Friday, November 17, 2006


A Printable Penguin

Using the 2.5mm CAPA filament that Forrest had made up, I've produced a 50mm tall penguin for my upcoming presentation at LCA 2007:

It's extruding fine, and the spareseness of the fill is intentional at this point. The deposition is done with Simon's latest version of the code that raises and lowers the head to avoid smears. It works well, but there are small drawing artifacts that cause the head to bounce up and down in the penguin's feet, causing it to look like the RepRap is trying to hammer the polymer in!

Vik :v)

YES!!! Go Vik! :-D
For some reason it hit me that you should make an icing nozzle and do some cake decorating.
The Linux/opensource lobby are going to love you for that. Demand free-as-in-beer drinks! ;)
I recently uploaded a full-blown 3D Tux to the Wiki:

Printing *that* would be an achievement!
Doesn't look too difficult. It will need a support material head to fabricate, however.
Yep - exactly.
Next stop, a Cthulhu..rather more challenging.
Actually - isn't a 'Cthulhu' what you get when you are printing a Tux and your X/Y stepper motor drivers go haywire?
I think this technology has huge potential. It will be interesting to see how it progresses.
I had a look at the Tux model. It has no relief features like eyeballs, beak joins etc. I did have a look at addign them, but the model would not convert into a valid Euler 3D mesh under AoI.

Attempts to manually validate it were, shall we say, less than successful.

Vik :v)
"Attempts to manually validate it were, shall we say, less than successful."

...meaning...what exactly? That there are gaps in it? That the tool couldn't run it's validation check for some reason?

We can easily add 3D eyes and such - but there doesn't seem to be a 'blender' tool for finding gaps (if that is indeed your problem). I can write such a tool though (if that's what the "validation" problem is).

Actually, now that I look at the model, I think it's possible that one of Tux's flippers intersects his that a problem?

I guess we could use some guidelines for model contributors so we know what is and isn't acceptable to the next layer of tools. I rather assumed that the 'volumising' tools would be able to take care of small gaps and overlaps...but maybe not.
I'm writing an STL exporter for 'blender' in an attempt to make a simpler tools path. It's a ridiculously easy program - but I need to be sure I'm generating legal STL files.

I've been surfing the web to ascertain the details of the STL format (Jeez - that's a TERRIBLE file format!) - I've updated the Wikipedia entry for STL to reflect my findings:

I have a couple of questions that I can't find the answers to online though - maybe you guys could answer them for me so I can fix wikipedia and get my exporter working correctly:

* In one of the descriptions of the STL format it says that all coordinate values inside an STL file have to be positive numbers. Other descriptions of the format don't say that. This seems amazingly inconvenient. Is it true?

* In the binary version of STL, are values always stored little-endian or are there also big-endian STL binary files? You can theoretically guess which it is by comparing the 'number of triangles' value at the head of the file to the length of the file itself...but that's kinda messy!
The Tux model contains over 100 invalid edges. These don't matter in a virtual world, but when you try to produce the model in real-world 3D you end up trying to print a collection of small Klein bottles...

My manual attempts failed because I don't know how to remove points or edges from a 3D mesh in AoI and the things I did try did not remove the invalid edges. THe Solid Editor has a fix-up algorithm, but that was unable to handle all the errors. I've asked for some help on their SourceForge forum.

Vik :v)
OK - so there are multiple avenues of attack going on here and it's getting confusing!

* My son worked on the model - he's gotten rid of a lot of problems - but I doubt they've all gone. I'll re-upload it.

* I worked some on an ASCII STL exporter for blender. I have it working but I'm not sure about the two questions I asked earlier.

* We've found another blender STL exporter that may "just work" - but not having tried it yet, I don't know whether it generates ASCII STL or Binary STL.

* The Wiki doesn't like us uploading files which are (a) large (>2Mbytes) or (b) sufficiently like an image file to confuse it. This appears to be why some of the formats I uploaded won't download properly - and why other formats wouldn't upload in the first place. I think we need to re-think how works because Tux is far from being the most complex object we'll ever want! This means I can't upload an ASCII STL of Tux...although we could try something simpler.
You can always ZIP the ASCII STL.

Vik :v)
That's really cool... nice one Vik. I can't wait to get creative with this thing...

Oh, the temptation eD. I want to play, and yet we must finish the monster first...

Vik :v)
Incidentally, there's a GPL STL-file fixer here:

I don't know how good it is.


AdMesh looks interesting - I'll play with it some today.

I've been cleaning up the main page and added a 'list of tools' page (including your link to AdMesh). If there are other tools you know about that might be useful, it would be good to add them.
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to
Posts [Atom]