Advanced Search

Thanks! That does it.

bladekel wrote:

I tried it as you said but nothing changed.
Again receving "USRMOT: ERROR: command timeout" errors and couldnt control the LEDs.

 

    "rpi5": "true",
->
    "rpi5": true,

 

I tried it as you said but nothing changed.
Again receving "USRMOT: ERROR: command timeout" errors and couldnt control the LEDs.

LinuxCNC can only run one toolpath at a time, so to do what you want you might have to get creative.

You could certainly give the two tools a different Z-offset in the tool table, so the two passes could be using the same workspace coordinates. If you have only one physical Z axis, though, that won't really help, as LinuxCNC can only apply one tool offset at a time.

A kludge might be to use X for tool1 and Y for tool2. (unfortunately, no diameter mode in the Y axis)

You would need to split the paths, with a simple XZ move for the start of roughing, then a combined XYZ until Tool1 is finished, then switch to ZY moves.

I don't think you can do arcs this way.

Even with an XZ/UW setup, ie with a separate W axis, I don't think you can do _simultaneous_ XZ and UW arcs. This is mainly a limitation in G-code, as there is no way to define an alternate arc centre.

bladekel wrote:

Hello everyone,
First of all, Meister, I’d like to congratulate you on this amazing work.

I decided to use this project on my machine.
I made some adjustments to use the TangNano9K with an RPi 5 over SPI.
I currently have two LEDs connected to the TangNano9K, and for now I just wanted to start by turning them on and off.

I configured everything through the Setup-GUI, then generated, compiled, and flashed it to the TangNano9K.
I also compiled rio.c using halcompile.
After that, I ran the test GUI from Setup-GUI, selected SPI communication, and was able to control the LEDs successfully from that interface.

However, when I tried to run and test my own configuration rioDeneme with LinuxCNC, I encountered some issues.

When the Axis GUI opened, I first pressed the E-stop button.
After pressing it, I received several consecutive "USRMOT: ERROR: command timeout" errors.
Then I opened HAL Show and tried to change the LED states from there.
Although I could make changes within HAL Show, I couldn’t actually change the state of the LEDs in the real world.

Where do you think the problem might be? 
 

you try to mix software stepgen with rio stuff,
there is a working/generated hal/ini template in Output/cgr/LinuxCNC/

you can start it with :

sudo halcompile --install riocore/files/rio.c
sh Output/cgr/LinuxCNC/start.sh
 

It might be safer to use an established interface to LinuxCNC, possibly modifying it for your needs.

linuxcnc.org/docs/stable/html/man/man1/linuxcncrsh.1.html
linuxcnc.org/docs/stable/html/man/man1/linuxcncsvr.1.html - this is not very informative...

Does it have the same USB ID as the previous one? Possibly you need to change the USB identifier in the config files, or possibly the udev rules.

linuxcnc.org/docs/html/man/man1/xhc-hb04.1.html

This does feel like something that could easily(?) be done with LinuxCNC.
It is basically an ELS (electronic leadscrew) system. There are many options for an ELS, but I don't know if any of those support EtherCAT.

You could start off using a standard LinuxCNC GUI for manual control of spindle speed. Then add a PyVCP box (or a similar GUI element) to set the leadscrew TPI. Later on you might want to look at either making a custom GUI, or going further with the CNC conversion.

You might want to consider fairly early on whether you intend to keep the separate feed-shaft and leadscrew arrangement. Most CNC lathes replace the leadscrew with a ballscrew and use that for both threading and feeding. If you were to retain the leadscrew and the existing half-nut for feeding too, I think you might wear it out. However there were lathes that worked this way (Smart and Brown Model M, for example)

I would suggest keeping the back-gears and possibly the intermediate gearbox, they give a big increase in torque for larger workpieces.
If you do that you would need to add a separate encoder for the spindle to account for losing the fixed ratio between the spindle motor and spindle.

You might find my blog on CNC-converting my Holbrook Minor interesting:
bodgesoc.blogspot.com/2015/08/holbrook1.html

hitchhiker wrote:

Warum brauch ich für die 7c81 ein usb-blaster?
Kann man nicht mit dem w5500 modul zugreifen und per mesaflash flashen per ethernet?
 

weil das ganze mesa zeugs nur mit der mesa gateware läuft, also auch das mesaflash,
du könntest noch rio darüber drauf flashen aber danach nicht mehr.
Flashen über Ethernet/SPI ist in rio nicht eingebaut.

@ bladekel:
i take look later

Hello everyone,
First of all, Meister, I’d like to congratulate you on this amazing work.

I decided to use this project on my machine.
I made some adjustments to use the TangNano9K with an RPi 5 over SPI.
I currently have two LEDs connected to the TangNano9K, and for now I just wanted to start by turning them on and off.

I configured everything through the Setup-GUI, then generated, compiled, and flashed it to the TangNano9K.
I also compiled rio.c using halcompile.
After that, I ran the test GUI from Setup-GUI, selected SPI communication, and was able to control the LEDs successfully from that interface.

However, when I tried to run and test my own configuration rioDeneme with LinuxCNC, I encountered some issues.

When the Axis GUI opened, I first pressed the E-stop button.
After pressing it, I received several consecutive "USRMOT: ERROR: command timeout" errors.
Then I opened HAL Show and tried to change the LED states from there.
Although I could make changes within HAL Show, I couldn’t actually change the state of the LEDs in the real world.

Where do you think the problem might be? 

Warum brauch ich für die 7c81 ein usb-blaster?
Kann man nicht mit dem w5500 modul zugreifen und per mesaflash flashen per ethernet?

I just finished a massive rework of the enable/disable logic for controls!

JT

The G code editor highlight line is updated by the current reported motion line from LinuxCNC. This is sometimes not the exact line that is being executed because of look ahead.



JT

generell finde ich es pratischer die konfiguration am pc zu machen (VM mit Linux geht auch), dann könnte man auch den TangNano9k programmieren.

Für das ECP5 board brauch man den programmer (usb-blaster, der funktioniert auch am PI und die toolchan auch (arm64) )

Zum betrieb am RPI4 kann man SPI (Jumperwire kabel) nutzen oder en W5500 Ethernet-Adapter (~4-5Euro).

Dann musst du aber noch schauen was du ansteuern möchtest, also z.B. Stepper,
da wird es mit den dev-boards etwas umständlich.

Wie soll denn dein test-aufbau aussehen ?

Oder wenn du eine 7c81hast kannst du die auch nutzen, brauchst aber wie einen PC/VM + USB-Blaster zum programmieren.

Ok!

Aber was könnte man direkt nutzen ohne zusätzliche hardware? Direkt am rpi4?

Danke