoders and table position

Weyland wrote:

>On Friday 06 June 2003 01:06 am, Jon Elson spake:
>
>
>
>>Weyland wrote:
>>
>>
>>>Okay, having looked into it a bit more, it doesn't look like the linear
>>>scales I have are going to work with the Kaluga/Mauch DRO board after all.
>>>
>>>
>>Why won't they work? Are they analog scales? You can get the basic
>>resolution of the scales pretty easily, but that might be too coarse.
>>As long as they provide some sort of quadrature signal, it should be
>>relatively simple to interface them.
>>
>>
>
>Well, for a few reasons as I understand (or not).
>First is that no one seems to be able to tell me with certainty what the wires
>do, and as far as I've found in my searches, there isn't any standard to the
>color coding of the wires.
>

You got that right, as I found out a while ago.

>Second, is that it seems I have a few different
>configurations, two are six pin connector, and one is nine pin.
>
>Now, in taking the six pin apart, I did find that they are actually only using
>five wires, but again they are coded differently, and I've not the knowledge
>nor equipment to figure it out I think.
>
>

Yup, I think I might be able to do it, but it would take some effort. I
have
done this on some encoders that came with no documentation.

>
>
>>>My reservation against using the itty-bitty QEDS units I have
>>>is that they measure screw movement and not table position.
>>>
>>>
>>Yes, if your screw is of low quality, then anything read out by an
>>encoder on that screw will be no better.
>>
>>
>
>Exactly my feelings...
>I am considering heavily the option of ballscrews and nuts at this point, but
>I'm still thinking I want my positioning done by table, not screw position.
>
>

But, a good ballscrew and encoder can have MUCH higher resolution that
a high resolution linear encoder of considerable length. The most basic
(roughly .0005" but metric) linear encoders in 30" length run $350 at least,
new. That is really insufficient for smooth motion and decent precision
work. I am using actual resolution of .00005" on my system, with ballscrews
and shaft encoders. (Yes, that's 50 micro-inches.) I was lucky enough
to get
high accuracy ground screws off a scrapped machine. But, in a small
area, most ballscrews with shaft encoders can do better than a bottom
of the line linear encoder. Over the full length, the basic accuracy of the
linear scale will be better than a rolled screw and shaft encoder, of
course,
but these errors can be compensated for.

>For that matter, is it possible that I might be able to simply buy
>the reader heads and use them with the scales I've already got?
>At $25.00 each from US Digital, it makes
>a very reasonable solution to this problem.
>
>

Not likely. The US digital read head is designed for a 7-mil thick plastic
scale with 360 line/inch (I think). The glass scale will be a lot
thicker, and
the stripes will be roughly 500/inch (.002"). So, the US digital read head
would not work without extensive modification.

Jon