Plasma Questions

Hello

I am using the THCad card from Mesa. So far I haven't used it with the plasma table as my control is still under construction. Seams to be plenty stable over the voltage range I need it work in.

Progress, I have my system probing for the top surface of the sheet and going to the set pierce height set from the GUI. This also means the new Z axis is functioning with limits and homing. Now I need to get the external Estop figured out along with reading control settings out of the Sheetcam generated Gcode file as generated by the CandCNC PP. I really only need a subset of the parameters because I don't have a fancy Hypertherm power supply that can be adjusted on the fly via a serial connection. I will post progress shortly.

Off to a trade show and selling for the rest of the week.

John

John, I've attached my work in progress hal file that has the estop working. The breakaway sensor is also tied to the estop. It also has two switches for Run/Step Pause/Resume as per this thread. forum.linuxcnc.org/47-hal-examples/13201...-hold-resume-buttons.

I've since added a working MPG to the X axis. I was going to run separate ones on X and Y. I posted the code here.
forum.linuxcnc.org/27-driver-boards/2773...-7i77?start=20#85131

I think from memory, I used your HAL as a starting point.

I see it but no sure about the resolution, do you think it is enough for the application?
also, I like to have the encoder interface free for others purpose.
does anybody have experience with this card?
thanks

Hello

I have another question for the plasma table and developer communities.

I have an adjustment and display in my GUI for a value to limit the Z axis speed for smooth torch height control. Presently this is just a percentage which will require some calculation to get a usable value. The two ways I have found to implement this are:

1. Used the limit1 hal component connected between the Z axis PID and hardware step generator (Mesa 7I76e).
2. Use the maxoutput input to the Z axis PID.

Which of these is the preferred method? Is there a better way?

TIA

John

John, is that
you are trying to control or something else?

rodw

Well that isn't what I was thinking of doing but it may have the nearly the same result. My original intention was to limit the maximum velocity command to the step generator. I really don't care about limiting the the THC.comp gain (thc.correction-vel). thc.correction-vel is the correction step applied for each servo period (I think). What I want to do is limit the maximum velocity command to the step generator with a value from the GUI and not limit the acceleration or rate of change from the THC.comp beyond that set in the .ini file. I am sure that someone out there has a functioning system with this feature.

John

I would have thought you have control over Z axis velocity in the.ini file settings. From what I could work out the THC component tells the system to move a certain distance and therefore the default axis settings in the .ini file would apply.

Part of me still says that THC should not update the position until the commanded movement has completed but until I have working hardware I can't explore this stuff. Maybe the range set by thc.voltage-tol is enough to stop seesawing movements.

rodw

This may well be a feature that other systems have to make up for other short comings, all I know is that the present leading commercial systems have this adjustment. It may not even be needed for a Linuxcnc system that is tuned correctly. I intend on having the control set a percentage of the maximum velocity from the .ini file. That way I can change things during a cutting session without editing the .ini file for every trial. I really don't need this because 98% of the time I am cutting the same type and thickness of material.

The THC.comp updates the offset for an arc voltage error once every servo period when enabled and the table is up to speed and the torch is commanded on. The correction (very small for each period) is added or subtracted from the offset if there is an arc voltage error. When enabled and the torch is on the THC.comp does not use the Z axis position or velocity to calculate the offset (correction). When the THC is enabled and the torch is not commanded on the change in Z position from the last period is used to see if the torch is moving in the correct direction to correct the offset. Then depending on the sign of the error and the offset the offset is either set to 0 or adjusted in the correct direction.

While this may not be quite as sophisticated as the PID servo correction advertised by the commercial offerings there are many here who claim it works well for them. I intend to try and use the stock THC.comp first to test this out. I think that by implementing the corner hold (called Velocity Anti Dive by one leading commercial offering) the way it does this thc makes up for its simplicity by the short correction (servo) period used.

John