Thanks for your reply. I've read through the entire forum thread, but I haven't found anything specifically related to my problem.

I suspect the OUTPUT_OFFSET value isn't working correctly. When I configure the axis in the pncconf wizard and adjust the offset value so the axis no longer drifts, everything works fine. However, as soon as I switch to the normal Linux CNC environment and enable the axes, they start drifting. I also tried disconnecting the analog signal connector from the 7i77 to see what effect it would have. There's no noticeable change. The axis drifts slightly faster, but that could be misleading.

Could it be that there are missing links in my .hal and .ini files? Or do I need to adjust other settings? I've also experimented with the calibration, increasing the P value or setting all values ​​to 0, but this hasn't helped. The axes are still drifting.

Or is it possible that the drifting is normal (approx. 3-5 rpm)? As I understand it, the axis should stand still when the OUTPUT_OFFSET is set correct, even in the normal CNC GUI. Is that correct?

Okay figured that out too, the "Tool Measure" needs to be ON in the probe tab. So now I am able to T# M6, it moves to tool sensor, I get the pop up, I change the tool, it starts to probe down...suspense... touches the tool sensor and I get the error below.

On a side note the tool sensor does work. I would say to exclude out the "probe tripped during non-probe move" because when using the offset commands "Go to Sensor" and "Tool Sensor" it works and I do not get that error, what is the difference?  

Made a solid jump in progress. The issue was the qt_dragon would appear to "freeze" thus looking at the terminal led me down this rabbit hole. The problem was the Message Dialog was not popping up (or maybe it was because you need to acknowelge it to do anything). Either way I am starting to think the "not callable" is normal?

After checking my .hal file and my postgui.hal file the problem was this wasy in my postgui.hal file, I was missing this in my postgui.hal file and I comment out all the unlinkp lines ande uncommented the toolchange signals.

Attached is what I have as the lastest and I was able to somewhat make progress, it goes to tool senosr and the message dialog appears now! However after I accept I get the next error


 

I corrected the program prefix, to be one and have the "~" at the start. Ran again I am getting the following from the terminal. The path is valid as well. The following is from the terminal

I compared to the sim .ini file and found that the "~" was not part of the program_prefix. Running it again without the "~", cleared the error

Still does not work but this was noticable

Running the sim actually in debug shows the same thing.. 




 

Hi Oliver, thank you! I post the configuration (json.txt), linuxcnc_print.txt output and a copy of the terminal output linux_cnc_out.txt

I generall I know that I have to do some HAL configuration to (software) wire the machine together. What for the beginning is not so easy to know is what exactly has to be done. I see what your rio-setup does, and there is also the possibility to make linuxcnc HAL net connections in the setup, but what is necessary to start with..

I'm looking forward to hear from you,
regards, Robert

PS: do you run robot os on RPI5 Debian Trixie?

Hi,

thanks a lot for the update. I’m always happy to see new progress in this thread!

Also thanks for the screenshots, everything is looking very promising.I noticed that you’re writing the manual in German as well, so I assume you’re German too.

If you’re interested in chatting "in real time" in German with like-minded people, there’s a German-speaking Discord from Peter aka Talla83. You’ll find many LinuxCNC users there, including quite a few lathe users. If you’re ever looking for feedback or beta testers, that could be a good place to connect.

Of course, this is only meant as an addition, please definitely keep posting updates here as well. If you are interested, here is a Discord server invite link: discord.gg/dPBgNR3t

Best regards

Tom

Hopefully not a silly question if i got this tool change config to work could I use it with easyprobe with min effort? (Axis Gui) easy probe not yet installed but like the GUI AXIS/Easy probe very basic.

Thanks

Hi everyone,

I’d like to give a short status update on LatheEasyStep, mainly to clarify what has changed recently and what my current focus is.

 What has changed

1. Interactive legend in the preview

The preview now includes an interactive legend.

The title “Legend” is always visible and acts as a permanent click area.

Clicking on it toggles the legend collapsed / expanded.

When collapsed, only the header remains visible so the legend can always be reopened.

The legend affects visualization only, not geometry calculation or toolpath logic.

2. Improved preview with safety areas

The preview now clearly distinguishes different geometry and safety elements:

Green (solid): contour / target geometry

Yellow (solid): active machining path

Grey (dashed): stock outline

Cyan (dashed): retract planes (XRA/XRI/ZRA/ZRI)

Red (dashed): machining / safety limit (e.g. ZB)

Safety limit violations are highlighted visually.

Stock and retract planes are drawn based on the actual configured parameters, not approximations.

3. Facing – chamfer or radius at the end

In the Facing tab it is now possible to define an optional chamfer or radius at the outer end of the facing operation.

Supported options:

None

Chamfer

Radius

This affects both the preview geometry and the generated toolpath.

All of these changes are documented in the README, in both German and English.

 Current focus: the preview

Right now, my main focus is the preview system.

The goal is to make sure that:

geometry is displayed correctly,

safety and retract areas are clearly understandable,

and the preview can be trusted as a visual plausibility and collision-check aid.

Once the preview side is solid and stable, I’ll move on to the fine-tuning of the actual machining / stock removal functions and further logic improvements.

 Next steps

Further testing and stabilization of the preview

Real-world test cases with complete programs

UI adjustments if needed

After that: refinement and extension of machining features

Hello,
I installed Debian 13 (Trixie) and the latest Linux CNC 2.9.8 today. Unfortunately, the Mesa Configuration tool won't install.
I get the message: mesact: Depends: phyton3-distro but it is not installable. Perhaps someone could give me an additional tip.

Hi Robert,
can you please post your json config and maybe the complete error output ?!?

OK
In your configuration folder (or the sim folder if you want to test there) there will be a set of folders qtvcp/screens/qrdragon/
in that last folder add the file i have provided. name it qtdragon_handler.py

When you run the config, there should be an entry in the status page about a customization.
The button should be clickable after that.

I would suggest you try the sim first.

Can you post your hal and ini files?

The sserial_port0=something stanza in the driver command line
in the hal or ini file will determine which sserisl channels are enabled.

Does the 7I74 have 5V power?

 

Eureka!
I think I've found the cause.
While I was testing various settings in the ini file, the errors occurred sometimes on one axis and sometimes on another.  Even with unchanged settings, the error was sometimes there and then gone again. Until I turned off my handwheel (WHB04B). 
With the handwheel turned off, there are no problems with referencing. 
For some reason, the handwheel seems to interfere with the reference run. 
 

The 7I84 should connect to 7I85 TB2 pins 10..16

There are also some difference in the VHDL source and mesaflash pin list

1. You have 2 null modules descriptors before the PKTUARTS. I don't think this will
work as the module scan normally ends at the first null

2. The source you posted has  sserial channels starting at channel 1 rather than 0
(so sserial_port_0=2xxxx" would disable the first channel (1) on the 7I85)