Post Processor for LinuxCNC and PlasmaC

phillc54 wrote:

pl7i92 wrote: i voter for having it as we here use it since years now and are very happy to enable thc first after piercing
so the thc internal got a 1sec space and i got a very nice cut after entry
no Hopping

If THC is enabled there is already a delay for THC after piercing.

Cheers, Phill.

What about at the end of a cut?

I have had problems cutting 20mm plate where the lead in was 10mm and when the torch came around to the lead in of the part the THC caused it to dive into the sheet because it had already been cut because there was no steel left. When I disabled THC 15mm before the end of the part I did not have any problems.

It really is a powerful tool to be able to control the THC with code.

Also off topic but those who say that 6mm plate up to 20mm plates does not need THC obviously has not done much cutting themselves with those thicknesses. I have seen 16mm plate buckle up like the sheet in Rods THC video. I had a batch of 16mm plate come in that was bowed in the middle, I needed the parts at the time so I used it but it was when I first started plasma cutting and it really tested my Sheetcam settings.

thefabricator03 wrote:
What about at the end of a cut?

I have had problems cutting 20mm plate where the lead in was 10mm and when the torch came around to the lead in of the part the THC caused it to dive into the sheet because it had already been cut because there was no steel left. When I disabled THC 15mm before the end of the part I did not have any problems.

It really is a powerful tool to be able to control the THC with code.

Also off topic but those who say that 6mm plate up to 20mm plates does not need THC obviously has not done much cutting themselves with those thicknesses. I have seen 16mm plate buckle up like the sheet in Rods THC video. I had a batch of 16mm plate come in that was bowed in the middle, I needed the parts at the time so I used it but it was when I first started plasma cutting and it really tested my Sheetcam settings.

Yes, I had some pretty brutal learnings from the little bit of 16mm I cut but that was before plasmac. I just wish I had a bit more time to play. This is where the kerf crossing should come into play as when a slug drops away, there is a violent dive at the end of an inside cut which motivated me to look harder at the kerf/void sensing. In fact now I think of it, I might add a second external disable pin to plasmac so I can hook up my almost working void sensing component..... if we could get that right, you might not need to insert the THC disable codes.

Rod,

Your void sensing component sounds interesting, I am just worried that adding all of these components will add a lot of complexity to in my humble opinion is a system that doesn't need to be so complex.

Just from my experience to get good cutting results from air plasma you need one main thing and that is full control of the torch height control. If we can do that with some simple G-code I believe that would be the best way forward until such a time when we have working components to take care of that.

At the end of the day if the codes are placed in Plasmac and they are not needed by the user I do not see any downside to having them added,especially when Plasmac gets added to the main Linuxcnc release.

And I imagine it wont effect Phills original plan to be able to use the branch without a CAM program.

My experience with plasma is quite big over the years.

If you have a g-code with top level info, a m3 and m5 for torch on-off, a S number for plasma power and eventually a tool number, and later on a M30 is the most simple way, and direct compatible with all kind of program's like inkcsape. This will result in a very clean g-code.

The component has to do all the work !

I prefer to extend the G-code only with the usage of G40 and G41 and G42 kerf width offsets. This way is used at the python
coded dxf2gcode program, wich is a perfect solution. Even a better solution then the actual Sheetcam g-code output.

I made a short list of how a component should work :

- Auto enable the component on program start, Auto disable the component on program stop.
- Arc oke simulate button is a must. Otherwise you can not simulate the program.
- Torch restart button (included a 1sec pat'h reverse run logic, connected inside the component to program pause inhibited).
- Pierce height input.
- Pierce speed input.
- Pierce delay input.
- Cutheight input, standard = 0.
- Torch height control correction speed input.
- Probe on-off button.
- Probe search limit input.
- Float switch offset input.
- Probe over distance input.
- Probe speed.
- Divided voltage input (digital 0-10V) - or up-down signal input for the proma version.
- Target voltage input.
- Torch on-off button, for cutting manual ( outside the component, no thc needed, 90% edge start )
- Modified "Adaptive Feed speedbutton" from -1 to 1.
- Optional : Plasma power input.
- And so on.

Attached to give GPL2 code back, my latest component THC2. Used on my latest Ethercat machine.
It has all the option's that i find usefull. Tested on a plate and rotary cutter. The component comment's need to be updated, all todo comment's are integrated and tested.

The nice think about the usage of a component is, you don't have to connect all the i/o signals to it.
A few connected signals are enough to make it working.

thefabricator03 wrote: Rod,

Your void sensing component sounds interesting, I am just worried that adding all of these components will add a lot of complexity to in my humble opinion is a system that doesn't need to be so complex.

My thinking is that complexity in the controller is OK if it makes the user experience simpler. In a perfect world the user should just load a file and cut it. The machine should look after all of the complexity in the background. Phill has not nailed void sensing in plasmac yet.

All of these components currently is just one, the plasmac.comp file. Keeping new features seperate is not a bad idea and Phill can decide if they get incorporated down the track either as a seperate component or by incorporating into the main system.

I might add that if I could also get Linuxcnc to calculate the arc radius, there would be no need to ever use sheetcam code snippets....
Now that would be cool

Rod,

I agree that the user experience should be simplified when possible.

One thing I like about Sheetcam is there is many settings that I can choose from to fine tune my machines cutting.

My main concern with simplifying the experience is that you loose some of the ability to make those small tweaks to some settings to dial in the cuts.

I would love to have a system where I can ask any one of my guys to go to my machine and have them cut great parts with little experience. I would say I am in a good position to help this community to get to that point as I cut steel from 1.6mm up to 25mm in thickness.

I worry though that the settings that work great for my machine wont perform the same on Rods machine or Phills machine. Grotius might be able to comment here on the ability of setting the cutting parameters for different machines and getting the same cut quality.