wire offset direction is wrong

[Guest]

i have a file that the wire is offsetting the wrong direction when you window contour.never had the problem before.can i send this file to some of you wire guys and see if you can duplicate the problem.i use the color mask,window contour the geometry,chain is set to ccw,offset is to the left,compensation is set to control.one file does correctly and another similar one does not.i can't figure it out. thanks del

[Storkman]

Turn your monitor over, that'll fix er.

[edmBosto]

Del send her over and I'll try to take a looky

[Guest]

quote:

---

Turn your monitor over, that'll fix er.

---

storkman, ok, i did that.now what.

 

 

edmbosto, mail is coming your way.

[edmBosto]

I have your file

[edmBosto]

Del,

Just real quick here on my part.....

 

It is doing the same thing for me (chaining to the right)

My 1st thought is maybe cause of the open contour.

[Guest]

edmbosto,

 

yep, i have another file almost exactly the same that goes to the left.thanks for looking.

[edmBosto]

Oh man I ferget to turn me monitor upside down.

[Nathan G. Baldwin]

Del,

Drop me a copy of that file as well and I'll have a looksie.

 

Thanks,

[Roger Martin from CNC Software]

I have some info about this.

It has to do the the distance between the Thread Pt and the start of the chain. The shortest distance wins...

This is working now (in X2) the same as it has since (at least) v9.1 - as I get the same results in either.

 

I can post up a more detailed explanation tomorrow if needed, but I have to leave now.

[Guest]

roger,thanks for taking the time to lok at it

 

it's not a big deal as long as i pay attention.it would be nice though if you could control it.

[Roger Martin from CNC Software]

Del,

 

You can control it, by the placement of the Thread Pt.

On your sample file, the Thread Pt appears to be equal distance from end of the possible (chain start) end points. But waaay down in the digits (this is floating point math ) one of those distances is 'closer' than the other. And it this case it happens to not be the one you wanted.

Easy enough to bias the system to the 'other' end point. Select - Only - Points (window the Thread Pts), Xform-Rotate (move) and rotate the points CCW by 0.001 deg. Now do your usual Window chaining of the contours and the chains are going the 'other' direction.

[edmBosto]

Roger,

 

Thank you very much for that information.

It will be very useful for me. And others!

[Guest]

roger, thanks for that tip.that explains it very well. i knew there had to be some reason,i just couldn't figure it out why one file did what i wanted and the other one did't.damn floating point math.