Linuxcnc motion backwards (adaptive-feed negative value) and nothing happens

Hi nerd, I did have troubles about 12 months ago with the Gmoccapy "all home" button on a gantry machine but Norbert pushed some fixes very quickly and its worked ever since.

Just check your ini file for this line

if it has "kinstype=both" on it, remove that setting. This will then prevent you from selecting joint mode and leave you with home all, home x, home y and home z buttons. All should work.

Also ensure that your paired joint axes have a negative value for the home sequence. in V 2.8, this will enable the joint axis homing feature
and all other joints have 0 or a positive value for this setting. I found it is wise to have the Z axis set toso it is the first axis homed to lift the torch up ot of the way before moving the X or Y axis.
Hope that helps.

@Hi Rodw,

How are you? Thank you for your comment's on the other topic.

Gmocappy has a powerfull code. Very good !! I am lucky that yesterday i had made a upgrade.

I am trying out bit by bit to change the screen to my desires. It takes a while, but in the end it will be practical for plasma.

Homing sequence is now done by one click for all axis.
Running the program is also one click.
Fixed axis position's without auto generating 4 and 5, and up to 9 (deactivated some line's off code).

This week i want to integrate the negative adaptive speed. This will take a while. I hope this is possible with this Linux version.
I think easyest is to make a button for negative speed command.

After that i want to integrate the THCud function for the torch height controller. This function works fine.

Its good that you are making such great progress.
I'm curious what you are using the A & B axes for. rotary axis or torch angle or something else?

I noticed earlier in this thread that you are using version 2.7. For a gantry machine, Master branch (v 2.8) is a much better choice as it handles the joint axis homing so much better. Some of my comments assumed V 2.8 but that might make the negative adaptive feed hard to implement. I wonder if you could somehow trick the system so that commanded velocity was a lower level than what it should be (eg at 0.5 the speed) so you can change settings up or down with just the positive adaptive feed?

And yes, Norberts code is very well written. I've dug into it to fix some bugs and to extend the screen handling.

Hi Rodw,

I implemented the adaptive-feed option in Gmocappy. It's not working negative on the version 1:2.8.0 pre. Pcw know's this also i think. The function is working only on value's 0.0 to 1 for full adaptive speed combined with the Macro code M52 P1. The only float in command left is analog in. But this is unusable for negative feed. Also a negative g-code value gives error on the g-code intepreter.
I have not looked in the intepreter source code. Maybe here is possibility.

Tricking the system is not possible also if you use a negative feedrate.

The most quick way is to change a few lines of code in the command.c file and a few other files. The log's for how to do this lives on github. It's not very difficult. Just copy past... en delete some lines of code.

But i can not find this "command.c" file on my system. I only see this file in the github download's. But on my system i don't know where it's located. Also i have no file search function in Thunar. Maybe someone know's how to.... Maybe the file is gone after compiling linux on your system?

The A and B axis are for machine's that need this axis. For example my tube cutter go's up to 5 axis. Also i have gantry machines
with double head's or side tube cutters.



For me is learning linux cnc a investment in the future of my company. It takes a time. But in the end you have almost
unlimited possibilities.

Here is a waterjet with 5 axis. ( 1 stepper for grid supply to the waterjet nozzle. ) Also a ceramic beared milling unit.
This is an old type of machine for me. The newest types are very cool and more hufter proof guide way protection.



This is a new type of machine, with integral guiding and drive system. Also 4 axis. Working area : 3000x6300x180mm



Maybe i must contact Norbert. He can do a great job for me.

Grotius wrote:

Hi Rodw,

I implemented the adaptive-feed option in Gmocappy. It's not working negative on the version 1:2.8.0 pre. Pcw know's this also i think. The function is working only on value's 0.0 to 1 for full adaptive speed combined with the Macro code M52 P1. The only float in command left is analog in. But this is unusable for negative feed. Also a negative g-code value gives error on the g-code intepreter.
I have not looked in the intepreter source code. Maybe here is possibility.

Tricking the system is not possible also if you use a negative feedrate.

The most quick way is to change a few lines of code in the command.c file and a few other files. The log's for how to do this lives on github. It's not very difficult. Just copy past... en delete some lines of code.

But i can not find this "command.c" file on my system. I only see this file in the github download's. But on my system i don't know where it's located. Also i have no file search function in Thunar. Maybe someone know's how to.... Maybe the file is gone after compiling linux on your system?

The A and B axis are for machine's that need this axis. For example my tube cutter go's up to 5 axis. Also i have gantry machines
with double head's or side tube cutters.

For me is learning linux cnc a investment in the future of my company. It takes a time. But in the end you have almost
unlimited possibilities.

Here is a waterjet with 5 axis. ( 1 stepper for grid supply to the waterjet nozzle. ) Also a ceramic beared milling unit.
This is an old type of machine for me. The newest types are very cool and more hufter proof guide way protection.

This is a new type of machine, with integral guiding and drive system. Also 4 axis. Working area : 3000x6300x180mm

Maybe i must contact Norbert. He can do a great job for me.

Reverse feed rate override is only implemented non-merged branch so currently you must fetch the reverse
feed branch you need (with git) and build it from source.

@PCW,

Thank you.

I downloaded with git, yesterday the Linux version i needed for "negative motion adaptive" on a pre installed linuxcnc pc.

The installation process was giving errors at the state off : ./configure.

This errors are more seen by people, if you google on this error code. ( missing tlc source path.... )

What do you advise me? If i have a new pc here tomorrow, how can i install this version without errors?
When you know how to do it, it's alway's simple, i know.
I need some help to keep installing simple. When the install is done, i can clone the hard disk's with a device no problem, that spares time in the future.

Is it possible to sent a pc to your home adres. Maybe you can fix it for me?
Or you buy a pc and sent it to me, that is maybe better. I live in Holland.

I don't know anything about the branches you want to use, but the author has pushed his branches over to the LInuxCNC git repository where you can select them from the drop down where it says "Branch:master".
github.com/LinuxCNC/linuxcnc
Look for the branches that PCW listed.
That means you can select this branch and do a git checkout of the code without resorting to working with a git repository outside of the linuxcnc ecosystem.

That also makes it possible to do a git merge to merge or rebase the changes in the branch with the current master code base. Of course this could break the code base as its been a while since the branch was written. Your goal should be to merge this branch with the current master branch so you are using the superior joint axis homing and gantry alignment that exists with master but was written after the negative adaptive feed was written.. If you worked completely within the git environment, then you could ask the developers to merge your code into master branch so it is maintained in future releases.

I don't know enough about git to advise you further but I have merged branches so its not that hard.

@Rodw,

Now it is more clear for me how to do things.

I will contact Mr. SebKuzminsky

Merging is not a future component to use. if you have update's they are not compatible....
It must be a clear statement in download or version i think to hold the strukture.

I have contacted Mr. SebKuzminsky waiting for responce.

With Linuxcnc you see him all around. He's a importand member in the club on git.
github.com/SebKuzminsky

I'm glad I helped. Bringing the branch up to date is probably not that hard as surely robEllenberg only modified a few files. There will be a way to run a "diff" against the current version of those files to see what has changed eg show the "difference". However, If I was in your position, I would find a developer to do it for you and simply direct the project. This will allow you to focus on higher value tasks that drive your business forward. (I tell myself to do that every day and sometimes it hurts to pay the bills for outsourced specialists but in the long run it will be what grows my business eg. let me do those things that make $1000 an hour, and not do the $20 an hour tasks that others can do.)