I have been thinking about what killer features Chilipeppr has * Open-platform pubsub provides

I have been thinking about what killer features Chilipeppr has

  • Open-platform pubsub provides fantastic extensibility. This means whatever other close platforms have won’t be too hard to implement.
  • New visualization platform and JS front end mean faster, cleaner and decoupled support for machinery. You seriously do not need those crappy computers with expensive software attached to run things.

So where should we take it from here:

  • GRBL/TinyG only support a subset of G-code. This is a good design. Many G-code support is for things like routines, variables, … which are all much easier done in higher languages. Insisting on doing it at the G-code level sounds to me like insisting on doing things in assembly to gain the minuscule performance improvement. Completely unnecessary.
  • JSCut is filling in on the CAM part of the toolchain. We should put more effort into this matter to make the toolchain from CAD models to finished object as smooth as possible.
  • Add Computer Vision and Machine Learning to take our machinery to the next level of smartness. Things like Homing, Work Zeroing, Work Inspections, … can already be achieved with computer vision. And with WebRTC, we should definitely explore this matter.

Overall, I am excited about Chilipeppr. These badly designed machine packages have been around long enough.

@Dat_Chu hey i would like to get you a tinyg can you email me your address? We all on the tinyg team have been working very closely with @jlauer and are impressed by your work. My email is just my username @http://gmail.com. I am doing orders now in fact. If you can email me quick I can get you one in the mail asap.

@Dat_Chu I couldn’t agree more. The pubsub was a core design principle to create a loosely coupled system. It’s paying massive dividends as people are starting to extend. It was also a core principle to have the UI include the code documentation. Sounds like that working well for you.

Some things hopefully we’ll see in the future from the community:

  1. Pick and place via WebRTC integration
  2. Advanced soldering techniques using CNC
  3. More laser control
  4. More Arduino extensions (after all the Serial Port JSON Server was designed to enable multiple serial ports from ChiliPeppr)
  5. Editing of Gcode on-the-fly using the 3D Viewer

The list goes on…

One thing that I know Mach3 can do is to control an A axis that allows a tool that has direction (such as a blade with one sharp edge) to point toward the move direction. I believe this is called tangential cut. How would we perform this task with our current PubSub architecture? I assume this is similar to the auto-level feature of TinyG with G-code being updated on the fly.

The guys at other mill did this with TinyG. It was for cardboard cutting. @Mike_Estee was the primary Dev on that project. He’s at @Other_Machine_Co now.

I would add a widget that modifies the Gcode by scanning through it. You can get access to the Gcode and iterate it by looking at the “Iterate Your Gcode” macro.

Then, look at the Start and End position of the Gcode line. Each line while you iterate has a p2 which is the XYZ point of that line. Look at the previous line’s p2 which would be the Start point. Then just triangulate using Pythagorean Theorem. The angle you get you set to the A value for that line. Or perhaps inject a line prior to each to make sure the angle is set on the blade prior to the execution of the next line.

I would just modify the Gcode the way the Auto-Leveller does and resend it to the workspace. You should resend because the 3D Viewer and Gcode Widget both have independent copies of the Gcode. Look at the macro “Load Gcode Into Workspace” for how to do this.

Honestly, you could have this whipped up pretty fast.

That sounds awesome John. I think I will add a sticky box (similar to checkbox) that one can choose within a new widget with 3 buttons: A,B,C. This allows the user to choose which extra axis will control the blade direction.