Cnc3020 1 day upgrade to open source G-Code

[Pauldg]

My friend Dan and I did a one day upgrade of the cheap Chinese eBay cnc3020 machine. We made a YouTube movie of that day and compressed it into 10 mins, enjoy!

 

[XD351]

The super Gerbil unit looks interesting !

 

[Pauldg]

The super Gerbil unit looks interesting !

Thanks! The idea is that many cnc machinists and hobbyist will dive into Grbl and develop it further on this new 32 bits platform. Examples are like taping and canned cycles... At this moment the controller parses the G-Code while streaming and executes it. Canned cycles will require a pre-scan of the G-code, store it if under 20kb static ram (or use some of the flash memory...still 50k available). Once checked and ok'd, it should executed the cycli. My biggest issue is that I don't have enough experience with e.g. helix cutting and threading to ensure the G-code parsing and execution are done correctly. So looking for collaboration since it's open source! Anyone is welcome.

 

[XD351]

Sounds like you need someone with a lot more experience than i ! I will keep following along learning as i go .

 

[Wizard69]

Nice work guys! You appear to be having lots of fun

Thanks! The idea is that many cnc machinists and hobbyist will dive into Grbl and develop it further on this new 32 bits platform.

There are a number of projects putting or trying to put GRBL on 32 bit ARM. Development seems to have slowed up a lot. I believe that this is due in part to one of the main developers trying to rewrite major portions of GRBL for a bright future in the 32 bit world.

Examples are like taping and canned cycles... At this moment the controller parses the G-Code while streaming and executes it. Canned cycles will require a pre-scan of the G-code, store it if under 20kb static ram (or use some of the flash memory...still 50k available).

I’m trying to understand why canned cycles would require a code pre-scan.

Once checked and ok'd, it should executed the cycli. My biggest issue is that I don't have enough experience with e.g. helix cutting and threading to ensure the G-code parsing and execution are done correctly.

Would it not be better to start with tiny steps? Beyond that is a router really the ideal machine for threading. I suppose you could make threads in wood. Drill and peck for example would be a good enhancement and likely usable on a router.

So looking for collaboration since it's open source! Anyone is welcome.

No machine at the moment.

Interestingly that little Chinese machine looks to be a lot better than many I’ve seen. Im sure many would be interested in the suppliers you used for this build.

As for super GRBL, I May have missed important info here, but is this a product you intend to sell?

The one thing GRBL suffers from is a proper board to run on that supports automation well. If you are trying to address that I could offer up a few suggestions. Right off the bat one suggestion would be 24VDC capable I/O. 8 bits IN and OUT would be a good start but 16 INs would be even better. Note these are general purpose and not the enables, E-astops and other required inputs. An ultimate board here would have an I/O coprocessor (possibly a PLC like environment) to execute user code outside of the G-Code pathways.

My first magination can wander significantly but these are basic ideas. I think I need to rerun you video when I’m more fully awake.

 

[Pauldg]

I’m trying to understand why canned cycles would require a code pre-scan.

If you stream G-code, the canned cycles are sub routines that have not been streamed yet!

As for super GRBL, I May have missed important info here, but is this a product you intend to sell?

Yes, I sell them on Awesome.tech

The one thing GRBL suffers from is a proper board to run on that supports automation well.

Mighty Gerbil is in the make (additional IO based on a STM32F105RC6)