Posted on

Serious Deep Plotting

I’m always amazed that I haven’t grown tired of this hobby yet, year after year I find new stuff I want to try or improve. There has got to be an end to the madness, right? I rationalize this by saying I’m learning very general implementations of stepper motor controls, and that is true. Or that the kids’ minds get sharpened from the interactions with such machines, and that is true as well. But deep down I’m aware that there is a large obsessive component that is much harder to justify, say to one’s life partner.

And so I have the gondola plotter deployed at home for the Christmas break. It’s over Nicole’s half of the craft room desk, but there were no clauses about air space so I’m in the green.

To push my luck with Nicole I have it draw 30+ hours plots, and so it goes through the night and we can hear it from where we sleep. But it’s soothing like a dishwasher, it’s the sound of work being done for you. I’ve had to be extra careful about power, it’s not rare in the Winter that we lose the inverter in the early morning before the Sun comes out again. This is a non issue with lights and internet on the always-on 12V circuit, but the plotter can’t pick up where it left off currently so I had to be careful to have enough power to go through the night uninterrupted.

Over time, I’m gaining an understanding of what algorithms will work well on what subjects. Vpype flow imager works really well for portraits & anime. For some reason, trees in particular come out really well with it.

An homage to Miyazaki: Totoro

An homage to Miyazaki: Princess Mononoke

In any case, with the Christmas break around the corner, I started getting excited about all the things I wanted to do in the plotting world. Besides just plotting, I’ve been wanting to improve the acceleration model to make it dependent on the angle being taken by the pen. And this was a very deep rabbit hole which got me to rethink everything about how a sequence of pen strokes is being drawn. After many failed attempts I brought some amelioration to the acceleration model but I wouldn’t say it’s perfect.

Another thing I wanted to tackle was to correct the odd way in which gondola plotters (which are polar based) move over cartesian planes (not in straight lines). Over a long travel distance, this fact is painfully obvious:

moving from (20,80) to (80,80) with an actual straight line between these points for reference

Yet for the things I tend to draw, there are enough intermediary points that it’s rarely an issue. And so the fix is obvious: when traveling over a certain distance, break down the travel to go through intermediary points on the same line. Of course this interacted with the acceleration improvements and it was an overall nightmare to get right… Oh well. It’s done! There will be a point at which my plotters can be improved no further.

This plot was supposed to be simple, but it ended up being the final boss. It made several issues percolate up, and I tested the new algorithms with it, after several failed attempts, it was finally complete:

It’s really absurd how accurate this gondola plotter is given how it’s driven. I’m putting a lot of miles on it and bring about many improvements along the way.

Although plotting on a much larger surface highlights some imprecisions of the algorithms I’ve been using.

What is this? A Mechanical issue?

Oh…

Well, that’s really just 3 plots, but very long ones with many failures and improvements in between. I was hoping to document the Gondola PlottyBot build over break, but this will have to wait a bit more. I’m definitely close.

Oh, and before break I had it do a bunch of snowflakes at work. They are very satisfying to watch get drawn.

Unfortunately for everyone around me who is sick of hearing about plotters, there are still many things I want to do with them.