Advanced Search

Hi, 

​​​​​​JT thanks for your reply.
additional for this situation question. I have two sliders. Usong Slider1 user may choose value 10..20. slider2 could be 0..100 . But if slider1 is more than 15, Slider2 could be only 80..100.
How i may realise this?

Thanks in advance.

jpg wrote:

Spumco hasn't said anything since... 

I've been on an LCNC hiatus for 6 months due to some other work eating up all my free time.

I've still not gotten the new spindle motor completely hooked up and tested, much less tried Aciera's fix.  Hoping to get back to it in the next month or so now that the outside work has finished.

I found the probe signal problem and got the EasyProbe screen to come up.
Something is still wrong - clicking any probe button gives an error:
EOF in file: /usr/share/axis/images/axis.ngc seeking o-word: 0<probe_x_minus> from line: 0
I also do not have the Probe Signal LED working. So, my "fix" of the probe input error must have broken something that does not cause an error message on startup.

It also gives the following error when I press the home button:
"emc/task/taskintf.cc 976: Error on joint 0, command number 74"
if that helps at all...

I made sure that they agree.

Do you know what
"(PREOP + ERROR) could be caused by? or "Invalid input configuration"?

Also, did you take a look at the .hal and .ini file? There are a few things in the .hal file that I don't know the purpose of, yet LinuxCNC doesn't run without them, so I'd appreciate an explanation of why they're necessary-if you're willing.

Personally I would suggest tp use the .comp format:

linuxcnc.org/docs/2.9/html/man/man9/userkins.9.html

But regardless,
Forward kinematic uses the joint positions to calculate the cartesian coordinate positions 
Inverse kinematic uses the cartesian coordinate positions (ie the position commands of the motion planner) to calculate the joint positions (ie basically the commands that get sent to your motor drives)

One needs to be the mathematical inverse of the other (otherwise the values will run off and you will get limit violation errors)

I was able to check the connection yesterday, you were right, [GND pulse - dir -] I had them bridged. !!

I always connected them, I understand that it's wrong! I like to learn because the board can get burned with this connection. I thought that the ground from both the driver and the board should be connected.

thank you!!!

I just checked and I was actually on LinuxCNC 2.9.4. QtPlasmaC version is v2.9-238.315. Is that version effected by the bug you mentioned? I'm going to assume so since the gcode works for you.

I've also just deployed a new version of MetalHeadCAM to metalheadcam.alan-gibson.workers.dev/. I'd really appreciate your feedback on it as I'm specifically targeting QtPlasmaC.

The basic workflow is:
- Click the hamburger menu and then Tools so you can add your tool parameters
- Import a DXF file on the Import stage
- Skip the Edit stage for now as it doesn't do much yet
- Click the topmost Apply button on the Prepare stage
- Add operations to parts and shape chains on the Program stage, making sure to select what you want to cut in the Apply To dropdown.
- Simulate the program on the Simulate stage or just skip it
- Download gcode on the Export stage

Hi All,
I've just released an update that provides improved lead in/out support. Any feedback is of course appreciated.
 

qtdragon.probe-error pin is an I/O pin not an input pin.

I'll have to think more on the best way to add external error checking/handling.
It's a problem spot for linuxcnc.

Ok, sorry for the images. Anyway, I performed all the math on Google Sheets, so you can enter the B-axis position on the ball screw, and it will provide you with the spindle angle, the tooltip x-offset, and the tooltip z-offset. The range of travel for the B-axis is 0 to 305mm. You can enter this in cell Y11, and it will spit out the results. I have also illustrated the drawings better here so that hopefully you can understand. Here is the link . I have looked at the Kinematics tutorial on the website, but I just learn better if I can see someone do a working example. After hours of reading that with Gemini helping with some of the explanations, I don't know how to start writing the code. Anyway, from the tutorial I posted earlier, there is an image of this person's kinematicsInverse function, which I will insert below.

From what I've understood, the job of this function is to understand the required tool tip movement and spit out the number of steps each axis motor needs to make, correct? If that is correct, I should be able to modify the trigonometry here to accommodate my ballscrew articulated B-axis and get this going, hopefully. Because his setup also has an XYZAB setup similar to mine.

If your're running a mes card there is no need to test with a base thread. Servo thread is all that is needed.
Usual advice, turn off power saving, anything that attempts to change clock speed in the BIOS, also turn off hyper threading if that is an option, for a 4 coremachine you might get some relief from isolcpus=2,3

It may help to supply some details regarding the machine you're using.

RADXA Cuby A7Z
radxa.com/products/cubie/a7z/

SoCAllwinner A733 CPU2x Arm® Cortex®-A76 up to 2.0GHz6x Arm® Cortex®‑A55 up to 1.8GHz1x Alibaba RISC-V XuanTie E902 up to 200MHz

GPIO40-Pin GPIO Header with I2C / SPI / PWM / I2S etc

MCU: RISC-V E902 (200MHz) Coprocessor, Real-time tasks and low power management

Unfortunately I didn't see documentation yet

I'm trying to get EasyProbe running on my mill. I had an older version running, but am having problems with the probe signal.
The report file shows: Probe_postgui.hal:14: Pin 'hm2_7i96s.0.gpio.040.in_not' was already linked to signal 'probe-in'
Probe_postgui.hal has:
#

---

Probe signal and LED control (see readme.md)

---

# dbounce_probe.in must be linked to the probe tool input of the interface card
 
setp  dbounce_probe.delay      [PROBE]DEBOUNCE_TIME
setp  xor_probe.in0            [PROBE]PROBE_SWITCH
 
net   probe-raw                dbounce_probe.in             <=    hm2_7i96s.0.gpio.040.in_not
net   probe-switch             dbounce_probe.out            =>    xor_probe.in1
net   probe-signal             xor_probe.out                =>    motion.probe-input pyvcp.probe_led

I can't find any other instances of probe_in. Where might it be?

 

That got me up - but now I've got "Unexpected realtime delay on task 0 with period 1000000"

Latency test gives 60624 ns jitter on the servo thread and 70111 ns jitter on the base thread.