This will be a short update, as the next one will be... good.

Started over with my all metal design.

Changed a few angles and dimensions. The issues I had with v8, the plugging and jamming during cold starts, are gone and I was able to keep all the good stuff. All metal, less parts, easier to produce, very happy.

Correcting the rounded corners issue, which was a tool accuracy gcode setting, has revealed another issue. The test cube I've been printing now comes out with the top evenly shifted towards the back/right. I believe this means I'm loosing steps on my B axis during fast upward motion. When I go to print the next part, the head also slams into the bed if I don't re-home. I'll be tuning my drivers and tweaking my acceleration and max velocity settings today. I've never done this correctly and I've been living in the world of 'good enough'.

In other news, I got to try actual CYMKW printing after a trip to Coex resulted in 5 spools of suitable filament for testing. Got a pretty good looking full color spectrum, but the saturation level (and balance) of the filaments requires some significant tweaking. Getting a pure blue for example is off by ~50% from where it should be, and ends up looking dark and muted.

So, I drove back there yesterday and tried RYBKW on-site. The color saturation was much better, but so was the color saturation of the source filaments. Obviously I could produce a pure blue now.

I'm thinking it comes down to mainly the saturation and color balance of the source filaments. As I just grabbed the filaments they already had and it really wasn't designed for this, nothing I'm using is a real 'pure' color. My magenta is too white/red, my cyan is actually aqua and seems to have black in it. My yellow also contains a lot of white. My red and blue both contain black.

All these colors are designed to be visually appealing on their own. They are not designed to be accurate or balanced.

Ideally, the source filaments should be fully and evenly saturated, and not include black or white in the mix to the extent possible. This would probably not make for good looking filament on it's own, and would require a significant investment on the part of my filament supplier to correct. I'm going to solve this to the extent possible in software first, so I can have some quantifiable results. There would also need to be a market for it. Working on that..