Remora for RP2040

I'm keen for this or something like this, just for GPIO even.

(E.g. for complementing an ethercat setup).

Watching with interest.

With step/dir even low step rates for additional things like auto tool changers would be quite helpful.

So, I've implemented a really super simple semaphore between the threads and the Ethernet data transfer, looks to have solved the data corruption. My hunch about the virtual machine causing the following errors was correct. The servo thread was just to slow for the communication and updates. The video is showing the hotel room test bed running with a 5000ms servo thread.

This is so close to be running now!

pippin88 wrote:

I'm keen for this or something like this, just for GPIO even.

(E.g. for complementing an ethercat setup).

Watching with interest.

With step/dir even low step rates for additional things like auto tool changers would be quite helpful.

The current 40khz step rate is not too bad. I think we should be able to get it up to about 100kHz. 

With the config file all of the pins could be configures as IO.

Below is the step and direction output, just to show it's actually working...

 

I'm not familiar with Remora - will it be possible to use multiple of these?

pippin88 wrote:

I'm not familiar with Remora - will it be possible to u,se multiple of these?

A good question and something that needs to be considered given the limited IO of a Pico. Currently the Remora Ethernet component has a fixed IP address it is looking to communicate to. If the LinuxCNC component is modified to accept a configurable IP address it would be possible to load multiple instances, in LinuxCNC, talking to different boards.

The controller board will also need to have a configurable IP address, a feature that has been requested for the NVEM, EC300 and EC500. This could be included in the uploaded config file (when I get it working), and a method to force a default config on boot to "reset" the board just in case we need to.

The 2040 is an interesting chip.
I hope it grows in popularity.
Do you think with the cores and the 4 state machines this could be faster than the current 3d printer variants?

Aaroncnc wrote:

The 2040 is an interesting chip.
I hope it grows in popularity.
Do you think with the cores and the 4 state machines this could be faster than the current 3d printer variants?

Hi Aaron, 

The port for the RP2040 is not utilising the PIO, so we will be limited like the 3d Printer boards.

Would it be possible to use one or is the way the code setup make it not possible?
Might be useful for high speed encoders.

The PIO on the pico will need a learning curve for a DDS style step gen but I would think a high speed encoder interface could nearly be off the shelf...

github.com/jamon/pi-pico-pio-quadrature-encoder