Advanced Search

I was able to make a basic gui to turn on and off 2 leds using cursor. Pretty cool tbh. It took 2 days of pain in the ass vibe coding to get it done but now that I'm here everything should be smooth sailing. This is really a new method to make gui that is not talked about much. In my opinion it's much easier then trying to learn a few programming languages. Pretty neat because cursor is pretty strong at using the ai to make python gui. I don't think I would use it for a mill but for basic input outputs it should be strong enough.

Firehunter wrote:

LinuxCNC version on the splash screen is 2.9.0~pre1+git20230208.f1270d6ed7

I really thought we had a newer version than that on Bookworm, but the package lists suggest otherwise. 

I don't think that anything has changed with M66 since that version (which was a prerelease that found it's way into the Debian archives during their testing phase, expected to be replaced, and then wasn't. 

I don't expect it to make a difference to this particular issue but there are other fixed bugs that I would be happier knowing you are not going to see. The current release version (outside Debian) is 2.9.4.  You can just download the .deb file and install it from our server: 
www.linuxcnc.org/dists/
What Linux are you running? (eg, Debian Bookworm (my first guess)) and on what platform? (PC? Pi? Oher?) we can then point you to the exact package file you need. 

Please tell us the advantages of gcode++

ihavenofish wrote:

This is still not the build with the jerk limiting right?
 

Does PathPilot have jerk limiting? Or are you talking about the Grotius version? 

automata wrote:

I am not really looking for the pathpilot UI.
The Tormach version has look ahead applied on all 9 axes (as opposed to only XYZ for vanilla linuxcnc v2.9).
Also I wanted to see if they have applied the dpll patch needed to operate Ethercat drives with lower jitter (github.com/sittner/linuxcnc-ethercat/blo...-functions-2.8.patch). My conclusion on this front is that they have not.
Further they sent over non-compilable code - there is a bunch of uncommented FIXMEs and TODOs littered over the code and it does not even compile.
So Tormach is definitely not playing by the rules of GPLv2.
And it uses python2 (and python3 in some places) to the point that it is not compilable on Bookworm. Neither it is compilable on Stretch.
-automata

Earlier in this thread you said you got 2.9 to compile. I am wondering a few things:
1. What steps you took to get that to work?
2. What version are you working with currently?
3. I wonder what the last version that would compile would be assuming you did get v2.9 to work.

The 9 axis look ahead should be in v2.9, so if you can make that compile, then it might be a start, and worth doing something like Andy did to see what all is different and go from there?

Just thinking out loud.

It's not, but I think a good step 1 is to add the 9 axis blending, and then step 2 would be to look at those changes. It would probably be easier to implement that way.

Since you're observing activity at hal pin “motion.digital-in-xx”, this confirms that the hardware input is functioning correctly. The issue likely lies further downstream—could it be that Gmoccapy is monopolizing this pin?

seuchato wrote:

I was looking for a way to control my vacuum table's vacuum pump from within axis.

To late, but:
Axis (and probably other GUI's) already have built in control for Mist and Flood so you could have just used those as that would also make it possible to use the M7/M8/M9 codes to control them directly, more info here
linuxcnc.org/docs/html/gcode/m-code.html#mcode:m7-m8-m9
For more outputs/stuff, you would have to go with what you did, and there are probably many ways of doing it, and you can use PyVCP panel to show info about it at all times:
linuxcnc.org/docs/html/gui/pyvcp.html
I use PyVCP panels a lot with Axis GUI, a lot, even for showing all Mesa IO and everything else, on industrial machines.
forum.linuxcnc.org/30-cnc-machines/31792...ofit?start=100#88523
See the right side, all is added as PyVCP panels.
P.S.
I was just searching for a retrofit i did to post the link here, bumped into my video from youtube posted by some russian company on rutube !!! And they wonder why the whole world hates them!

langdons wrote:

Please post output of ip a

He did, so why ask again?
-
On topic, that seems an issue with the EtherCAT key, it has nothing to do with networking, also supported by "modprobe" in the error as the module could not be loaded/inserted.

ATC design is final. Also did up a quick 4th axis rotary table using a surplus harmonic. The body is a steel block, faceplate is steel. printed nylon dust seal, and then an 60mm motor shaft coupling. It takes a  52mm round quick change plate for vises. Just enough centre height to spin the small xindian h67 vise with a 4" long work piece. Total cost with the harmonic drive, 200w ethercat servo, machined parts from JLC is in the $800usd mark. Not bad!   There are a small amount of these surplus harmonics, so if someone wants to build their own high quality rotary, I can give you the files. (I cant post publicly on here though). Note that the precision of the housing block is probably not going to be perfect, and you may way to scrape in the bottom to be dead square square.

This is still not the build with the jerk limiting right?

A parallel port has 12 output pins, that's a lot to work with.

But in case more than 12 outputs are needed, this method is great for non-latency-sensitive stuff.

Please post output of ip a

You may find netplan useful.

netplan.io

What about the BSD latency command?

Never hurts to have some extra RAM though, for I/O buffering, cache, etc.

I have 16GB because I have an insane amount of RAM from old PCs.