[dorkbotpdx-blabber] Silhouette Cameo update

[Erik Walthinsen]

Thought you all might be interested in what I've just found:

graphtecprint actually successfully drove the Cameo from Linux. The
code is horribly organized and not documented at all, but it looks like
everything is actually there. I have a stencil cut for a board I just
ordered that looks mostly OK.

<https://github.com/jnweiger/graphtecprint>

I'm going to do some experiments to see if I can actually square up the
cuts that are happening by overshooting the corners (the blade rotates
along the vertical axis and "drags" to change direction, afaict). It'd
actually look something like a badly compensated scope probe (since I
just did that on mine...).

With a major restructuring of the code (the gui is deeply entangled in
the driver code and not appropriate to a Cameo anyway), I think it could
very easily work to do batch stencils. ;-)

[Erik Walthinsen]

On 04/04/2013 01:54 PM, Erik Walthinsen wrote:
> I'm going to do some experiments to see if I can actually square up the
> cuts that are happening by overshooting the corners (the blade rotates
> along the vertical axis and "drags" to change direction, afaict). It'd
> actually look something like a badly compensated scope probe (since I
> just did that on mine...).

I have to share this update, since it blew me away it worked so well:

I just cut a perfect 0.2mm square in 3mil Mylar, using careful
blade-angle pre-cuts and overshoot. It lifted off the adhesive mat with
the fragment still on the mat. This is done from my own very simple
python script.

It shouldn't be very hard then to get *insanely* accurate stencils made
straight from Gerbers (just need to look into existing ways to get
convex hulls out of them, or write them).
_______________________________________________
dorkbotpdx-blabber mailing list
dorkbotpd...@dorkbot.org
http://music.columbia.edu/mailman/listinfo/dorkbotpdx-blabber