Sample HiRes STL output file

[Ken_Shea]

Loaded fine in OneCNC XR5 this time Art.

[ArtF]

Ken:

  Thx. Its good to figure this stuff out early on. Most file formats have their up's and downs. STL is really a bad format, it has mutiple vertices in it instead of
indices tables. Coudl be much more efficient really..

I think I have it right now though, it was just a typo in the ouput code, rather simple format to do really. Only the question is, why doesnt slicer load it?
Ill have to experiment, but Im thinking it may be the size is just too large, though that seems funny, its not that big. I plan on allowing a single STL output
of entire mechanisms, so the triangle count would be in the millions. Im hopeing general 3d printing software will allow such large files. I should have my
printer around the end of the month to use in testing, so Ill know better then I guess.

Thx fer the test
Art

[Richard Cullin]

new file loaded into netfabb studio ok,
as you say it crashes slic3r  i used netfabb to save it as a binary stl , which again crashed slic3r
the beta version of the latest slic3r (9.02 i think) accepted both versions happily , i did not try to print it as the resultant gcode looked a bit too small to be valid .
do you really need that sort of resolution / can a 3d printer really resolve that much  detail anyway

[ArtF]

Hi Richard:

Thanks for the test. Must be an issue with size then. You say the Gcode was too small, does that mean it seems
like its incorrect?
  Im not sure the ideal resolution, but unlike most things one can build, gears need quite a bit of resolution to be truly
accurate enough in many cases. Ill likely allow lower resolutions to be made in the end, the ones Im generating now are
maximum res based on what GM put out as contour data in the project file. GT doesnt yet create anything, it loads a
GM project file and re-extrudes them to create a proper stl. My problem in GM was that a licensed library created the
STL's, which were bad resolution if made too large, GT is my own extruder with unlimited res, but the files as a result
are quite large at the moment.

Art

[Ya-Nvr-No]

Nice clean HiRes file lines, loads fine now.

Did have quite an issue getting Assimp to install and work but did manage finally, that's a cool tool too. Exports a very nice .obj file.

The major issue was the DirectX 9 installation. Had to extract the dxwsetup.exe to a folder and then run as admin.

Thanks again.

[ArtF]

Thx..

  3D printers are now commonly in the 150 - 250 micron resolution. The example helical is about the same in terms of its resolution, which is one of the reasons
I selected it for a test. Doesnt seem logical to me to create models with less resolution than the printer itself. We can always have it decimate to an acceptable level for a particular user, but I want to be able to tell any user that his STL will always have enough data to make a clean gear.

Assimp is pretty powerfull and a nice way to convert formats around but not for the faint of heart to get running. A problem Ive noted for most unix written 3d processors.  Ive been playing with Slicr and Im pretty sure its a memory issue there. On the other hand I no
longer am of the opinion that the files are overly large. They are , after all, simply ascii files, its the underlying data thats important, and at 256K vertices,
thats not really that much for a proper 3d file. I can allow for stack setting to control the resolution and make them more corse, but I think when you have a
sweeping path such as a helical you need lots of information to give that swept curvature. Yoda's and dragons for example run into the multimillion vertex sizes.
Bevel will likely be around 500K vertices when done to any acceptable level of resolution.

  Slicr , while it seems a pretty good little program, is written in perl of all things, so it likley has trouble with certain memory allocations. Its display for example is horrible
and could be much better. Ill wait till I have more experience with actual printing before I decide what to do about that situation. I know most 3d printers come with
software for printing so Im not sure yet what standards are in the field. We'll definitely talk more on this topic in the future though as I struggle to educate myself to the level of you 3d printing guys..

  So now that helicals are proper, Im redoing the bevels for the same level of detail, that and a couple of fixups and GT will be at the stage where its basically
a hi-res stl recreator for the low res components of GM's output. Thats the first stage in its development, and probably when Ill release an alpha so that
at least Ill know users CAN have such stl's if they require them. (Something Ive felt bad about since I discovered just how bad the large gears were in stl form..)


  I think I see the direction Im headed and we'll discuss that more after an initial release of the postprocessor.

Thx
Art

[Ya-Nvr-No]

A programmer with a conscience  :o
Your a real find, and a very impressive gentlemen.
Thanks for sharing your loves & interests with us.
I'm sure I speak for many, that without your generosity we would never know what is possible for us hobbyist & hackers.

[John S]

Bobcad V24 and Cut 3D both opened the file up OK.

What's Gearotic thought ?
new version of Gearotic or just the next stage ?

[ArtF]

Hi John:

  Gearotic Thoughts is really to add functionallity I could no longer add. The STL's from gearotic were not high enough quality
and that has bothered me since I found that out about 6-8 months ago. The SolidCore library i used simply couldnt extrude
more points and its limit wasnt enough to make hi tooth count. Also , the helicals and bevels ( which most of the 3d printers
wanted) couldnt be done form their STL's. So I decided to write my own 3d stl based program. In its first stage it will load a
GM project, re-extrude the gears to hi-res toleances and prepare them for 3d printing. All of us have seem 3d printed mechanisms
that move after printing. ( bearings and such), but the truth is such things are very hard to build unless your a CAD/CAM expert,
or at least very good at it.
  SO stage 2 of Gearotic Thoughts will be the modification fo models to include things liek keyways and collars, and modifiction
of shafts and gears so that proper clearances are in them, and tie them all into one STL file that cant print as a complete mechanism.
  Stage 3 is to add 2d CAD capability so that a person can draw and extrude shapes and mechanisms of their own design. Cant promise
exactly what it will do, just that thats the road Im currently walking. I tend to let the road lead me to whatever place I end up.

  Gearotic Thoughts will be a free addon to GM and Ill start releaseing Alpha's and beta's as soon as I figure it has enough
functionallity to be usefull.

:)
Art

[John S]

Bollocks,

Now gotta go buy a 3D printer - sigh...................

[ArtF]

lol..

  Look at the bright side, you loved the laser you got. You may want to wait till Ive had time to play with 3d printing though, Ive been very
slow to jump in because I truly dislike many of the photo's I see. Poor resolution, step artifacts and such led me to believe quality is
a big issue. Thats why I spent over a year experimenting to see if I could design my own with everything from ink to laser. I failed. 

  The quality appears to be getting much much better though on FDM. Polymer UV seems the highest resolution, but I suspect we
can all imagine what priting resin smells like and is like to clean up. Supports in the print are a concern, hard to say how hard they
are to deal with, also hard to say how hard it woudl be to control some of these machine without the manufacturers software..one reason
the one I ordered isnt a reprap , I want to see the current software protocols and how they function. I want to push the envelope of a
machine like that to see if its a toy or if it really fits properly in a shop and can make good bevel gears or not. ( I must admit, one of my
priorities is bevel gears, their too hard to make, and take too much prep, a 3d priter may be able to solve that one for me as a programmer
since Id be very pleased to at least be able to say one can 3d print a nice working bevel. So I'll let you know soon enough what I like and dont like
and will be able to tell you if its a shop necessity, or just cool to play with. If it DOES work, I have a 3' x 3' robotic arm Ill make into a huge 3d printer
by adding an extruder to it.

:) , bah, you probably want one anyway, stop being cheap.... lol

Art

[BobL]

Bollocks,

Now gotta go buy a 3D printer - sigh...................

Join the club John...lol


Bob

[John S]

lol..

 Look at the bright side, you loved the laser you got.

:) , bah, you probably want one anyway, stop being cheap.... lol

Art

The laser is awesome.
Just bought a new milling machine to replace this POS Bridgeport, still a Bridgeport type turret mill but on steroids, the knee weighs about as much as the POS.

But what to call it ?
The POS Bridgeport was famous over 18 forums as being made out of licorice.

So I thought...... and though..... and then thought of you Art.



;D

[ArtF]

Stop linking your damn photo's.. , upload them so I can actually see them. :)

Art

[John S]

Updated the pic.