Advanced Search

Hello everyone,
first of all, thanks to Wischi for the touch gesture version of gremlin.py — it works perfectly on my hardware (Melfas LGD AIT touch panel, LinuxCNC 2.9, Gmoccapy).

The only limitation for me was the one HansU already mentioned: with the mouse you can only rotate, dragging/panning does not work in any setting, because motion() returns unconditionally.

I'm attaching a version that is Wischi's file unchanged, with one addition: instead of disabling motion() completely, it now filters events by their source device. Only events that clearly come from a real physical mouse are processed — the event must not be pointer-emulated, the source device must be classified as MOUSE, and the device name must not contain "touch". Everything ambiguous is discarded, so the touch behavior stays exactly as in Wischi's version.

The background: many touch controllers expose a second X input device that does driver-level mouse emulation (mine shows up as "Melfas LGD AIT Touch Controller Mouse" in xinput list). Its events look like genuine pointer events to GTK — no emulation flag set — and those are what caused the original "preview only moves a couple of mm then stops" bug in the legacy motion() path. That's why the filter also checks the device name, not just the emulated flag.

Result on my machine:

Touch: unchanged from Wischi's version — one finger rotates (perspective view), two fingers pan and pinch-zoom
Real mouse: full mouse_btn_mode behavior restored — left rotate, middle drag/pan, right zoom (mode 0), shift+left drag also works
Since motion() is active again, the mouse button mode setting in Gmoccapy works again too
Tested on LinuxCNC 2.9 with Gmoccapy, capacitive multitouch panel. As before: just replace gremlin.py in /usr/lib/python3/dist-packages/ (keep a backup of the original — it gets overwritten on updates).

Changes:
1. Lines 40–50 — header comment (documentation only)
2. Lines 676–699 — new helper function _is_real_mouse_event()
3. Line 700 — motion() 

Dislexic

Its in Master branch.

May I ask how the project is going? Any update?

QEMU?

QEMM on Linux is more to my taste than Virtual Box in Windows. On Linux VB kept getting in a tangle over UEFI

Simple answer would be find the old image and do not change it, but no idea how or where or anything as i do not use RPI's.
From what i read here (hence the i do not use them as they are way to much trouble), there is something that needs to make GPIO active, but i do not recall what or how.

Take some pictures of ...everything and post them here.

This is the second time i am reading this, most if it is your statements, not questions, the only actual question i can find is about the dual Y axis, and that can be done in multiple ways, from using the same output for both, to have two separate outputs with two separate home switches and the added possibility of tramming the axis every time the machine homes.
The Mesa 7i96S should be all that is needed from what i gather, but you will not have position feedback to LinuxCNC and that is fine, but if you do want that and the drives have encoder outputs, you can add a Mesa 7i85 or 7i85S as it has 4 encoder inputs, and there is another one on the 7i96S for the spindle feedback.

We would need much more info to be able to help.

Nice job! keep on iterating!

Hello @Kuzma30,

I'm looking at starting a EtherCAT build and considering the AMSAMOTION EC3A IO1632. Have you been able to make it work on your setup and could you share the EtherCAT-config.xml file ?


Thanks ! 

FabLabRacing wrote:

Wins feel good, even if they are small ones

 

Yes, they do.

I have made a few updates to the code & I feel like I hit a pretty big milestone, I was able to get a full circle trace (for the first time) I set the software to use "to many" move points so I got an overrun of ~77 deg but for a camera-followed trace on a printed line, this seems pretty respectable. Especially because this is not post-processed yet — this is raw captured polyline.

This is the raw data:
Points:             250
Average segment:    0.02922"
Total trace length: 7.276"
Fitted radius:      ~0.9501"
Radial RMS error:   ~0.0108"
Radial p2p error:   ~0.0495"
Angular coverage:   ~436.9°

The overlap:
First full-lap return point: ~0.0167" from start
Overrun overlap average:    ~0.0114"
Worst overrun overlap:      ~0.0179"

The circle was drawn with a 1" radius

Apologies, if this more exciting to me than anyone else, and if this is boring data, but I felt it was "share worthy"
Wins feel good, even if they are small ones
 

Yep, that's it. Should be in ~/linuxcnc/nc_files/examples/remap-subroutines on your actual pc. Just know that it's changing the tool length offset (tool table Z) instead of the actual WCS Z offset. So you put the height of your tool setter above the machine bed in "probe HT" of the probe tab (leave "block HT" at 0 as mentioned before). Then you put the height of your workpiece above the machine bed in for your desired WCS Z offset. 

So example would be the toolsetter is 1 in above the machine bed and your stock top is 3 in above the machine bed. You set "probe HT" to 1 and your G5x Z offset to +3. You have to turn on the "tool measure" button in the probe tab and now when you run your program, the remap automatically takes over once the M6 line is reached and it will bring the spindle to your defined change position in your .ini, then after you change the tool and hit OK, it will drive over to the defined tool setter position in your .ini and probe the tool which will automatically fill in the tool length offset in the tool tab. Same diagram as figure 7 in section 14 of qtdragon docs, but "block height" is your g5x z offset. If you want to run without the remap, just turn off the "tool measure" button in the probe tab. It will still bring the spindle to the change position, but doesn't probe the tool and doesn't create a tool length offset. Nice thing about running the remap this way is it creates a universal tool measurement independent of the part setup allowing you to run multiple WCS offsets at once. 

You need to make sure your .ini has all the things shown in section 14.3.2 of the qt dragon docs for this to work. it's also very important to make sure you are using the qt_auto_probe_tool.ngc file from 2.9 and not the 2.10 master branch as the 2.10 version is different formula and won't work the same. Included my .ini, .hal, and personally modified .ngc remap if you want an example of what it looks like.