topic Re: in Visual LISP, AutoLISP and General Customization Forum

Getting Arc-midpoint in an easy way..

Anonymous — Mon, 27 Sep 1999 12:54:03 GMT

Hi there !
Is there some easy, reliable method of getting the coords for a midpoint
on an arc. I tried to do it this way:

1) Get the start-angle (DXF-code 50)
2) get the end-angle (DXF-code 51)
3) Get the radius (DXF-code 40)
4) Get the center-point (DXF-code 10)

From this data I just calculate the point using the Polar-function, like
this:

(setq Coords(polar Center (+ (/ (- endangle startangle) 2) startangle)
radius ))

This works, if the angles are in the 1:st and second quadrants, but I
encounter problems if the arc-span is between the 4:th and 1:st
quadrant..

Thanx !

/Erik

Re: Getting Arc-midpoint in an easy way..

Anonymous — Mon, 27 Sep 1999 13:09:21 GMT

Here's a quick and simple method using (polar):

(defun get-arcmid (enam / elst st-ang ed-ang rad cpt p1 p2 midpt)
(setq elst (entget enam)
st-ang (cdr (assoc 50 elst))
ed-ang (cdr (assoc 51 elst))
rad (cdr (assoc 40 elst))
cpt (cdr (assoc 10 elst))
p1 (polar cpt st-ang rad)
p2 (polar cpt ed-ang rad)
midpt (polar p1 (angle p1 p2) (/ (distance p1 p2) 2))
)
midpt
)

--
Phillip Kenewell
CAD Systems Technician
Air Gage Company
pkenewell@airgage.com

Erik Nilsson wrote in message
news:37EF68EB.5FD4D65D@systeam.se...
> Hi there !
> Is there some easy, reliable method of getting the coords for a midpoint
> on an arc. I tried to do it this way:

Re:

Anonymous — Mon, 27 Sep 1999 13:15:17 GMT

Correction - I forgot a line of code:

(defun get-arcmid (enam / elst st-ang ed-ang rad cpt p1 p2 midpt)
(setq elst (entget enam)
st-ang (cdr (assoc 50 elst))
ed-ang (cdr (assoc 51 elst))
rad (cdr (assoc 40 elst))
cpt (cdr (assoc 10 elst))
p1 (polar cpt st-ang rad)
p2 (polar cpt ed-ang rad)
midpt (polar p1 (angle p1 p2) (/ (distance p1 p2) 2))
midpt (polar cpt (angle cpt midpt) rad)
)
midpt
)

--
Phillip Kenewell
CAD Systems Technician
Air Gage Company
pkenewell@airgage.com

Phil Kenewell wrote in message
news:7snqdh$fuf3@adesknews2.autodesk.com...
> Here's a quick and simple method using (polar):

Re: Getting Arc-midpoint in an easy way..

Anonymous — Mon, 27 Sep 1999 13:45:56 GMT

I believe Phil's code is essentially the same as mine. However, I chose to
forgo the extra checking he did. Basically, I just calculate the start point
of the arc and a pply a midpoint snap to it. Hope it helps. P.S. There is no
error checking what so ever in this code.

(defun c:test()
(setq elist (entget (car (entsel)))
rad (cdr (assoc 40 elist))
cen (cdr (assoc 10 elist))
ang (cdr (assoc 50 elist))
mid (osnap (polar cen ang rad) "_mid")
)
(princ mid)
(princ)
)

Erik Nilsson wrote in message
news:37EF68EB.5FD4D65D@systeam.se...
> Hi there !
> Is there some easy, reliable method of getting the coords for a midpoint
> on an arc. I tried to do it this way:
>
> 1) Get the start-angle (DXF-code 50)
> 2) get the end-angle (DXF-code 51)
> 3) Get the radius (DXF-code 40)
> 4) Get the center-point (DXF-code 10)
>
> From this data I just calculate the point using the Polar-function, like
> this:
>
> (setq Coords(polar Center (+ (/ (- endangle startangle) 2) startangle)
> radius ))
>
> This works, if the angles are in the 1:st and second quadrants, but I
> encounter problems if the arc-span is between the 4:th and 1:st
> quadrant..
>
> Thanx !
>
> /Erik
>
>
>
>

Re: Getting Arc-midpoint in an easy way..

Anonymous — Mon, 27 Sep 1999 20:58:25 GMT

Erik,

Since you asked for an easy way, how about:

(setq mpnt (osnap (cadr (entsel "\nSelect arc: ")) "mid"))

This will work for any arc with the midpoint in any quadrant.
--
Dave E

Erik Nilsson wrote in article
<37EF68EB.5FD4D65D@systeam.se>...
> Hi there !
> Is there some easy, reliable method of getting the coords for a midpoint
> on an arc. I tried to do it this way:
>
> 1) Get the start-angle (DXF-code 50)
> 2) get the end-angle (DXF-code 51)
> 3) Get the radius (DXF-code 40)
> 4) Get the center-point (DXF-code 10)
>
> From this data I just calculate the point using the Polar-function, like
> this:
>
> (setq Coords(polar Center (+ (/ (- endangle startangle) 2) startangle)
> radius ))
>
> This works, if the angles are in the 1:st and second quadrants, but I
> encounter problems if the arc-span is between the 4:th and 1:st
> quadrant..
>
> Thanx !
>
> /Erik
>
>
>
>
>

Re:

Anonymous — Mon, 27 Sep 1999 21:55:38 GMT

Nicely done! I'm gonna have to start paying more attention to this group.

David Doane wrote in message
news:01bf092a$b32228e0$4456d2d0@lms-1...
> Erik,
>
> Since you asked for an easy way, how about:
>
> (setq mpnt (osnap (cadr (entsel "\nSelect arc: ")) "mid"))
>
> This will work for any arc with the midpoint in any quadrant.
> --
> Dave E
>
>
> Erik Nilsson wrote in article
> <37EF68EB.5FD4D65D@systeam.se>...
> > Hi there !
> > Is there some easy, reliable method of getting the coords for a midpoint
> > on an arc. I tried to do it this way:
> >
> > 1) Get the start-angle (DXF-code 50)
> > 2) get the end-angle (DXF-code 51)
> > 3) Get the radius (DXF-code 40)
> > 4) Get the center-point (DXF-code 10)
> >
> > From this data I just calculate the point using the Polar-function, like
> > this:
> >
> > (setq Coords(polar Center (+ (/ (- endangle startangle) 2) startangle)
> > radius ))
> >
> > This works, if the angles are in the 1:st and second quadrants, but I
> > encounter problems if the arc-span is between the 4:th and 1:st
> > quadrant..
> >
> > Thanx !
> >
> > /Erik
> >
> >
> >
> >
> >

Re:

Anonymous — Tue, 28 Sep 1999 00:13:51 GMT

Thanks.

Sometimes the more advanced programmers look deeper than they really need
to.
--
Dave D

Frank Oquendo wrote in article
<7sopgi$l8h5@adesknews2.autodesk.com>...
> Nicely done! I'm gonna have to start paying more attention to this group.
>
> David Doane wrote in message
> news:01bf092a$b32228e0$4456d2d0@lms-1...
> > Erik,
> >
> > Since you asked for an easy way, how about:
> >
> > (setq mpnt (osnap (cadr (entsel "\nSelect arc: ")) "mid"))
> >
> > This will work for any arc with the midpoint in any quadrant.
> > --
> > Dave E
> >
> >
> > Erik Nilsson wrote in article
> > <37EF68EB.5FD4D65D@systeam.se>...
> > > Hi there !
> > > Is there some easy, reliable method of getting the coords for a
midpoint
> > > on an arc. I tried to do it this way:
> > >
> > > 1) Get the start-angle (DXF-code 50)
> > > 2) get the end-angle (DXF-code 51)
> > > 3) Get the radius (DXF-code 40)
> > > 4) Get the center-point (DXF-code 10)
> > >
> > > From this data I just calculate the point using the Polar-function,
like
> > > this:
> > >
> > > (setq Coords(polar Center (+ (/ (- endangle startangle) 2)
startangle)
> > > radius ))
> > >
> > > This works, if the angles are in the 1:st and second quadrants, but I
> > > encounter problems if the arc-span is between the 4:th and 1:st
> > > quadrant..
> > >
> > > Thanx !
> > >
> > > /Erik
> > >
> > >
> > >
> > >
> > >
>
>
>

Re: Getting Arc-midpoint in an easy way..

Anonymous — Wed, 29 Sep 1999 00:06:21 GMT

To deal with angles you need always
to get into consideration the 2pi factor.

When you try to find the midpoint, you're
calling (polar ctrpt (/ (+ a1 a2) 2) r) where
a1 and a2 are starting and ending angle of an arc.
Here, would we add 2pi to one of angles, the result
will differ by 1 pi, which is what's causing an error
for you.

To deal with angles, I often use the function to
confine the periodic value to some given interval,
for example:

(defun make-pi (a) ; make it be in [-pi,pi] interval
(while (< a (- pi))
(setq a (+ a pi pi)))
(while (> a pi)
(setq a (- a pi pi)))
a )

Now it's safe to code

(defun midang (a1 a2)
(+ a2 (/ (make-pi (- a1 a2)) 2)))

and

(polar ctrpt (midang a1 a2) r)

So the key here is changing the angle value
to right interval _before_ dividing it by 2,
which could otherwise introduce a pi size error
(just what happened for you in that case).

HTH

On Mon, 27 Sep 1999 14:54:03 +0200, Erik Nilsson
wrote:

>Hi there !
>Is there some easy, reliable method of getting the coords for a midpoint
>on an arc. I tried to do it this way:
>
>1) Get the start-angle (DXF-code 50)
>2) get the end-angle (DXF-code 51)
>3) Get the radius (DXF-code 40)
>4) Get the center-point (DXF-code 10)
>
>From this data I just calculate the point using the Polar-function, like
>this:
>
>(setq Coords(polar Center (+ (/ (- endangle startangle) 2) startangle)
>radius ))
>
>This works, if the angles are in the 1:st and second quadrants, but I
>encounter problems if the arc-span is between the 4:th and 1:st
>quadrant..

--
Vlad http://www.netvision.net.il/php/vnestr/

Re:

Anonymous — Wed, 29 Sep 1999 13:30:10 GMT

Thanx all of U !
This helped a lot !

/Erik N

Vladimir Nesterovsky wrote:

> To deal with angles you need always
> to get into consideration the 2pi factor.
>
> When you try to find the midpoint, you're
> calling (polar ctrpt (/ (+ a1 a2) 2) r) where
> a1 and a2 are starting and ending angle of an arc.
> Here, would we add 2pi to one of angles, the result
> will differ by 1 pi, which is what's causing an error
> for you.
>
> To deal with angles, I often use the function to
> confine the periodic value to some given interval,
> for example:
>
> (defun make-pi (a) ; make it be in [-pi,pi] interval
> (while (< a (- pi))
> (setq a (+ a pi pi)))
> (while (> a pi)
> (setq a (- a pi pi)))
> a )
>
> Now it's safe to code
>
> (defun midang (a1 a2)
> (+ a2 (/ (make-pi (- a1 a2)) 2)))
>
> and
>
> (polar ctrpt (midang a1 a2) r)
>
> So the key here is changing the angle value
> to right interval _before_ dividing it by 2,
> which could otherwise introduce a pi size error
> (just what happened for you in that case).
>
> HTH
>
> On Mon, 27 Sep 1999 14:54:03 +0200, Erik Nilsson
> wrote:
>
> >Hi there !
> >Is there some easy, reliable method of getting the coords for a midpoint
> >on an arc. I tried to do it this way:
> >
> >1) Get the start-angle (DXF-code 50)
> >2) get the end-angle (DXF-code 51)
> >3) Get the radius (DXF-code 40)
> >4) Get the center-point (DXF-code 10)
> >
> >From this data I just calculate the point using the Polar-function, like
> >this:
> >
> >(setq Coords(polar Center (+ (/ (- endangle startangle) 2) startangle)
> >radius ))
> >
> >This works, if the angles are in the 1:st and second quadrants, but I
> >encounter problems if the arc-span is between the 4:th and 1:st
> >quadrant..
>
> --
> Vlad http://www.netvision.net.il/php/vnestr/