Filter NCI chook

[Guest]

Not sure how many of you use/used this feature but I was just updated that the function is not being brought back....

Quote

Hi John,

There are no plans currently to revive the FilterNCI C-Hook, so I logged enhancement request R-14050 to see if it can be brought back.

Thank you!

Ticket Status changed from In Progress to Resolved

It's been re-added as an enhancement request but seeing as though bringing it back was passed over, I am thinking the chances of seeing this back are about nil.

It leaves me wondering if by not bringing back the ability to filter these paths whether this is a signal that the wireframe toolpaths themselves may be deprecated in the not too distant future.

 

Edited September 11, 2017 by Guest

[Thad]

18 minutes ago, JParis said:

It leaves me wondering if by not bringing back the ability to filter these paths whether this is a signal that the wireframe toolpaths themselves may be deprecated in the not too distant future.

 

I hope not.

[Matthew Hajicek - Singularity]

Everything useful that is not new and shiny will be depreciated.

[Guest]

Well, in response to my idle speculation

Quote

I checked with the team and here is their response. "No plans to retire the Wireframe toolpaths. If anything we will bring them up to a modern interface." I'm checking to see if the modern interface includes filtering. I'll be back when I get an answer. Thanks!

 

[gcode]

I recall a poll some years ago about retiring the wireframe toolpaths.

There was a rapid retreat once the results came in,  though I have to say I personally

have used one of the wireframe surfacing paths once in the last ten years.

The old legacy surfacing toolpaths are an entirely different matter.

I don't use the rough toolpaths much anymore, but for many applications,

it is impossible to beat Flowline.

[huskermcdoogle]

Before I knew how to use the moduleworks paths, I used to use wireframe and classic surfacing paths quite a bit.  Don't think I have used either in about a year now other than to post them out on old projects with minor updates to other paths.  Once you get those old paths massaged into what you want, they are hard to beat.

On the subject of filternci, I wish they would give us the ability to filter the moduleworks paths.  If every toolpath in the program had the filter suite enabled, I think it would be fair to take it away.  In the mean time, sometimes it is just plain hard to get a smooth path without filtering code to some level.  I have had to write some nci filters of my own that I run in excel with vba, then importnci back into x6 to check for erroneous motion and post out.

[pullo]

1. post the file and post out the nci file

2 .open 2017 and with the Import NCI activated import the nci into 2017. (2018 dies immediately if you try to backplot this toolpath , 2017 is just fine)

3 .backplot and convert to geometry.

4 . using 2d profile chain the 3d  geometry and with compensation off you can filter anything using the profile's chaining filters.

 

A parallel toolpath with a 0.015mm precision weighing in at 924 KB filtered with 0.045mm precision lost weight down to 390 KB

 

If you need to do it in 2018/2017 , it can be done No need for reviving anything. 

Gracjan

[pullo]

It's called the McGyver approach to Mastercam 

Gracjan

[huskermcdoogle]

7 hours ago, pullo said:

with the Import NCI activated

I guess its time to get this activated again in 2018, if they will do that for me as offered before.

[huskermcdoogle]

7 hours ago, pullo said:

1. post the file and post out the nci file

2 .open 2017 and with the Import NCI activated import the nci into 2017. (2018 dies immediately if you try to backplot this toolpath , 2017 is just fine)

3 .backplot and convert to geometry.

4 . using 2d profile chain the 3d  geometry and with compensation off you can filter anything using the profile's chaining filters.

 

A parallel toolpath with a 0.015mm precision weighing in at 924 KB filtered with 0.045mm precision lost weight down to 390 KB

 

If you need to do it in 2018/2017 , it can be done No need for reviving anything. 

Gracjan

This is fine for toolplane type paths.

[pullo]

I left the obvious  out  Husker

checked out the filterNCi in X5 and it  also worked on 5-ax nci. 

Well anyway here goes McGyver 2018 for 5x-afficionados. 

I created a 5-ax  toolpath  in 2018 , created it's NCI file AND THEN PROCEEDED TO USE THE x5  filterNCI on the 2018 nci file and it worked !  Although something weird happened . I had to INCREASE the accuracy to DECREASE the size of the 5-ax file. I am doing this for the sheer fun of doing it, but

I'd be careful here.... but jumping back in time is possible (X5 circa 2010) . I guess you could use this solution for 3-ax files too.

Gracjan

 

[Watcher]

4 hours ago, pullo said:

I had to INCREASE the accuracy to DECREASE the size of the 5-ax file.

Not weird at all. We assume that a coarse tolerance will generate less points, but filters like small tolerances because they are able to convert a lot of colinear moves to lines, and with linearized arcs, the center points match each other within a small tolerance (See it as a square window centered in the arc center, where the length of the sides is the filtering tolerance and all arcs centers must lie within this window).

Mathematically, an arc can be defined by 3 points in a plane, and the filter checks the next two points and see if an arc can be built, and if so, where the arc center lies. If it matches the center of the previous arc within a given tolerance, then it connects the new arc segment with the previous one, and it does that thousands of times if necessary, turning thousands of points into a single, accurate arc. This is the essence of how the filtering algorithm works. That´s the reason when you have linearized toolpaths, small collinear segments tend to be converted to lines, or linearized arcs into real arcs. Large tolerances make the motion to be to erratic and the filter cannot make the connections, so they keep the motion intact.

 

 

[pullo]

I was talking  about the 5-ax path file size reduction Watcher, so arcs are not the thing here .  I went rather quickly yesterday thru this stuff, but today  I played around  a bit and the 

X5 filter seems to work fine with a path created in 2018 or rather it behaves in the same way with an original X5 path and one created in a newer version of Mcam..  So if Mama Bear don't give the ol' nipple, here is this one.

Gracjan

[gcode]

You've  got to be careful filtering 5X toolpaths.

NCI filters are neither part nor stock aware, it's just filtering points and can easily violate your part.

I got badly burned filtering legacy 5X toolpaths in early X days

Unfiltered the files were huge and choppy.

Filtered they ran clean and smooth, but violated the part

 

[Newbeeee™]

On ‎12‎/‎09‎/‎2017 at 2:47 PM, gcode said:

The old legacy surfacing toolpaths are an entirely different matter.

I don't use the rough toolpaths much anymore, but for many applications,

it is impossible to beat Flowline.

+1 BillTrillion.

And I still do not know what anyone was thinking, when they decided that the new finishing toolpaths should rollover edges...

[Watcher]

2 hours ago, pullo said:

I was talking  about the 5-ax path file size reduction Watcher, so arcs are not the thing here .  I went rather quickly yesterday thru this stuff, but today  I played around  a bit and the 

X5 filter seems to work fine with a path created in 2018 or rather it behaves in the same way with an original X5 path and one created in a newer version of Mcam..  So if Mama Bear don't give the ol' nipple, here is this one.

Gracjan

Sorry, I tried to use your point to make a general explanation about how filters work, and in 5 axis most of possible filtering is done for collinear moves, as explained above too.

 

[pullo]

No harm here  .  However  I share gcode's point about being careful with filtering 5ax code. Actually  I am totally the wrong person to preach about it as I just  yesterday  found out about  5ax code filtering ,courtesy of this fact finding mission.  

I have always fine tuned my 5-ax paths either by: 1. tweaking the inputs during the programming , 2. if possible tweaking the geometry 3. tweaking my CYCLE DEF 32.0 parameters

(rounding of the toolpath on the control) .  For those blessed with a Heidenhain , You can  give the  A  value (angle ) in CYCLE DEF 32.0 a value of  2.0 degrees without fear of violating the part. This info came courtesy of one German dealer over a beer in Hungary (X9 meeting) and it is stated so in the  530 manual .  I thought I was reckless working with 0.05 degs :).

Gracjan

 

 

 

[Matthew Hajicek - Singularity]

13 hours ago, newbeeee said:

+1 BillTrillion.

And I still do not know what anyone was thinking, when they decided that the new finishing toolpaths should rollover edges...

Has anyone here ever wanted a surface toolpath to roll over edges?  Never in my 22 years of programming have I ever wanted this.

[Newbeeee™]

40 minutes ago, Matthew Hajicek™ - Conventus said:

Has anyone here ever wanted a surface toolpath to roll over edges?  Never in my 22 years of programming have I ever wanted this.

This has been asked quite a few times, and I can't recall a reason for anyone wanting it...

[pullo]

The topic seems to have changed direction and now the old saying 180 just doesn't do it justice.  But here goes:

The reason why I use old school toolpaths is that with Contour you have control over it . You can roll all over Beethoven or not , you CONTROL it . With High Speed it's roll over and nothing else.

So to recap it both are useful as long you have a choice.

Gracjan

 

 

 

[Matthew Hajicek - Singularity]

Except that with the new versions you have to jump through hoops to contain the toolpath, whereas the old ones just work.