30116pb03, Panel 14x14x2.667 with Millennium Falcon Pattern

[Daniel Goerner]

Can someone please check these files if I have overseen anything? Edger2 says the subfile is okay, but who knows?
Maybe some advice what I could do better (sorting the colors or whatever). And any help with the Header would be appreciated (Filename, Tags maybe?)...

[Orion Pobursky]

When did I make that "without faces" subpart? 8 years ago? And you beat me to this pattern. lol. Good job man!

[Philippe Hurbain]

Very good job indeed!

Edger2 says the subfile is okay

Well... not quite, I see 4 red lines that at closer look are caused by needle shaped gaps.

And LDDP does report a few errors (see below).

Code:

Line 32: Quad collinear (angle = 179.972756397366): 4 16 279.75 56 -40.9 277.239 55.083 -43.509 275.819 55.086 -48.649 275.184 55.088 -50.951
Line 32: Quad points not coplaner (angle = 7.29169626473399): 4 16 279.75 56 -40.9 277.239 55.083 -43.509 275.819 55.086 -48.649 275.184 55.088 -50.951
Line 37: Quad points not coplaner (angle = 5.72276032249363): 4 16 220.519 24.409 -32.432 273.579 53.063 -42.512 279.75 56 -40.9 220.517 24.407 -32.428
Line 43: Quad points not coplaner (angle = 3.74873199767183): 4 0 220 24 -31.4286 219.7523 24.00001 -32.32164 220.449 24.37128 -32.419 220.5167 24.40729 -32.42815
Line 107: Quad points not coplaner (angle = 3.4744190042692): 4 16 204.751 24 -86.398 208.011 29.414 -100.995 210.525 30.771 -101.745 211.741 31.426 -102.101
Line 113: Quad points not coplaner (angle = 1.73648810918471): 4 16 230.611 47.914 -131.22 222.962 50.54 -154.126 223.561 50.822 -154.149 224.104 51.071 -154.148
Line 153: Quad collinear (angle = 179.950153191285): 4 320 185.594 27.471 -132.186 186.593 28.086 -132.767 186.164 27.296 -130.557 180.756 24.49 -129.366
Line 153: Quad points not coplaner (angle = 8.83407612376923): 4 320 185.594 27.471 -132.186 186.593 28.086 -132.767 186.164 27.296 -130.557 180.756 24.49 -129.366
Line 183: Quad points not coplaner (angle = 2.99758107262041): 4 16 212.124 44.983 -151.99 211.549 44.614 -151.598 204.994 43.32 -158.015 219.625 49.843 -157.25
Line 309: Quad collinear (angle = 179.938276819139): 4 16 205.928 23.34 -72.879 205.147 23.341 -75.71 204.995 24 -85.519 207.471 23.339 -67.309
Line 309: Quad points not coplaner (angle = 5.95382889817837): 4 16 205.928 23.34 -72.879 205.147 23.341 -75.71 204.995 24 -85.519 207.471 23.339 -67.309
Line 431: Triangle collinear (angle = 0.0215881040453166): 3 0 149.974 8.235 -62.301 192.22 20.41 -81.11 193.076 20.655 -81.471
Line 602: Quad points not coplaner (angle = 1.13742545117368): 4 16 34.736 28.948 -228.727 34.231 27.003 -225.076 32.42 24.404 -220.513 35.281 31.204 -232.975
Line 616: Quad points not coplaner (angle = 1.94343745738227): 4 16 98.143 52.16 -256.386 97.549 51.376 -255.022 45.531 31.202 -230.127 100.003 54.645 -260.714
Line 788: Quad points not coplaner (angle = 1.52608539348902): 4 0 50.341 14.921 -179.357 54.981 20.98 -201.691 55.189 21.212 -202.541 58.631 24 -212.454
Line 792: Quad points not coplaner (angle = 7.65794381496949): 4 16 55.189 21.212 -202.541 31.19 21.903 -211.889 31.316 22.147 -212.807 31.802 23.064 -216.246
Line 813: Quad collinear (angle = 179.951873944033): 4 16 23.864 8.254 -160.709 48.374 13.593 -174.727 47.547 12.673 -171.369 46.777 11.819 -168.254
Line 813: Quad points not coplaner (angle = 7.4258139162545): 4 16 23.864 8.254 -160.709 48.374 13.593 -174.727 47.547 12.673 -171.369 46.777 11.819 -168.254
Line 885: Quad points not coplaner (angle = 1.1101878957145): 4 0 113.882 14.297 -146.256 113.326 14.057 -145.535 129.422 21.296 -167.899 129.939 21.52 -168.571
Line 953: Quad collinear (angle = 179.914789113821): 4 16 61.974 8 -149.149 58.005 7.633 -148.344 52.601 7.634 -149.846 49.571 7.635 -150.693
Line 953: Quad points not coplaner (angle = 5.91787820710034): 4 16 61.974 8 -149.149 58.005 7.633 -148.344 52.601 7.634 -149.846 49.571 7.635 -150.693

Most can be split into appropriate triangles to solve the problem .
As for the header, I suggest the use of Datheader... that will not only correct headers issues (missing space in front of 2.667, ö of your name converted to UTF8) but also most of previous problems (only 2 very thin triangles remain).

How did you do projection on surface? Using Slicerpro? then maybe playing a bit with "precision" and/or unification thresholds would have eliminated some of the issues from the start?

[Daniel Goerner]

Very good job indeed!

Edger2 says the subfile is okay

Well... not quite, I see 4 red lines that at closer look are caused by needle shaped gaps.

That must have happened after I used Slicerpro. Before using it everything was alright.

And LDDP does report a few errors (see below).
[...]
Most can be split into appropriate triangles to solve the problem .
As for the header, I suggest the use of Datheader... that will not only correct headers issues (missing space in front of 2.667, ö of your name converted to UTF8) but also most of previous problems (only 2 very thin triangles remain).

I now used Datheader on the original Patterm, before projecting it. 6 Errors remain:

Code:

Line 87: Colinear vertices found. 4 6 23.86372 0 -160.7092 48.374 0 -174.727 47.547 0 -171.369 46.777 0 -168.254
Line 163: Colinear vertices found. 4 7 61.97413 0 -149.1488 58.005 0 -148.344 52.601 0 -149.846 49.571 0 -150.693
Line 338: Colinear vertices found. 4 4 185.594 0 -132.186 186.593 0 -132.767 186.164 0 -130.557 180.756 0 -129.366
Line 560: Colinear vertices found. 4 6 205.928 0 -72.879 205.147 0 -75.71 204.9948 0 -85.519 207.471 0 -67.309
Line 618: Colinear vertices found. 4 7 279.7504 0 -40.89976 277.239 0 -43.509 275.819 0 -48.649 275.184 0 -50.951
Line 754: Colinear vertices found. 3 1 149.974 0 -62.301 192.2205 0 -81.11035 193.076 0 -81.47137

I don't know what this means (like the stuff you posted, remember, I never did this before). I think there are vertexes placed on top of each other. But I don't know how to find, correct or eliminate them.

How did you do projection on surface? Using Slicerpro? then maybe playing a bit with "precision" and/or unification thresholds would have eliminated some of the issues from the start?

Yes, i used Slicerpro. Very cool tool. I made the part with stickers (using Sticker Generator) before and it was torture to place them right. I'll try to play around with the settings tomorrow.

[Philippe Hurbain]

That must have happened after I used Slicerpro. Before using it everything was alright.

Yes, there are some less-than-perfect things in Slicerpro, and sometimes it easts a triangle or two... I have a slightly better version than the one on my website, but not yet published since some problems remain.

I now used Datheader on the original Patterm, before projecting it. 6 Errors remain:

Here too I may be the culprit. LPC uses a slightly modified version of Rectifier to assemble triangle pairs into to quads. The first 5 errors with very flat angle quads should not happen (except if the triangles in your pattern had very flat angle). The last problem comes from a very sharp angle triangle that you created in your pattern (needle shaped triangle). This should be avoided, there are generally ways of organizing triangles differently to avoid this.

Coud you share the .lpc file of your pattern? (rename it to .txt so that it's accepted on the forum, or email it directly to me...)

[Daniel Goerner]

You've got mail.

Oh, and Chris, if you're reading this, could you give me a partno. for this?

[Daniel Goerner]

Short question: Can I split a quad into two triangles again, and if so, how?

[Philippe Hurbain]

Yes, with a text editor... For example:

Code:

p1                       p2                    p3 (too flat angle)   p4
4 6     23.86372 0 -160.7092     48.374 0 -174.727     47.547 0 -171.369     46.777 0 -168.254

gives

Code:

p1                       p2                    p3
3 6     23.86372 0 -160.7092     48.374 0 -174.727     47.547 0 -171.369
        p3                       p4                    p1
3 6     47.547 0 -171.369        46.777 0 -168.254     23.86372 0 -160.7092

or

Code:

p2                       p3 (too flat angle)   p4
3 6     48.374 0 -174.727        47.547 0 -171.369     46.777 0 -168.254
        p4                       p1                    p2
3 6     46.777 0 -168.254        23.86372 0 -160.7092  48.374 0 -174.727

Of these two possibilities, the 1st one is the right since it splits in half the very flat angle.

[Travis Cobbs]

Doesn't LDDP have a command to split a quad into two triangles?

[Philippe Hurbain]

Doesn't LDDP have a command to split a quad into two triangles?

No, it's the other way around (merge 2 triangles into 1 quad). Splitting poses a UI problem (especially at text level) since there is a choice to do between the two possible splits...

[Daniel Goerner]

What the...? I just had it down to 0 errors. And then I changed the colors in LDDP. After saving the file there are 3 new errors. I just changed the ... colors!

Code:

Line 35: Coplanarity. 4 16 220.519 24.409 -32.432 273.579 53.063 -42.512 279.75 56 -40.9 220.517 24.407 -32.428
Line 107: Coplanarity. 4 16 204.751 24 -86.398 208.011 29.414 -100.995 210.525 30.771 -101.745 211.741 31.426 -102.101
Line 782: Coplanarity. 4 16 55.189 21.212 -202.541 31.19 21.903 -211.889 31.316 22.147 -212.807 31.802 23.064 -216.246

What is this and how can I correct it? This drives me crazy...

Oh, and for some odd reason there appears a optional line in the file, although neither the pattern nor the 3D file contains any...

[Philippe Hurbain]

Ah - yes. Another LDDP quirck. It does auto rounding, something BAD on intermediate files. On really tiny quads, rounding can warp things enough to go above recommended warp (1°) or even above maximum warp (3°). For more background on coplanarity, see LDraw specification, and documentation of Planarcheck.
This is not a big issue, since it's easy to correct by splitting the warped quad (a triangle is always planar
Planarcheck and Datheader can do it for you automatically. I'd recommend Planarcheck here, because Datheader always inserts a condline between the triangles after the split, something not useful here (the surfaces are supposed to be planar!). Planarcheck also defaults to add condline, but there is an option to prevent that. Maybe the mysterious condline you see was added during a Datheader check?

Now let's come back to LDDP. In Tools > Options > Configurations values tab, there is a check box that prevent LDDP to do automatic rounding. I strongly suggest you to do that... Unfortunately, there is also a bug: when you do a color change, rounding occurs regardless of the checkbox!!! Because of this I mostly do color changes using MLCad... You may also increase the number of digits of "position decimal accuracy" to 5 digits.

Only in the last step, try to reduce precision to 3 digits (the recommended value for parts), and run Planarcheck to clear coplanarity errors.

BTW, looks like the parameters tweak in slicerpro was successful? I see no more gaps!

[Travis Cobbs]

I guess that makes sense. However, with non-rectangular quads, a "better" split can be determined (the one where the seam line between the two triangles is shortest). For concave quads, there's only one valid way to split. Of course, if the quad is significantly warped (non planar), then perhaps an automatic splitting algorithm isn't such a good idea.

[Daniel Goerner]

First of all: Thank you Philo. You're a great help for a noob like me.

BTW, looks like the parameters tweak in slicerpro was successful? I see no more gaps!

I didn't tweak anything in slicepro. The first 5 gaps were eliminated by changing the colors of the triangles so they could not be merged into quads. And the for the last one I changed the colors also so the triangles (which had the gap between them) would be merged to a quad on purpose, which I manually split into triangles again. Therefore my question a few posts ago.

Only in the last step, try to reduce precision to 3 digits (the recommended value for parts), and run Planarcheck to clear coplanarity errors.

Is there a program you recommend to do the rounding? Because I just did that with DATHeader and now I have gaps and a lot of optional lines again (See attached file)...

[Chris Dee]

30116ps0.dat

[Daniel Goerner]

Thanks. The Subfile would be 30116ps0s01 then?

[Daniel Goerner]

So, I went back to LPC to rearrange some triangles.

And now I think I'm finally done. 0 Errors in LDDP, DATHeader or Edger2, as well as rounded to 3 digits.

But please check again.

[Philippe Hurbain]

First of all: Thank you Philo. You're a great help for a noob like me.

A pleasure, especially when the "noob" learns fast like you

Is there a program you recommend to do the rounding? Because I just did that with DATHeader and now I have gaps and a lot of optional lines again (See attached file)...

Mmhhh... the optionnal lines are easy to get rid of, just select one in MLCad, then "select same type" and delete. But there are also a few problems reintroduced by rounding - You've got really tiny surfaces that gets easily warped by rounding. Simplest solution, don't round!!! Probably you could have avoided these tiny surfaces by increasing Slicerpro vertices unification threshold to 0.05 or so. No big deal, the unrounded surface is fine!

The "gaps" are not really gaps but just two roundings that behaved differently on adjacent surfaces. Not really annoying (they disappear if you raise Edger2 precision value, to 0.002). You might also suppress them using unificator, but that's not really needed here.

[Philippe Hurbain]

Ah - the posts crossed...
One word for your new file: Perfect !!! Congratulations

[Daniel Goerner]

Thank you.

The correcting part was exhausting and took almost longer than actually creating the pattern, but it was totally worth it. I just submitted the files.

[Daniel Goerner]

A pleasure, especially when the "noob" learns fast like you

Oh, I just have learned to use more tools. What all these error messages mean is still beyond me.

[Philippe Hurbain]

What about the torso?

[Daniel Goerner]

I'll have to check it first but I'm busy making the next part. Maybe this weekend.

[Daniel Goerner]

Another quick question: How do I cut a part in half or, lets say, a round part into 8 pizza slices? ;-)
Anything special I have to do if I want to use these as subparts, like edges etc?

[Steffen]

look here: http://www.philohome.com/ldraw.htm

you can use intersector to cut it,
or you could use SymSplitter

[Philippe Hurbain]

When it's possible to use Symsplitter (cut only by planes), Symsplitter should be preffered: it gives better cut quality, and it does eliminate unwanted stuff automatically (with Intersector you have to sort things by yourself manually).

[Daniel Goerner]

Yes, I just don't get how Symsplitter works.

I want this


to be split like this:


What do I need for this, since I was only able to split the file in half? How does a splitterfile for this have to look like. Do I need lines or something?

Sorry, I worked the whole weekend to correct all the errors in the file above, so I'm a little exhausted.

[Philippe Hurbain]

Ah - for a pattern it's even simpler! Slicerpro will project things only if there is something to project on... So if you make a former file containing one of the colored sectors in your image and project the whole pattern on that, the projected pattern will be nicely cut on the boundaries of the sector.

Sorry, I worked the whole weekend to correct all the errors in the file above, so I'm a little exhausted.

Sorry for you too - but your pattern with gradations looks gorgeous!

[Daniel Goerner]

Oh Boy, I din't try that, because I thought Slicerpro would put the complete pattern on the slice. Silly me.

[Daniel Goerner]

When I mirror a subfile in the partfile, do I have to add something concerning BFC?

At this moment the body of the partfile looks like this:

Code:

1 16 0 0 0 1 0 0 0 1 0 0 0 1 s\30366s01.dat
1 16 0 0 0 1 0 0 0 1 0 0 0 1 s\30366px4s01.dat
1 16 0 0 0 -1 0 0 0 1 0 0 0 1 s\30366px4s01.dat
1 16 0 0 0 1 0 0 0 1 0 0 0 1 s\30366px4s02.dat
1 16 0 0 0 -1 0 0 0 1 0 0 0 1 s\30366px4s02.dat
1 16 0 120 0 -1 0 0 0 -1 0 0 0 1 s\30366px4s02.dat
1 16 0 120 0 1 0 0 0 -1 0 0 0 1 s\30366px4s02.dat
1 16 0 0 0 1 0 0 0 1 0 0 0 1 s\30366px4s03.dat
1 16 0 0 0 -1 0 0 0 1 0 0 0 1 s\30366px4s03.dat

[Philippe Hurbain]

No, a subfile has its own BFC, nothing to do. Only surfaces (triangles and quads) are directly concerned. If you mirror a surface, the result changes winding, so you have to restore proper orientation.

[Daniel Goerner]

And I thought the creation of the pattern took long. The correcting took almost longer. I did not find any errors in the files above (using LDDP, DATHeader, Edger2 and Planarcheck), but since I'm looking at these parts for the past two days now, please, can someone else check them again before I submit them?

[Daniel Goerner]

I just checked the Qui-Gon Torso and Edger2 says 16 unmatched edges. But these are around the front and the back, not in the Pattern. Is that normal for a Torso or did the template (from LPC) I used not work correctly?

[Philippe Hurbain]

It's expected for a torso. You would need to expand subfiles in Edger2, but the torso itself has some mismatches in its structure. If you don't see red lines in the middle of the pattern (and here there are none), it's OK.

[Philippe Hurbain]

I'm speechless! This is pure awesomeness. I understand that the creation of this pattern took forever!
The only problem I see is that the conditionnal lines of the front are missing. The simplest thing to do is to copy them from 30366 into 30366ps0 (technically it would be better to copy only the relevant condlines in each subpart, but it's not as easy because of mirroring).

[Daniel Goerner]

If you're happy, I'm happy too. I copied the condlines from 30366 to my file.

[Daniel Goerner]

Okay, thanks.

[Philippe Hurbain]

(almost) perfect! Comment should be "0 // Conditional Lines".
Again - Wow!

[Daniel Goerner]

"Lines do not end with <CR><LF>"

Um, what?