Communication solution on my own CNC Machine

Great writeup. Lots of enthusiasm and anticipation, good start!

To my knowledge anything can be done in the automation sphere, including making your own proprietary motion controller(s) & HMI. Many have done it for their robots and other machine-tools. It's no secret to anyone that much of the modules required to get you started are already-done by those gate array manufacturers and serviced to their developers via "IP cores". (stepgens, pulse waves, encode/decode, communication protocols, filters, mipi/dsi, sram, dma, everything's there). Have a look at the ip cores and hdl example sources from Lattice's Radiant, AMD's Vivado, Intel's Quartus, Gowinsemi Designer, etc etc... option list is endless if you add in the second tier manufacturers of those semiconductors...

And so can a mesa-like solution be achieved? of course it can. Question is, though; is your goal to be productive, as in making parts? or be a night-time developer who waste time reinventing what already works? I.e MESA, Litexcnc, RIO, Remora... to name a few

(Side note) I've always been under the assumption that those raspi were engineered to play video games on a 7" lcd and for making youtube unboxing vids (aka marketing toy). Not particularly aimed at the automation signal gen & processing market, at least not at the data integrity level one you require and deserve. That being said, you might want to re-think that strategy!

happy-happy chip makin'

Nah the Raspberry Pi can be used, quite successfully. Pretty sure that Ol Gene has been using a RPi3 successfully for years on his lathe, think he uses a 7i90 via SPI.
The 5 is a little rocket, shame it’s lacking an SPI driver.
As for the OP proposal, I’m skeptical. I also think he’s a little naive on what’s involved in “rolling you own” motion controller from scratch.

Did I get it right he’s also going to build his own servo motors as well ? That sounds like a serious engineering challenge, if he gets that done I’ll give him a right decent tip of my hat.

Sure. Dunno, can't chime on their potential I've never had one never will. Must be their repulsive branding and silly blackberry-success-wannabe stagnating in the back of my mind that makes them underperform lol Still a yt toy...

True. Correctly tapping into the RT of Lcnc ain't exactly easy at the first place, OP to find out the hard way soon enough. Much (muchhh) easier to go proprietary (ground up) with all this... if it would be me. Just who got the time for that though, what's the point/advantage.

Started messing with a Spartan dev board cos I had one on hand that was in the useful box. Plus I needed something to keep the black dog at bay.
Plus shipping from the US to Down Under almost doubles the price, plus we have a government who’s shutting down coal fired power stations hoping that wind and solar will take up the slack, thereby pushing up power prices and it flows onto to everywhere else, add to that a new excuse why petrol and diesel have gone thru the roof every 2 seconds.
If I could afford the 350k drive down to the capital I’d start a coup. Hahahaha

At least we haven’t been invaded, the snakes kinda protect us from that.

Spartans are killer, just not cheap anymore though (ironic). First functional project I made was on a spartan, a ov5640 door bell hack that was loool Got into the lattice during their debut, got whacked out from their free license, since moved to the M3 hard core + FPGA fabric "all-in-one-silicon" (GW1N, from Gowin). Check'em out, some seriouzpwr for a few bucks (search keyword "minimal schematic" in their doc center for putting up a dev board in 2hrs).

Leave DC alone man, they're lost enough chances they'll coup their own selves is more likely than not...

No fancy 15million gigabytes bs vivado dwloads required. Cortex core part can be written in eclipse/keil/whatever (for those of you familiar with MCU c/cpp), and the hdl's carried using their designer. Easiest, cleanest, fastest and most efficient free licensed IDE I know of. /A f f o r d a b l e, and it just freaking works. Raspi... [rolleyes], come on now @OP...

Ain’t no yank man.

Down Under = Australia.

No need to move away from the Spartan 6, my project is done.

I am aware that the io management I mentioned is not real time. Does anyone have any information about the size of this job? My goal is not to convert a huge vertical machining center, but to make my own hobby machine by sacrificing some real-time features. I'm wondering if it's possible to do this with ordinary non-realtime TCP packets. I think there must be a solution between simple grbl builds and "superior" mesa builds.

Two computers very close together can exchange data between them very quickly, although not in real time. What are the challenges of driving axes with this data? I want to discuss this.

I wonder how possible this could be. My conclusion from this topic is that I can only approach the idea of using this solution as a simple handheld terminal or keypad.

Once again, I want to know "to what extent" this is possible. Will there be a huge millisecond delay or momentary drops? Will a communication not in RT mode suddenly crash? Or could using a hobby machine at low speed and acceleration with a few milliseconds of delay be "instantaneous enough"?

Most importantly, can anyone explain how to send these pulse signals as ethernet packets with Linuxcnc? Please don't worry about the real time. Maybe I'm making a turtle feeding device, you never know slow but how slow? I want to know this.

You’re building a fixed gantry router to process aluminium of about 600mm x 400mm.

So no you’re not building a turtle feeding machine.

I think your best solution is to ignore all the advice you’ve been politely given and go and do it your own way. It may work it may not.