1. Hey guyz. Welcome to the All New Phlatforum!



    Sign Up and take a look around. There are so many awesome new features.

    The Phlatforum is a place we can all hang out and

    have fun sharing our RC adventures!

  2. Dismiss Notice

0.923 Acting weird

Discussion in 'SketchUcam Bugs' started by TigerPilot, Apr 10, 2010.

  1. TigerPilot

    TigerPilot Well-Known Member

    Offline
    Messages:
    1,578
    Trophy Points:
    48
    In the attached SU file you can see two fuse halves. The bottom one is the original and the top one is the bottom one copied and flipped on the green axis. If you look at the cnc file, you'll see that the bottom one is being cut perfect but the top one is going up and down, up and down in order to cut in the right direction, as if the face is reversed, which it isn't. I thought that the problem is because of the flipping on the green so I flipped on the red and the results are the same.

    I wonder if I'm the only one that run into this problem and if so, why? What can I do to make the top fuse cut clean like the bottom one. Attached files Rival 30 Fuse.skp (183.6 KB)Â Rival 30.cnc (27 KB)Â
     
  2. Jnida63

    Jnida63 Member

    Offline
    Messages:
    749
    Trophy Points:
    18
    I have run into that, I had to just redo each piece seperatly
     
  3. rcav8r

    rcav8r Moderator Staff Member

    Offline
    Messages:
    1,193
    Trophy Points:
    38
    I had a very similar issue, and it was because the face was not really flat...although it appeared to be so as it was a face.
    About 1/2 way down on this page KYYU has some screen shots explaining it.
    viewtopic.php?f=98&t=1732&start=25

    I don't understand why this would happen when flipping faces, but it sure does appear to be the same issue.

    Did you try to flip before applying the phlatcode? Not sure if it will help, but worth a shot.

    Also when this happened to me, I noticed the face flipped (changed color) as soon as the Phlatcode was applied. If I got that I knew to look for junk lines, or to flatten the piece.
     
  4. kyyu

    kyyu Active Member

    Offline
    Messages:
    1,183
    Trophy Points:
    36
    Location:
    Maryland
    Yea, it looks very similar to that issue that rcav8r had. But I really didn't know for sure why, at the time. So looking at what I said before, really won't explain it. I just tried a few thing which seem to fix it. I think what it comes down to is, redoing the phlatlines corrected the problem.

    I have a better guess this time, but really only the guy that wrote the code will know for sure. And that is gasmasher. I did look at tiger pilot's drawing and verified the problem. Redoing the phlatlines, after you flip the part, fixed it. Normally, in cases like this, I would use a much simpler drawing to check problems. So I just drew a simple square and flipped it and it still created the problem. My guess is the outside and inside phlatlines know the direction they were made. Or they know that the head of this line goes with the tail of the next. And when you flip the line, that cause the jumping back and forth. Because they still try to connect the old way.

    One thing I notice, that supports this, is if you reverse that very narrow face between the original outline and the outside phlatline. That fixed the flipped drawing, except the outside cut now goes in the wrong direction, but the gcode is correct. And if you do this to a good drawing, that is not flipped; then reversing that narrow outside cut face will cause the exact same corruption. So it was because the flipping, reversed the face and the face direction told the phlatscript to reverse the direction of cut. And reversing the cut direction, caused the problem because of the way the phlatlines are linked together.

    -Kwok
     
  5. TigerPilot

    TigerPilot Well-Known Member

    Offline
    Messages:
    1,578
    Trophy Points:
    48
    Thanks, guys, I'll know now how to fix the problem. It may be that flipping it before applying the phlatlines will prevent it but I was kind of lazy and wanted to save the trouble on a complicated part with lots of tabs and so on. Maybe a programer will address the issue in the future, but until then I'll gladly live with this bug. :D
     
  6. Flashsolutions

    Flashsolutions Active Member

    Offline
    Messages:
    1,123
    Trophy Points:
    38
    Location:
    Leesburg, Florida
    This problem seems to come up more and more. Unfortunately, it is difficult to duplicate. Every time I think I know what I did to cause it, and then try to repeat it, the problem disappears.

    Tim needs to know how to duplicate the problem so that he can fix it and thus far, none of us have been able to come up with a set of steps that will allow us to produce the problem at will.

    If anyone does find a hard and fast rule that can be applied to cause the problem, they need to contact Tim.

    These kind of problems are the worst to solve, because you need to be able to understand what caused it in order to fix and verify your fix really did solve the problem.
     
  7. kyyu

    kyyu Active Member

    Offline
    Messages:
    1,183
    Trophy Points:
    36
    Location:
    Maryland
    Flash, like I reported above, I can duplicate the problem. Just draw a square and do a outside cut. Then reverse the thin white face. That will reproduce the problem.

    Likewise, flipping a part in effect reverses that same face and hince I believe why flipping causes the problem. Again reverse the thin white face and that corrects the problem with the gcode jumping around to cut the segments. But of course, now the cut direction is wrong.

    -Kwok
     
  8. Flashsolutions

    Flashsolutions Active Member

    Offline
    Messages:
    1,123
    Trophy Points:
    38
    Location:
    Leesburg, Florida
    Must be more to it than that... I did exactly as you said and can not duplicate.

    1. Draw square
    2. apply outside cut
    3. generate gcode
    4. test
    5. reverse face
    6. generate gcode
    7. test

    Also tried
    1. Draw square
    2. apply outside cut
    3. generate gcode
    4. test
    5. reverse face
    6. delete all paths
    7. apply outside cut (applies inside cut due to reversed face)
    8. test

    1. Draw square
    2. apply outside cut
    3. generate gcode
    4. test
    5. reverse face
    6. delete all paths
    7. apply outside cut with shift key down to effect an outside cut
    8. test

    Also tried flipping along blue axis

    Aside from reversing the direction of the cut, it does not cause the problem which is
    as Tim explains it, a bug that causes it to cut the segments in one direction but selecting the wrong segment to cut next. For example if you had a line that looks like this with two segments (AB and BC):

    A B C
    .------.-----.

    Segment AB should cut from A to B and then BC should cut from B to C. The bug is triggered on that particular part and what happens is it cuts BC from B to C and then cuts AB from A to B. So it cuts them in the correct direction (A to B) but cuts them out of order. Tim says this has been a bug in the last few versions but it doesn't turn up often and he hasn't had time to look into it yet. He states that he is not sure how to work around it since he does not know what triggers it.

    So, if you can duplicate it, Please carefully record the steps you used, the version of the platscript and email gasmasher (Tim) with the details so that he may reproduce it.



     
  9. kyyu

    kyyu Active Member

    Offline
    Messages:
    1,183
    Trophy Points:
    36
    Location:
    Maryland
  10. Flashsolutions

    Flashsolutions Active Member

    Offline
    Messages:
    1,123
    Trophy Points:
    38
    Location:
    Leesburg, Florida
    Ahhh! That should be very helpful. I would never have thought to reverse the offset face like that. When you spoke about reversing the face, I assumed you to be reversing the large area of the face, not the face exposed after applying the outside cut path.

    You did say "thin" white face in your post, but it didn't sink in.

    The video also shows very clearly what I have seen occur quite often regarding the flipping of the objects to produce a reverse side such as the fuse.

    Hopefully Tim will be able to use this information to discover what is causing the problem.


    Thanks for putting that together Kwok!



     
  11. navionflyer

    navionflyer New Member

    Offline
    Messages:
    220
    Trophy Points:
    0
    Location:
    Richland, MI
    Thanks Kwok. This is the same problem I was having a while back but I didn't associate it with the reversed face.

    Thanks,
    Tim
     
  12. kyyu

    kyyu Active Member

    Offline
    Messages:
    1,183
    Trophy Points:
    36
    Location:
    Maryland

Share This Page