Advanced Search

thanks Tommy,

Using the same linuxcnc post in sheetcam that i have been using for ages.

grijalvap wrote:

yes it works if you use 2 tools one with piercing delay and one with 0 delay.
I'm using the tool from sheetcam (tempoary tool created from file). so i need 2 files or edit the file after pierce and reload.
what I'm thinking is just press a button and make the Pircing delay 0,

In the Sheetcam processor, there are some variables at the top of the file to select between Sheetcam and qtplasmac tools. Sorry about that but had to humor Les at Sheetcam seeing its distributed with his software (and he wrote the post.)!

There is also a script that lets you import a Sheetcam  toolset into qtplasmac

snowgoer540 wrote:

Temporary materials were meant to be temporary, QtPlasmaC and PlasmaC before were not built with a temporary materials workflow in mind.

Nothing more permanent than a temporary solution! Quote from someone on the internet...he/she was bloody right!
-
I absolutely and strictly use the materials in PlasmaC/QtPlasmaC, using anything else is asking for trouble.
I have modified the Inkscape gcodetools to output the required plasmac-feed-rate in gcode, and that was it.

The thing is, a round 20mm tube is roughly 63mm in perimeter (or in this case the distance the machine must move) while a 100 round tube is 314.15mm in perimeter, so if you have the same rotation speed, the 100mm will travel much faster under the spindle.
The easy way to avoid any of this is to have multiple normal configurations for each diameter, and use Y as rotary but in hal it remains linear. Not using A.
I have several rotary plasma machine in use with this type of configuration, so normal 2D CAM (Inkscape with gcoodetools), and just run the configuration for the diameter being cut.

spincycle wrote:

Ah using Linux to entirely tune the feedback loop?

Yes, the position loop is in LinuxCNC.
Velocity loop is usually on the drives when motors also use tacho generators.

spincycle wrote:

Am I right in thinking that if you run analog signal and rely on linuxcnc to close the loop, all tuning happens within Linux, as per your link?

No auto tuning through the drives etc?

Yes, again only the position loop is handled in LinuxCNC, velocity and torque loops are handled usually in the drives, although LinuxCNC can also handle all of the above on it's own.
The tutorial is for using older existing drives, whenever possible i try to use them as they are very good quality and rarely fail, and most of the tuning is already done in the drive as they were in use and working perfectly fine.

You could just stop using temporary materials and let the materials in QtPlasmaC be your master.

Then you only have to perfect settings in one place. Furthermore for your situation you’d go to another tab, change the delay,  do your pieces go back to the tab change the delay again and run the program. No editing gcode files necessary.

Temporary materials were meant to be temporary, QtPlasmaC and PlasmaC before were not built with a temporary materials workflow in mind.

Might not be the issue, but you should really avoid using Mach3 post processors for LinuxCNC, if you use any for plasma, it will shows each consecutive cut on screen lower and lower. Try something with 200 cuts!
If you must, use the plainest post processor, mill should do probably, again with no gcode for tool changes n'stuff.

I have a cnc router with a laser and a rotary axis.  I modulate the laser output (using the laserpower HAL) according to motion's requested vs current velocity, so the laser power is reduced during starts and stops.  Works fine for XY work.  For XA work, I'm running into the "which F" problem.  I'm using inverse time mode for the gcode but... how do I get inverse speed for the laser modulation?  motion.current-vel follows the F rule - it's inch/min if there's linear moves, else it's degrees/min.  If you do a coordinated move, you lose the angular velocity info.

In my case, I know the Z value will be "accurate" as far as "how far from the rotary axis are you" *in this case*.

Is there a way to do this in HAL?  Preferably without writing my own laserpower module

Or is there a better way to modulate laser power with a rotary axis?

Or should I just assume "it will work out" because the acceleration for X and A is synchronized, so reading either velocity fraction would suffice?
 

SO.......
wifi adaptor died on the mac I'm using for the cnc.

Easy, new Mac.

load linuxCNC 2.9.4.

forget to save a backup of old config (dumb ass !!!)

Spend last night making a whole new config for the machine.

Save a backup on usb this time and place in a safe place.

Having a few glitches, hopefully sorted soon.

Just downloaded and installed 2.9.4.

Having a few issues with the loaded job.

It cuts once then deciding it wants to show the job 2500mm low on the Y axis. Although it will still cut where it is supposed to.
Also throwing error codes if i change any of the .tap file through the edit option.

Is there a new one I need to install?

Looks like your min and max limits on X and Y (and joints 0,1) are reversed.
 

 

Ah using Linux to entirely tune the feedback loop?

I am interested in Ethercat communication, but I am getting slightly cold feet as it appears to be a path far less travelled in this forum.

Am I right in thinking that if you run analog signal and rely on linuxcnc to close the loop, all tuning happens within Linux, as per your link?

No auto tuning through the drives etc?

Thanks tommylight.

yes it works if you use 2 tools one with piercing delay and one with 0 delay.
I'm using the tool from sheetcam (tempoary tool created from file). so i need 2 files or edit the file after pierce and reload.
what I'm thinking is just press a button and make the Pircing delay 0,

Silverdragon is based on gscreen and gtk. When I ported it to Qt, I renamed it QtDragon. I have since created an updated version on github, but it's still QtDragon. I'm not creative enough to come up with a cool new name.
Jim