topic curve chamfer in Visual LISP, AutoLISP and General Customization Forum

curve chamfer

Anonymous — Mon, 09 Mar 2020 15:18:38 GMT

i need a lisp which can make (multiple chamfer) in one command,

also i need it to solve the issue of making chamfer between curve and line

Re: curve chamfer

CodeDing — Mon, 09 Mar 2020 15:44:05 GMT

@Anonymous ,

The CHAMFER command already has a "Multiple" option within the command. This will allow you to do more than 1 selection while in the command.

Also, if you're working with 1 entire polyline that you would like to Chamfer, you can use the "Polyline" option.

What is the 'issue' you are having between a curve & line? Can you provide an example? Give more detail? or post an example drawing?

Best,

~DD

Re: curve chamfer

Anonymous — Mon, 09 Mar 2020 15:51:30 GMT

i want to select all polylines in same command to make all the vertices chamfered with specified distance...

even if one of the vertices is curve not a polyline

Re: curve chamfer

Kent1Cooper — Mon, 09 Mar 2020 16:02:28 GMT

@Anonymous wrote:

.... the issue of making chamfer between curve and line

Are you aware that you can Chamfer between a curve and a line if they are an arc segment and a line segment of the same Polyline? [It even works if they're not adjacent segments, though the result can be pretty wild.] HOWEVER, doing so changes the route of the arc segment, which I assume you would not want. Starting with the red Polyline on the left [duplicated in dashed dark red on the right so you can see the original shape]:

it seems to find a point as if Chamfering the line segment with the chord of the arc segment [dotted yellow], and changes the arc segment to hit that point [green result]. I don't know how it determines the new route of the arc segment to hit that point -- I expected it would keep the original's bulge factor, but that's not the case, nor does it keep the starting or ending direction of it.

BUT the question arises: If something can be worked out to calculate a point on an Arc to Chamfer to, would that distance need to be measured along the curve of the Arc, or as a straight-line [chord] distance? The difference would be small in most cases, if the Chamfer distance is small relative to the Arc's radius, but there would be a difference.

Another question: By "curve" do you mean only Arc or perhaps Polyline arc segment? Or might you want to use it with an Ellipse or Spline?

Re: curve chamfer

Anonymous — Mon, 09 Mar 2020 16:20:35 GMT

its not what i want , this way change the curve ....

Re: curve chamfer

Kent1Cooper — Mon, 09 Mar 2020 16:47:27 GMT

@Anonymous wrote:

i want to select all polylines in same command to make all the vertices chamfered with specified distance...

even if one of the vertices is curve not a polyline

You have a contradiction there. A "curve not a Polyline" would presumably need to be an Arc [or maybe Spline or Ellipse], which won't be included in a selection of Polylines. Do you really mean "even if one of the segments within a selected Polyline is an arc segment, not a line segment"?

Also, your image doesn't Chamfer all corners of the one with the arc segment [if that's what it is]. I set up a similar situation, and the Chamfer command applied to a Polyline like that of course does the other corners, but also does both ends of the arc segment, which does not appear to be the situation in your image:

[And note that the arc-segment's path is more significantly altered.]

Even if a way can be found to preserve the path of the remaining curve, if you would not want both ends of an Arc or arc segment Chamfered, by what criteria would a routine be able to decide which end to do?

Re: curve chamfer

Sea-Haven — Mon, 09 Mar 2020 23:41:06 GMT

For the arc segment a manual approach is draw circle with rad, then trim pline, erase circle, draw line from ends use pedit join back in. So a arc segment check is possible. Maybe use this method for all chamfers.

Re: curve chamfer

Anonymous — Tue, 10 Mar 2020 04:09:42 GMT

Yes thats true this is the manual way and thats what i need to create in a
lisp so as to make all of this moves in one command

Re: curve chamfer

Sea-Haven — Wed, 11 Mar 2020 11:19:20 GMT

Try this change the rad value or make it a getreal. No error checks.

(defun c:champl ( / rad plent x pt ent obj st end oldsnap) (vl-load-com) (setq rad 30) (setq oldsnap (getvar 'osmode)) (setq plent (car (entsel "\nPick pline"))) (setvar 'osmode 0) (if plent (setq co-ord (mapcar 'cdr (vl-remove-if-not '(lambda (x) (= (car x) 10)) (entget plent))))) (repeat (setq x (length co-ord)) (setq pt (nth (setq x (- x 1)) co-ord)) (command "circle" pt rad) (setq ent (entlast)) (command "trim" ent "" pt "") (command "erase" ent "") (setq obj (vlax-ename->vla-object plent)) (setq st (vlax-curve-getstartpoint obj)) (setq end (vlax-curve-getendpoint obj)) (command "line" st end "") (command "pedit" plent "join" (entlast) "" "") (setq plent (entlast)) ) (setvar 'osmode oldsnap) (princ) ) (c:champl)

Re: curve chamfer

john.uhden — Thu, 12 Mar 2020 17:01:09 GMT

It works!

Re: curve chamfer

Anonymous — Thu, 12 Mar 2020 18:53:08 GMT

now i can tell that you know what i want but am sorry it still not the
perfect solution or i can say it got some errors,
first of all when i apply this on a new cad file on one rectangle or
something it worked very well but when if i draw alot of rectangles (even
if they are not intersected) and try to apply this on them i got an error
or i dont know the process stops anyway in the middle of the action....
second by considering this rectangle as a residential block with 4 vertices
in the end which i need to chamfered them so i need only 4 chamfers for the
4 points of rectangle, but when one of the ends is arc it will make the
rectangle with 5 vertices that the lisp will chamfer them and thats not
what i want, i want to chamfer the 4 ends only if you know what i mean

Re: curve chamfer

Kent1Cooper — Thu, 12 Mar 2020 19:25:22 GMT

@Anonymous wrote:
.... by considering this rectangle as a residential block with 4 vertices
in the end which i need to chamfered them so i need only 4 chamfers for the
4 points of rectangle, but when one of the ends is arc it will make the
rectangle with 5 vertices that the lisp will chamfer them .... i want to chamfer the 4 ends only if you know what i mean

Going back to your image at Message 3:

The shape with a curve on it does not show the 3 right-angle corners Chamfered in the "after" part, but I assume you want them to be. But the bottom edge of that is clearly made up of both a line segment and an arc segment, so the Polyline in question has [at least] 5 segments & 5 vertices, however much you may consider it to have 4 edges. Are you saying that where an arc segment is tangent to an adjacent line segment, as near the middle of that bottom edge, you don't want that vertex Chamfered? [Not that I can imagine easily how a routine would make that distinction, but it may be possible....]

Re: curve chamfer

Anonymous — Thu, 12 Mar 2020 20:02:08 GMT

Yes in case of the rectangle has an arc i need to chamfer only the 4 ends
of the rec. Not the whole 5 vertex of the rectangle

Re: curve chamfer

john.uhden — Thu, 12 Mar 2020 23:26:54 GMT

I like your approach because it does exactly what a user should do to obtain the desired result.

But I have 2 observarions:

1. entmake the circles with '(60 . 1) so they are invisible.

2. Instead of

(command "trim" ent "" pt "") try (command "trim" ent "" (list plent pt) "")

because with potentially overlapping polylines the pt may pick up a different entity.

I was also concerned that trimming a piece of a polyline might result in creating two polylines, but that seems not to be the case.

Again...nice work.

Re: curve chamfer

Sea-Haven — Fri, 13 Mar 2020 08:02:56 GMT

Thanks for the trim method

Unless you run it multiple times on user objects then you find the headaches. Its a quick and dirty way I acknowledge that, it would probably be better to use a "Break" function using pline segment but code will triple in size and after all its all done for free.

I know you can see the circle thought it was just interesting to watch, using entmake for the circle should speed it up also.

Re: curve chamfer

Anonymous — Fri, 13 Mar 2020 09:13:32 GMT

sorry i cant understand the 2 observations although i applied the second to
change the command trim as u said i did it but i dont know what changed i
still facing the same issues,
1) if there are many rectangles need to be chamfered it cause an error , it
only works if the file has one rectangle,

2) still when the rectangle has an arc which represent 5 vertices not 4
still i dont want to chamfer the fifth vertix i need to chamfer the 4 end
points only of the rectangle as shown in attached jpg file ( the clouded
area i dont want it to be chamfered like the others)

Re: curve chamfer

Kent1Cooper — Fri, 13 Mar 2020 15:26:29 GMT

@Anonymous wrote:
....
2) still when the rectangle has an arc which represent 5 vertices not 4
still i dont want to chamfer the fifth vertix i need to chamfer the 4 end
points only of the rectangle ....

First, if it includes an arc, it's not a rectangle, but let's say "outline" or something.

Second, would I be correct in assuming that "the fifth vertex" is not always going to be specifically the fifth one in terms of the drawn order of vertices of the outline? That is, that a routine could not just step along the Polyline vertices starting at the first one, and do its chamfering thing for the first four vertices and leave the last one alone, because sometimes the vertex you don't want chamfered could be somewhere other than the last one? To me, that's the big challenge -- to get a routine to determine that a vertex is at a tangent continuation between a line segment and an arc segment, or the reverse, no matter where in the sequence, and to therefore ignore that vertex.

Also, might there ever be more than one arc segment, such as this 6-vertex outline?

or that an arc segment could be tangent at both ends, such as this, so that you would want to chamfer only 3 corners?

Re: curve chamfer

Anonymous — Fri, 13 Mar 2020 16:32:32 GMT

So u think there is no solution for that ??##- Please type your reply above
this line -##

Re: curve chamfer

Sea-Haven — Fri, 13 Mar 2020 23:24:46 GMT

A dwg would help to work out what is happening, I will check the multiple not working.

Re: curve chamfer

john.uhden — Fri, 13 Mar 2020 23:27:39 GMT

Oh, c'mon. There's almost always a solution. It's usually just a matter of scraping your knees to find one.

Where we usually fail in the engineering approach is to define the problem. Once the definition is clear, the design becomes more apparent, despite the fact that you may need umpty-seven conditions to account for varying possibilities.