This is interesting. The timeframe is about 6 +/- 1 years ago:
What have I been doing with my Sherline CNC mill system? Well,
bowhunting season wiped out 3 months of my spare time... but I'm back
on it now and you may find this interesting. Keep in mind I am completely new to machining...
I'm writing my own CAD software for this... I decided I want
something I can just keep adding functionality to as problems come up.
Maybe there's some way to do most of what I want in BobCAD, but I had
problems understanding how to make even simple parts with the v18 trial
I downloaded without hours of tinkering. I was very uncertain about how
much was just my lack of understanding and how much was just lack of
functionality. I didn't want to spend that much money and just hope the
functionality was there. The lower-priced or shareware/freeware
packages I tried were so horrible I wasn't going to take anything for
granted.
The funny thing is I just bought an expensive math program to help
me solve equations for this CNC software's algorithms... If I spent a
little more and I could have bought BobCAD... but now I'm hooked on
writing my own stuff.
Here are a few test parts I have cut:
They're not functional, just fun. Why not metal? Well, early on I
figured out that the leftover plexiglass I used to build a box for the
mill is much cheaper than metal to play with. Wood and plastic are
cheaper, easier, and less stressful to learn on. At first, I was
terrified of destroying the mill with a typo in NC code. :)
My first long-term project goal is to make a mostly flat wall clock
in wood. The idea is for all the gears to be exposed, lying as spead
out and flat as possible. Here's a screenshot of some test gears in my
software:
Here are the results on the mill:
Just in case you're interested in the properties page for a gear object...
I like seeing parts of the screen exactly the size they will be
(from experimentation, 109.2 DPI on my laptop screen is perfect). You
can see the gears in a shaded work area which represents the 'safe'
zone given my equipment. Under that is the mill tooling plate, and the
mill table itself.
The screenshots actually show the cut-out section. The gear object
algorithm was the first thing I did... and the software wasn't
sophisticated enough at that point to take an arbitrary outline and
figure out how to cut it out. So, the algorithm directly calculated the
cut path only, and by displaying that I could see what was left. The
screen always shows exactly what will be cut... the drawing process
itself is similar to the EMC backplot, in that it uses the NC code
itself to draw tool paths. For clarity, it just doesn't show anything
with Z > 0. The black progress bar on the left side with "Line
5026/5026" next to it tells you this is what the final piece should be.
You can grab the black progress bar and scroll it up and down to see
how it is going to be cut it out. The NC code is 5000 lines long
because the spokes are first approximated into very very small lines,
and are then duplicated inside to cut the pockets. (cutting more than
the outline is not necessary for the final part, but it looks a lot
better on the screen).
My current short-term project goal (they help me develop the
software and my machining knowledge towards the clock project) is to
make a chinese chess set. My girlfriend is Chinese, and her parents are
visiting from China for a few months. I figured it would be a nice
gift. Anyway, here is a screenshot of one of the four pages of chess
pieces:
I haven't cut them out yet because my program cannot currently cut
out the inside of the text letters. I've cut a single piece, but the
rough inside of the outlines made it look bad after my girlfriend
painted it. (No, it wasn't her paint job!) I could draw tool path lines
inside the letters manually and do it, but I want the program smart
enough to cut out arbitrary pocket shapes for me.
One really interesting bit (to me, anyway) with the chess set is
going to be the cover of the box. I want two Chinese characters for
Elephant Game (the Chinese words for Chinese chess) to be cut into the
cover in light hardwood inset into a dark hardwood. In order to do
this, the border of the Chinese characters must be modified so that no
corner is taken too sharply for the given tool radius. I've got a
"Round to tool radius" algorithm in the works... this screenshot is of
that.
It's incomplete... I need to write the algorithms for correcting
line-arc and arc-arc vertexes... so simple letters like these are all
it does correctly now.
Last, I'll throw in this screenshot... which shows various ways to
cut a shape out with a 1/8" end mill. The thinner red lines are just
there to show the tool path.
Anyway... I've got a -long- way to go with the software, but I decided to show you where I'm at.
NOTE: I did manage to cut out two Chinese chess pieces, but no more. My girlfriend (wife, now) painted them both, and they looked quite good. The tool pathing problem was more symbolic math than the poor MathCAD program could take. None of the other pieces looked good without a better tool path solution. Mathematica (what I should have bought) cost too much and was rejected by my girlfriend. I thought that might happen, but I then I also thought I'd compromise. Don't try to save money by compromising on tools, if you have the choice. It rarely works out. I don't really regret it though. Relationships are much more important.
Besides, it's almost always the process that matters. I learned a lot about a field completely new to me at the time. The equipment even paid for itself (eventually).
For CAD these days I use Alibre CAD. It's very intuitive (as just about all SolidWorks-like systems with even a handful of tutorials is), and it has WAY more functionality than I was ever planning on writing myself. It's more or less as functional as a couple-year-old SolidWorks package (which is damn good). The only problem (and it's potentially a big one) is that nobody is going to be able to read your files. Well, there's always IGES and STL. I basically work alone so it's not a problem here.
Anyway... I've decided there's not much of a future in the software mentioned here for any commercial purpose, so if anyone is interested I've placed the code and a Windows installer on SourceForge, under a BSD open source license. I will probably make another blog post in the next few weeks describing how to use it in more detail, and releasing some sort of rough user manual.
The gear creation is very simplified. It works, but there are much better ways to create gears.
The code may have a few glitches, as it was in transition from MFC+GDI calls to a custom cross-platform framework. (The Sherline mill came with Linux-based CAM software) These days I'd absolutely use Qt for that sort of thing (I may eventually port it). The tool pathing is also probably in some random state of partial broken-ness. That said, I have loaded it up and used it on several occasions to cut holes for various panel mount things.
It should not (never has for me) generate any catastrophic NC code (like ramming the Z axis into the stock at high speed)... but in any case there is ABSOLUTELY NO WARRANTY.
As for the clock, well... I hope to someday make it on a 3D printer. I did say it was a long-term goal. :)
If you want to try the software, go here to download it.
