Re: NAN Bug with circular moves

[Ray Henry <rehenry-at-up.net>]

Jon +

I copied Jon's sample hole bore code from his site and it runs with the 
new emccanon.cc without any modification.  The use of r words and small
arcs is still fine.

The NaN Ian found shows up for helical z moves where an xy arc is greater 
than 180 degrees and the z move has to be computed for it.  IMHO -- make 
that a guess -- older nc code should still work because we didn't define 
arcs larger than 180.

If you've got old helical or spiral programs but haven't got the new 
canon or backplot, post the relevant code and we can test run it.  Virtual
crashes are relatively easy to recover.

Ray

At 06:20 PM 2/9/2000 -0500, Jon wrote:
>
>
>Ian Wright wrote:
>
>> Hi Will,
>>
>> I finally got round to compiling in your 'fix' file and playing around with
>> the system again. The program you suggested runs fine and has obviously
>> sorted out the initial problem, however, being of an adventurous nature I
>> seem to have found yet another way of goofing it up! - sorry.
>>
>> Being still a relative dimwit at this cnc stuff and yet being dazzled by
the
>> apparent magic with which these things are bestowed, I reasoned that, now
>> the tool can be made to describe a complete circle with only such a simple
>> command, could it also do something really clever like machine a circular
>> spiral - as in producing a snail cam for example. So I stuck a 'z'
component
>> in and guess what - same problem again and nothing told me I couldn't do it
>> before the (virtual) tool ascended into orbit with the 'x' and 'y' axes
>> trazzing off at a 45 deg. angle at high speed. Try it - the line I used was
>> g2x5y0z1i-5j0 in place of your second line, a simple change but very
>> effective!!!  ;o)
>
>Oh, oh!  This is helical interpolation, and is necessary for milling threads.
>I have a program that writes RS-274D code that uses this, and it worked
>some months ago, last time I used it.  If we have loused up helical
>interpolation, then we need to get it fixed.
>
>As for spirals, there is no standard command for that.  There are special
>codes used on various controls for spirals, but I wrote a program that
>approximates a spiral with circular arcs in quadrants, to open up
>a hole in something.  See the makebore program on my web page
> http://www.artsci.wustl.edu/~jmelson/gcode.html .  The first
>program is the one that does this.  The sample run repeats the
>expanding spiral at two Z levels, and then makes a final, full circular
>pass to clean up the bore.
>
>You can extract the spiral generating logic from this program, and
>make a few simple adjustments to make it do smaller segments, if
>you want to cut real spirals with accurate curves.
>
>Jon
>
>
>
>