Vertically Stacked Prints

[Ryan]

Over the last month or so I've been working on a method to increase
production of my Makerbot 170. Making 608 idler pulleys is pretty
time consuming for such a small part and there are a lot of pieces to
print if I want to rep-strap a Mendel. A conveyor belt that pops
parts out the back would be ideal, but this is a ways off. My
solution is to stack parts on top of each other. It turns out that
you can effectively separate two parts using a little over 1mm between
them such that they can be pulled apart by hand. Of course parts that
won't sit on each other won't stack. In between the parts I add a
layer much like the raft interface which I'm calling a "float" for
lack of a better term (for the time being I map out this layer in a
spreadsheet and write it out to text file).

After some experimentation I've automated the process of converting a
skeinforge produced gcode into a stack. The Perl script is called
MakeNBreak.pl and is documented over at the wiki: http://wiki.makerbot.com/makenbreak
As I say several times this is really just a first step and I'm
surprised that it seems to work as well as it does.

So far I've managed to take the 608 production code and create a code
to make 7 at a time. This cuts about 3 minutes off the build time for
the raft and decreases failures due to warping. As far as I can tell
they're the same as individually printed parts, and a couple days ago
I sent the gcode and instructions over to the makerbot guys. I've
also managed to make a stack of 3 frame vertex. I figure you can do 6,
but it's taking between 3 and 4 hours to do 3. Also, I tried to print
whistles, but my skeinforge settings aren't airtight apparently so
they didn't work. I'd also like to try the makerbot coin or a
derivative. You could probably print a mess of those at once.

Try it out and give me some feedback.
Ryan

[jet]

Ryan wrote:
> It turns out that
> you can effectively separate two parts using a little over 1mm between
> them such that they can be pulled apart by hand. Of course parts that
> won't sit on each other won't stack.

But as long as it's smaller than the part below it they should print,
right? So if you had N objects, where On+1 is smaller than On, you
could just stack them: O1, O2, ..., On, right?


--
J. E. 'jet' Townsend, IDSA
Design, Fabrication, Hacking
design: www.allartburns.org; hacking: www.flatline.net; HF: KG6ZVQ
PGP: 0xD0D8C2E8 AC9B 0A23 C61A 1B4A 27C5 F799 A681 3C11 D0D8 C2E8

[Tox]

Nice! I'll give this a shot.

Tox


--
Scott Small

[Ryan]

Jet, right now I'm only thinking of repeating the same part over
again, but in theory that's right. It's possible to you could print a
series of whistles from large to small.

[TeamTeamUSA]

This is a great idea, and could conceivably be turned into a
skeinforge tool: stack.

Doing this would optionally add it to the skeinforge tool chain and
make it available in the course of creating any gcode file on any
platform; a double win.

Since you know perl and have the core logic, porting this to python
and converting to a skeinforge tool shouldn't be too challenging. If
you decide to do this, use skeinforge-0006 and wipe.py as a starting
point. Once you get it running in 0006, then you can tackle the beast
that is 0007.

I wrote my first skeinforge tool recently, so let me know if you need
any tips.

Go!

=ml=

[Ryan]

Thanks for the encouragement. I'd like it to go into skeinforge, but
I've never written in python. As an engineer most programing I do is
like self-defense. Also, I need to think about how to make it fail
gracefully if it doesn't see a good way to create a stack.

[Ian Fischer]

If you don't feel comfortable doing it, I imagine one of the many CS
people who are on this list might do it. I may take a stab at it
myself once I get over this cold... (Don't let that stop anyone else
from picking this up, though!) Good work on this, by the way.

Ian

> --
>
> You received this message because you are subscribed to the Google Groups "MakerBot Operators" group.
> To post to this group, send email to make...@googlegroups.com.
> To unsubscribe from this group, send email to makerbot+u...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/makerbot?hl=en.
>
>
>

[Jeremy Ruhland]

That's awfully clever, when I get some time I'll see if I can compile a set of (stackable) stacked mendel parts.

--
Jeremy Ruhland
Jeremy....@gmail.com
Ruhlandpedia.dyndns.org

[Ryan]

Ian, that's my hope if this catches on that someone would be good
enough to translate it for skeinforge. I am interested in python now
since it seems to be the choice more and more often.

Jeremy, I'm planning on doing that as well. It appears that over half
the parts are stackable, those that aren't usually only require one or
two to be printed anyway.

Ryan

On Dec 1, 1:37 pm, Ian Fischer <ian.fisc...@post.harvard.edu> wrote:
> If you don't feel comfortable doing it, I imagine one of the many CS
> people who are on this list might do it.  I may take a stab at it
> myself once I get over this cold...  (Don't let that stop anyone else
> from picking this up, though!)  Good work on this, by the way.
>
> Ian
>

[jet]

Ryan wrote:
> Ian, that's my hope if this catches on that someone would be good
> enough to translate it for skeinforge. I am interested in python now
> since it seems to be the choice more and more often.

Is skeinforge really the place to do this? It seems like it should be
a general tool that could combine different g code instructions by
changing the Z-offsets.

[Tox]

On Wed, Dec 2, 2009 at 10:39 AM, jet <allar...@gmail.com> wrote:
> Ryan wrote:
>> Ian, that's my hope if this catches on that someone would be good
>> enough to translate it for skeinforge.  I am interested in python now
>> since it seems to be the choice more and more often.
>
> Is skeinforge really the place to do this?   It seems like it should be
> a general tool that could combine different g code instructions by
> changing the Z-offsets.
>

A "Height" value in the multiply tab would seem reasonable - it
already expands to a 2d matrix, this would bring it to a third.
Guessing the interface layer parameters would get buried in Raft.

Tox


--
Scott Small

[TeamTeamUSA]

Ryan already has his perl script which is a standalone solution.
Conceivably it could be extended to support stacking multiple
different gcode files. This would probably be a good thing.

Creating a skeinforge tool would integrate stacking easily into the
skeinforge workflow/tool chain, and provide a familiar UI from which
to change parameters.

In addition to a stack tool, a distribute tool to layout the same
gcode file horizontally within a given area would be nice too.

Both could feasibly be combined into one skeinforged file to optimize
printing time and material consumption: stacked and distributed.

Go!

=ml=

On Dec 2, 11:13 am, Tox <toxwander2...@yahoo.com> wrote:

[jet]

TeamTeamUSA wrote:
> Ryan already has his perl script which is a standalone solution.
> Conceivably it could be extended to support stacking multiple
> different gcode files. This would probably be a good thing.

[...]

> In addition to a stack tool, a distribute tool to layout the same
> gcode file horizontally within a given area would be nice too.

My vote (we're voting?) is for a separate tool, that way people aren't
tied to skeinforge for stacking/arranging parts. (Note that skeinforge
is unreliable on G5 macs and possiblly other platforms.)

Also, I am biased: once I'm dialed in and get a few simple projects off
the ground, I'll probably be moving more into parametric modeling and
generating my own g-code.