Mesa thcad control loop

I talked to Peter at Mesanet trying to get a feel for the mesa thc and linux interaction with the information it provides. At present it is my understanding that linux just compares the voltage and bangs out an up or down move at a predetermined speed.. Knowing that linux is fully capable of sampling this information and with a proper closed loop is it not possible to get a better control over the THC? why would this information not be analyzed and adjust the torch z speed according to it's needs? It would seem that PID loop should be able to compare voltage and adjust such at whatever rate needed to track the material. IE the further target voltage is out of range the faster the correction is made and the closer the slower. If this is in fact how linuxcnc is addressing this then I am ready for the move. If it is not the way it is handled then I don't really see it as being significantly better than any of the other low cost (not so good IMO) solutions. I really want to know answers before i tear down a functioning machine ( although not as good as I would prefer) to make a serious changes and arrive at the same place I already am. there are many well proven stand alone thc that use a pid loop to follow the metal but being stand alone leaves the z axis out of control to the motion controller , linux or even mach has no idea where the z axis is. making an integral plate marker attached to the z axis almost unusable. While mach doesn't even have the capabilities to handle internally controlled loops, linux does if it can be implemented. And if I am completely mis understanding this whole scenario please lead me to the proper documentation so i can get on with making a better machine.
Can we have a plasma category?

You might take a look at the current implementation:

git.linuxcnc.org/gitweb?p=linuxcnc.git;a...ecffe87556e2;hb=HEAD

This is in the THC comp.

To add proportional control one possible approach would be to modify the THC
comp to use an external PID component (Linuxcnc's standard PID comp)

cwalker1960 wrote: It would seem that PID loop should be able to compare voltage and adjust such at whatever rate needed to track the material.

This should be relatively easy to set up. You would basically just need a PID component where the command value is the required voltage and the feedback is the actual voltage.
I would expect this to be an I-heavy PID loop.

The complication is likely to be setting the behaviour when the arc is not yet established. What is the THCAD voltage output then, and what behaviour is required?

per Andy
The complication is likely to be setting the behaviour when the arc is not yet established. What is the THCAD voltage output then, and what behaviour is required?

i would think there should be a way to let linux handle thc until arc is established , likewise turn control over to linux in tight corners (ie below set speed %) , short moves and overpasses as this is the way regular thc functions are handled
I'm not looking to do anything a typical low cost THC wouldn't do for anything other than following the metal at different feed rates based on a set target. another approach that has crossed my mind.. it is well documented what torch height will give the best cut for said material. although should the metal be rusty or pitted the voltage may not be accurate,, if the IHS set the height , then read the voltage from that height , let the machine maintain that height from that voltage , not a pre determined voltage that may or may not be right.
I'm hoping that I am explaining myself clearly here ,, in other words (just as a for instance) say 1/4 mild steel plate needs to be cut at .050 torch height .. the voltage may be 146 on a nice clean plate or 162 on a rusty plate , but both cut equally well at that height,, so the controller needs to sample the voltage of the cut that the touch off has established and maintain that voltage , what ever it may be ,,

BTW , the neuron thc does exactly this , only problem is , it is a stand alone controller that has complete control over the z axis , linux or mach is only the coordinate plotter with no idea or control over z at all ,, the neuron is a very good thc , makes some of the finest cuts consistently that I have seen on a hobby class machine , but leaving the machine controller out of the loop is unacceptable to me.

PCW wrote: You might take a look at the current implementation:

git.linuxcnc.org/gitweb?p=linuxcnc.git;a...ecffe87556e2;hb=HEAD

This is in the THC comp.

To add proportional control one possible approach would be to modify the THC
comp to use an external PID component (Linuxcnc's standard PID comp)

It looks as if this is getting close to the way I would like to see thc handled. i will be ordering my mesa cards soon , I was really waiting to see the latest info on the 7i76e card before committing to anything.

Hello

Have you made any progress on using a PID loop for THC? It is something I want in my THC. I don't want to have to reinvent the wheel here if it isn't necessary.

John

I have not. I bought an old hurco vmc that I'm working on getting converted so for the moment I've just left my plasma table running on mach3 with the proma thc.
This is still something I want to work on when time (and money) allows so keep us posted on your progress.