Another plasma component...

All my machinest experance centers around lathes and mills. I am quickly realizing plasma is a diffrent beast. I have never used G64, I will play with it.

JTknives wrote: Anyone know why it would start cutting round corners.

And sometimes it’s still not using the THC. It touches off and just takes off but never adjusts the Z.

John has set you straight on the rounding. When I first got my table moving, I drew a large square at maximum speed and it rounded the corners by many inches. So i put it in the startup code (which in plasmac's case is in a .ngc file in the config directory)

If you have feed rates in the gcode, they will be used. To use the one set in the plasmac gui, you need to add where Pn contains the plasmac tool number. If you have a lot of tools (50 or so), you might need to increase the Q1 to Q2

If you have feed rates in the gcode, they will be used. To use the one set in the plasmac gui, you need to add
M190 Pn
M66 P3 L3 Q1
F#<_hal[plasmac.cut-feed-rate]>
M3 S1

Rod is on the right track here but you do not have to use the M190 and M66 lines, they are only used if you are specifying material numbers. You could omit these and it will use the parameters of the currently displayed material. You do need the F#<_hal[plasmac.cut-feed-rate]> if you want to use the displayed materials cut feed rate. If you specify a feed rate in gcode say F2000 then if that is less than 99.9% of the displayed materials cut feed rate then cornerlock will stay locked because the machine will not reach the displayed materials cut feed rate.

I hope that makes sense...

Edit:

Anyone know why it would start cutting round corners.

Have a look in imperial_startup.ngc, I have specified G64 P0.04 (1mm), it should have been 0.004 (0.1mm) it is correct in metric_startup.ngc.
Should I change these numbers???
I would really love for some input on setting the defaults to some sane numbers...

Cheers, Phill.

Phill

Ok, I made newbie mistake and didn't save the whole plasmac branch directory.

So I had a little time waiting for the boss to respond to a design proposal this morning and thought well my Plasmac is well out of date and has several known issues. So I blindly plowed ahead : This was more like a fresh clone than an update. I think every source file recompiled and linked.

Well it seems the whole directory structure has changed.

So went to build a new configuration after testing a sim, ok.

A new directory for my stuff and fixing all the broken links, the sim still works. I still couldn't get an actual configuration to run on hardware. After much fussing around I found that it worked by removing this from the .ini file:
# required for remapping gcode f word
#REMAP = F prolog=plasmac_feed_prolog ngc=plasmac_feed epilog=plasmac_feed_epilog
So is this bit now redundant?

I found the changes in syntax for defining the user button actions. It really is one of the confusing parts of LinuxCNC where the syntax changes for doing the same thing from file type to file type.

The next thing is the material file. When I start up there is only the default material available. I can edit it and save it but still only the default. I couldn't add to it either using the GUI or text editing the imperial_plasmac_materials.cfg file. How do I add new materials? The Kerf width spin box is broken.

The nice color coding of the torch enable button appears to be gone. Well I thought I'll just edit the glade file, no. Some how the button as setup doesn't let you set the active/not active colors.

Thanks for all the hard work, I know I should have kept up with the daily updates.

I still need to make a new PP to support multiple materials. Figure out how I enable/disable the THC (I am sure my old hack is broken).

John

To make thing easier (for me) when I got Axis going again I moved the configs from configs/sim/blah to /configs/by_machine/plasmac.
In this directory there is a file named configurator.py, it will (should) create a config from an existing vanilla ini and hal file for the base machine. So if you have a base machine config in ~/linuxcnc/machine you would run configurator.py and create a new config named something different to your base machine config, fill in all the required blanks and it will build a new config in ~/linuxcnc/newname with the new ini file called newname.ini. It will split the original plasmac.hal into plasmac.hal and newname_connections.hal and also create a postgui.hal file. Your original hal file will be untouched but a new copy called newname.hal will be in the config directory. The original ini and hal files will have a backup created in the newname directory. Keep all your hal customisations in newname_connections.hal or postgui.hal. When you want to update you just git pull (run make if requried) then run configurator.py and choose update and your config will be updated with newname.hal, newname_connections.hal and postgui.hal left untouched.

Clear as mud...

Cheers, Phill.

Edit:
Where I say base machine I mean just the machine itself without any connections to torch, ohmic, float switch, breakaway switch etc.

If you specify a feed rate in gcode say F2000 then if that is less than 99.9% of the displayed materials cut feed rate then cornerlock will stay locked because the machine will not reach the displayed materials cut feed rate.

Phil, a clear case where the state tags branch would benefit Linuxcnc immensely as plasmac would always know the current feed rate as its passed to motion where we can see it. If only it was merged into master...

I found that it worked by removing this from the .ini file:
# required for remapping gcode f word
#REMAP = F prolog=plasmac_feed_prolog ngc=plasmac_feed epilog=plasmac_feed_epilog
So is this bit now redundant?

John, yes, this is redundant (ask me how I know?). Its important to review your ini file against the current one in the sim while its still evolving. And yes, I learnt the hard way...