Advanced Search

Do a new full install and start over.
Never use root or SU to install stuff as everything will be installed in the root folder and nothing will work as a normal user, always use sudo.
LinuxCNC will refuse to run as root due to having full hardware control and stopping misuse.

MIN_FERROR = 0.010 - This is the value in machine units by which the joint is permitted to deviate from commanded position at very low speeds. If MIN_FERROR is smaller than FERROR, the two produce a ramp of error trip points. You could think of this as a graph where one dimension is speed and the other is permitted following error. As speed increases the amount of following error also increases toward the FERROR value.

FERROR = 1.0 - FERROR is the maximum allowable following error, in machine units. If the difference between commanded and sensed position exceeds this amount, the controller disables servo calculations, sets all the outputs to 0.0, and disables the amplifiers. If MIN_FERROR is present in the INI file, velocity-proportional following errors are used. Here, the maximum allowable following error is proportional to the speed, with FERROR applying to the rapid rate set by

If you are on the factory settings of the PID regulator. They are almost always very soft. They must be set up.
When the driver is activated, the motor axis must be locked - by hand, the axis must not move.

the rmii code is not mine, unfortunately my knowledge of verilog is not sufficient for that, i have only adapted it to rio

github.com/sipeed/TangPrimer-20K-example...Ethernet/verilog_UDP

you can run the rio-test tool to see the timeouts:
all in all, it works quite well with the w5500, but with the esp32 I get no timeouts at all

EDIT: the ping is completely handled inside the W5500, so it's not a good test

Comparison added for the sake of having something to compare too. Nothing scientific ofc, however goes to show. Here's me brutalizing the gateway the same way, yet here we are talking, lket alone streaming 4k vids, dwloading gigabytes of data from int'l servers. 

I'm not an network expert but, this W5500-over-RIO seems to be bullet proof!! Fairly impressed already. 

I can now say what condition I am in

Save this as 'feed_zero.py', mark as executable and save to a folder called 'python' inside your machine config:



Then add this line to the [HAL] section of your ini file:
This should then zero the feedrate and rapid-feedrate when the machine is first enabled and also whenever a new gcode file is loaded.
Note this will not zero the feedrates if the same file is loaded/reloaded.

meister wrote:

i am not yet satisfied with the W5500, better not to plan for it, there are packet losses and i have no idea why, it can't be because of the cables as i have test setups without them.

I also have a working RMII interface, but so far I have only been able to get it to run on a TangPrimer20K (with gowin toolchain) and even there I still have a few problems.

Are you sure about that w5500 pkts losses issue? Perhaps my silly jumper leads are less inductive/destructive than anticipated, I don't know. To me the losses are within 'functional specs' when a brutal ping beat-up is applied (see visual). Not 0% perfect ofc, but then no interfaces are perfect under that stress, to my limited knowledge mind you. 

The RMII reference design is an excellent one indeed. Not easy to implement though, you have my respect. 


 

How relevant is this option at the moment?

Not relevant, I was just wondering if I could run the output on a simulator.

thank you, you make me embarrassed

i am not yet satisfied with the W5500, better not to plan for it, there are packet losses and i have no idea why, it can't be because of the cables as i have test setups without them.

At the moment a UDP2SPI bridge (esp32) is still the best solution.

In principle, you could also programme the FPGA via an esp32 (JTAG or SPI flash), but you would have to adapt the firmware first.

I also have a working RMII interface, but so far I have only been able to get it to run on a TangPrimer20K (with gowin toolchain) and even there I still have a few problems.

apt purge linuxcnc*

Reading package lists... Done
Building dependency tree... Done
Reading status information... Done
The package 'linuxcnc' is not installed, so it will not be removed
0 packages upgraded, 0 newly installed, 0 to be removed, 0 not upgraded.

the second option worked but the first one didn't


1To install from a fresh Debian 12 with NO LCNC installed yet, just use: sudo dpkg -i (package name here without brackets) for each deb and reboot. Choose the RTAI kernel in Grub, enjoy.

2To install on Debian 12 with linuxcnc-uspace already installed use : sudo apt purge linuxcnc* and wait for apt to completely remove every trace of LCNC.
Then follow the instructions for a fresh Debian 12 install above.

@MennilTossFlykune do you know if it's possible to zoom to the machine extents (the red box) with a pushbutton?

Note: Prolly not going to work, or be super-reliable with those inductive antena jumper leads lol. Good enough for testing anyway. If this all works out I'll amend the dev board and swing on a wiznet w5500 part with the magnetics, route cross-talk outta of it, add the highspeed opto and level shifters for ins & outs. Guessing the SPI is plenty sufficient for comm and so no need for the realteck gigabit, methink.

I'll keep you updated. Thanks again, meister.

So, "slave" has a fairly specific meaning in EtherCAT -- it's basically just an EtherCAT device that isn't the master device that's in charge of timing for the bus.

In this specific case, it looks like it's just a 3-axis CiA 402 device. The first axis uses objects in the 0x60XX address range, the second is at 0x68XX, then 0x70XX.

You can either take a single-axis CiA 402 XML file and duplicate things, or you can give the new-ish basic_cia402 driver a spin. Out of curiosity, what happens if you run lcec_configgen on the box? It should spit out a mostly-working config file and you can cut-and-paste into LinuxCNC's config, but I don't think I've ever seen a 3-axis CiA 402 device before. I have 1, 2, and 6 axis devices here, though, so 3 should probably work .

I am with you on the "i have hardware" so use that, see how it works for you.
I was in that position and the amount of stuff i learned is astonishing so when i got my first Mesa boards (5i25+7i77+7i74+7i70+7i71+7i71) took me half a day to have everything wired and working, on a 5 ton machine.
You can always order Mesa gradually, get the 7i96S as it is very versatile and has expansions for everything, or get the 7i92TM even cheaper, then add 7i83 for analog and 7i85 for encoder feedback, add 7i84 for 16 outputs and 32 inputs, 7i85S for step/dir, etc, etc there is a huge number of combinations and expansions options.
As for used ones, i got a 7i96 back during the pandemics, in total cost me 330 Euro here, the next day i got an offer to sell it for 1000 Euro, i did not sell it despite not needing it right then.
Luckily that time is over.