Extend/Shorten Contour = added code?

[mike_f]

I'm also firmly in the 1 line camp.

I've been burned by blindly trusting verify more than once, and while single blocking through an entire program might be wasteful, I believe there are certain areas where a "new" program should be single blocked. distance to go can't be my friend if it's sailing by so quickly I can't respond to it.Besides, shouldn't it work in mill as it does in lathe ?

[Troy Gossett Indepak Inc.]

This is silly fix it.

[Allan]

Turn your filter on; the only extra co-linear moves you will get are the lead in moves. They will not change that because that is where the comp occurs. It is not broken, there is nothing to fix IMHO.

 

Allan

[mike_f]

Allan, I've never thought of this as a huge issue, but turning on my filter has no effect. I've tried multiple settings and none of them affect the code output that I get, I still get extra lines of code.

[Allan]

Check that again, turning on the filter joins all co-linear moves. When using the extend contour option and a tangent lead in I get one lead in line move.

[Scott Garrison@Weiler Engineering]

You might get one "lead-in move" but how many total?

 

I just tried using both extend contour and tangential lead-ins (separately) with a facemill, no comp on either, just center of the cutter.

 

I still get (3) X-moves with each way.

 

I had the filter on as well, isn't that filter for arcs anyway?

 

To me, the extend/shorten feature should work regardless of comp/no-comp. That's what the lead-ins are for. The lead-ins should be applied after the extend/shorten contour is taken into account.

[DavidB]

Allan

I've never seen the filter reduce the g-code for lead INOuts on a face toolpath or a contour toolpath with tangent leads with NO arcs.

 

The only way to get 1 line of g-code is to extend the geom.

[Allan]

So if I toolpath a line and extend the contour, I will expect 3 moves total, One for the lead in (joined with any extend I apply) one for the line and one for the lead out (joined with any extend I apply). These moves are seperate because in the NCI the first has start of contour information and the last has end of contour information. This is the way it is.

 

If you want to force one move extend the geometry and use no lead in/out.

 

The filter not only creates arcs where possible it also filters out co-linear moves, try breaking your line into many pieces with the filter on youll get three moves with it off youll get two plus however many pieces that you broke your line into.

[Scott Garrison@Weiler Engineering]

Ok, we see that filtering doesn't work for tangential lead-ins or extend/shorten contour and we know the workaround.

 

Would it be that hard to get the NCI to add those moves by changing the start & end of contour info internally and then filter it? Maybe I'm oversimplifing but it doesn's look that tough to me to fix it.

 

Scott

[Allan]

Could be done in the post something like:

 

code:

---

plinout         #Output to NC of linear movement - feed

if (nextx$ <> x$) & (nexty$ = y$) & (cutpos2$ = 1 | cutpos2$ = 4), subout = 1

if (nextx$ = x$) & (nexty$ <> y$) & (cutpos2$ = 1 | cutpos2$ = 4), subout = 1

 

pcan1, pbld, n$, sgfeed, sgplane, `sgcode, sgabsinc, pccdia,

pxout, pyout, pzout, pcout, `feed, strcantext, scoolant, e$

 

if subout$ <> 0, subout$ = 0

clearsub

 

---

That should do it for simple moves where the leadout is parallel to x or y. It will however affect your ability to use subprograms.

[mike_f]

I'm just curious...why does this function work ( at least the way some of would like it to ), in lathe programs ?

[Scott Garrison@Weiler Engineering]

Allan,

 

How does it shorten contours??? Wouldn't it have to change both the start & end locations?? Unlike extend where you can add moves, this HAS to change the basic contour start//end locations.

 

Scott