Hi Andrew, welcome, that's a pretty sweet looking plywood machine you've got there! CNC milled? I'm jealous.

You don't mention which motor driver you're using, so I'm going to assume it's an Adafruit Motorshield v1, and that you've loaded the hex firmware from polargraph_server_a1_adafruit_v1.cpp.hex.

Easy answer to your flipping question:

AccelStepper motorA(forwarda, backwarda);
AccelStepper motorB(forwardb, backwardb);

(https://github.com/euphy/polargraph_server_a1/blob/master/configuration.ino#L53-L54)

Swap the parameters of each AccelStepper constructor:

AccelStepper motorA(backwarda, forwarda);
AccelStepper motorB(backwardb, forwardb);

- The stepsPerRev should be 2x your motor step, so try 400 instead of 800 there.
- You've got your machine width set a little wide compared to your diagram (653mm on the diagram, 658mm in the machine).
- Your mmPerRev is at least twice what it should be. This value should be the exact amount of cord that is wound off the sprocket with one revolution, with a diameter of 33, then the mmPerRev couldn't be more than 100. The exact dimension of the sprocket isn't actually that useful to measure this - better to just count the beads, and then measure a stretched piece of beaded cord with that many beads on it. I get about 93mm on mine.

When you modify these settings, you're better to close the controller, and change the default.properties.txt file directly, then reopen the controller.

- Your home point is 262mm across, so is not in the centre. It doesn't have to be, but if there is a mismatch between what you tell the machine (when you do a "upload machine spec"), and where you physically place the pen when you do a home, it'll complain. Maybe you've already got this, because I don't see a horizontal distortion.

sn

Remember you can select your area and then use "draw outline" to get an idea of the outline, without having to wait through a full pixellated draw.

Actually looks like it's drawing pretty straight and true now. One minor flub in the right eye when, in explaining how it worked, I gesticulated into the pen.

One thing that is odd... it gets quiet fuzzy at the end. Do you all think this is just the pen working itself loose in the tube or the fact that the strings are longer and more wobbly? Any ideas on how I might fix it?

I have to think it's the pen working loose. I'm using a .5in brass tube with 3 roller bearings around it (the same setup as in the kits I think). I just wrapped the pen in electrical tape and stuffed it in there. Thinking about getting a small section of plastic pipe to put in and fit a thumb screw into. How have you found is best to lock down the pen?

Hey Sandy! Just got a chance to try your fix for the mirroring... it seems that while it does fix the left right mirroring problem it creates an up down mirroring problem.

Any ideas?

Switching the ports did the trick! Excited to run a full page drawing in the right orientation tomorrow.

Side note: I 3D printed a pen holder based on the one you linked to. It's designed to press fit into the shaft and be clipped to lock.

Will report tomorrow with results.

Side note: I'm going to be doing a lot of documentation of my particular variation of the machine (CNC milled frame, laser cut arduino mount, 3D printed sprockets, 3D printed ball chain/weight connectors, 3D printed gondola arms (with integrated seats for ball transfers, roller bearings and ball chain), etc...) Is there a collective hub where people post this work?

Thanks Sandy!

Binder clip on the tabs indeed. Yea layer lines would be tricky with this one. I printed it in the orientation shown on a dimension 760 abs printer and it came out fine. The layers running perpendicular to the binder clip seems to be most important. I'm also running one off on a friend's FormOne as those machines don't seem to produce layer lines and the parts tend to be strong but flexible.