Advanced Search

The serial steppers are nice because they are easy to setup and they are also very quiet. I dont know if there is an fpga firmware for it thought. The SPI ones, someone has done a custom firmware to work with those.

What is W13 doing? Is it just for swapping between the usb or the green connector?
Consider adding a jumper that can disconnect power to the FPGA or separate the 5v rail between FPGA and RPi5. I found it convenient having a separate jumper for that, incase you needed to either separate the power between the 2, or if you need to reset the MCU/FPGA, you dont need to turn off the RPi

When you are done will you post the design on that easyeda website so I can clone it to modify it? Im also hesitant about a 1 board thing, but I wanted to make something with a swappable MCU/FPGA , and the formfactor looks pretty decent sized.

Can you reduce the encoder scale? If not, probably best long term solution
would be for  the encoder driver to use 64 bit intermediate calculations.
I can take a look at this.

nice! please post up when its done, that is an interesting machine, look robust for a bench mill/

Thank you guys for the interest and the suggestions! I will try to respond to a few of them here:

"I have occasionally thought that there should be a niche for a similar board, but with on-board stepper drivers (Trinamic or some other serially-controlled driver)
That would, of course, need a dedicated HAL driver, but those can be surprisingly straightforward."

I was actually planing on designing a seperate IO Board featuring the Trinamic 5160 Drivers, as they would allow me to run high current stepper motors. They come with UART and SPI interface functionality but I haven't really seen an advantage in interfacing them serially. Maybe someone has a good argument as to why it would be worth it to go through the integration troubles?

"I like having the boards seperate, just in case my butter fingers cause one the voltage translation chips to blow a smoke valve and damage the PCB."

That was my thought process as well. I'm a little hesitant to put everything on one big PCB, as i feel like that would make errors more catastrophic. I also like the modularity of having seperate PCBs, especially in these early prototyping stages. I did actually bring the JTAG out, but realize that better labeling was neccesary, which I've since added.

"I am liking the look of this board. Did you do this in easyeda?
Suggestion, add the rpi 40pin header. Aside from being able to add stuff rpi side, I find it valuable for debugging the MCU/FPGA side of the board using a regular rpi without having to keep removing a cm module and you can use it as a backup incase you damage the 100pin hirose connectors or theres another RPi/chip shortage"

Yes, I did do this in EasyEDA. I started the project with EasyEDA becuase I foud the integration with the JLCPCB services appealing.
Including the 40pin Header is a great idea, which I will definitly add. Unfortunatly that will require major rerouting. Before rerouting everything I will continue collecting and thinking about additions/optimizations. That way I'll hopefully save money on design itereations.


I'd love to here if you guys have any other sugestions. The design philosophy behind this board is to make it the Compute Module 5 version of the Mesa Electronics 7c81, and I will probably finish this project making exactly that but I'm also open to brainstorming other ideas for potential future projects.

Thank you everybody!

Die Siemens und Heidenhain Steuerungen basieren auf Werkstatt orientierter Programmierung (WOP), im Gegensatz dazu die PAL-Programmierung.
LinuxCNC basiert auch auf PAL und hat sehr viel gemein mit Fanuc, aber eben nicht alles. Heidenhain oder Siemens Code kann man überhaupt nicht unter LinuxCNC laufen lassen.

"Pal ist die Sprache die in de rAusbildung gelehrt wird und die auch die Basis für die Sinumeric ist."
Ja, die ganz alten Siemens Steuerungen, aber die neueren (auch schon relativ lange) nicht.

I'm also designed a mesa clone for a Raspberry Pi, currently in the process of testing everything.
Except for the reversed Sserial Pinout it looks like I got everything right.
You can take a look at my schematic for reference: github.com/jstjst/stmbl-fpga-master/blob...tmbl-fpga-master.pdf
Be careful the SPI In Out labels in the 7c81 schematic are reversed compared to the ucf file. Luckily I could fix that in the ucf file.
I routed some signals under the fpga, I see no problem with it. But try to have no completely hidden tracks so you can fix them if you got some connections wrong. In my opinion one of the more important things is the placement of the decoupling caps.

There are some pitfalls with the whole (fpga) programming (especially when using the Ethernet host interface). But with some reading through the forum I got everything pieced together, including recompiling mesaflash and linuxcnc. So it is totally doable (for reference, this is my first deep dive into linuxcnc and mesa/hostmot2, only used the RPi GPIO interface before, didn't use/own a mesa card).

Big thanks to Peter for providing the schematics and all the information that is on the forum!

Have you had any time to give a Dell OptiPlex 7050 a rub yet?
It would nice to get some feedback and recomendations regarding eventual tweaking :0)
Kind regards/Stefan

Can you reproduce this in one of the included simulation configs?
If yes then please provide a minimal gcode example (ie one that can actually be run when loaded) along with a clear description of when you are stopping the program and tell us what linuxcnc version you are using.
If you cannot reproduce this in a sim config then please also attach your zipped config folder.

As it is I cannot reproduce this on either 2.9 nor on master.

91 turns is possible very much so ,

this is just a simple small part so a complex job would very much be way over that turn number.

wrapping sounds right as number does a odd sign change in the feedback counts

Unfortunately not yet, but am looking to get this cataloged for others as I have completed quite a few 3d printed parts to help support the PM728VT machine to CNC.

As such this can help others move across more quickly.

I want to make sure there is a positive ending before releasing the build program, just to make sure others get to a realistic endpoint.

Currently using a laptop, and seems to be Ok. But may need to get a miniPC to run it all so just incrementally addressing each issue.

For the cross slide I oredered a 12mm ballscrew. If I can find room I will mount the ball nut under the cross slide table. If not, the screw I bought is long enough that I could mount the ballscrew on the back side of the cross slide table.

I'll be going out of town for a couple weeks, but when I get back I think I will do some examining for the encoder situation.

I saw a video of somone doing threading on linuxCNC using a missing tooth scheme. I may look into this more when I bored in my hotel room next week.

Thanks again for the comments
 

Leave them at default 5000 till you have everything working properly, then you can decide if you need to mess with timings. Usually there is no need.

www.aliexpress.us/item/3256805790238823.html

I am waiting for the manual from the seller, will add when i receive it.

 

No., unless you need a new component to provide the pins.
(which is not the case here)

If you are learning, I would definitely poke around a bit with halshow
(in axis it's: Machine --> Show Hal Configuration)

 

Greg,

Right now, you have a machine in good mechanical state. The only problem you have with this machine is the controller. My advice, do not remove or dismantle any component at this stage. 

1-Gather the manuals wiring diagrams and any information you can find. 

2-Study the servo drive CONTROL METHOD (analog, digital, industrial communication bus etc …) 

3- Study existing wiring diagrams and determine the number and type of IO’s (sourcing/sinking) 

4- study the type of controller you will need, the type of daughter boards needed and type of PC (pci/pcie/ethernet/Mitsubishi bus communication bus … etc 

5- Modify the existing wiring diagrams to reflect the changes required. Keep in mind that the machine was designed and assembled by highly skilled craftsmen, at this stage you will focus on keeping the modifications to existing components to a strict minimum. 

6- With all the above knowledge and mods you will be ready to configure the controller and choose GUI 


Note: machine integration requires patience especially if you have not done it before on a machine such as vmc. (Bounding, grounding, shielding can be nightmare, modifying the safety control system can become a liability and changing from sourcing/sinking io can be laborious, etc …), keep it simple stupid (KISS) is the answer to a successful project.