Advanced Search

Gotta start with solving that latency issue first. It won't show up when running as it only pops up once...but rest assured the issue is still around.

forum.linuxcnc.org/38-general-linuxcnc-q...ead-and-irq-affinity

Try the first four code chunks in that post (lscpu, then editing the grub file) that takes about 10 min and solved the issue on 3 of my 4 computers. You can dig deeper into that post if that doesn't do it.

If you are not comfortable with command line edit it with visual studios code
(Download it, navigate to the folder, then install with sudo dpkg -i "file name)


That might solve it by itself.

If not PCW is correct. THC is seeing a voltage spike and diving the torch. This can happen for a number of reasons, cutting too slow, in corners, bad pid tune, end of cut, etc.


My machine is set P to 3, I to 1 and D to 0 but yours will be different. On your main screen make sure void anti dive is enabled. Whatever your void anti dive is bump it by 100 increments until it goes away. Watch the screen while it is cutting to see if the void anti dive is active. You don't want it active all the time or THC won't work. About 20-40 percent of the time works well for me but I cut much thinner steel.


Side note what are you cleaning that mill scale with? Looks great.

Are you trying to use the encoder with your mpg thing? post your halfile

for the rest of the mpg, i dont have a good answer.
personally i plan to do something like this with the hal_gpio module for the rpi 
community.carbide3d.com/t/cncjs-handwhee...owto-for-nerds/18726
linuxcnc.org/docs/stable/html/drivers/hal_pi_gpio.html 

Was für ein 'Powerbutton'?
Wo kommt welche 'Fehlermeldung' ?
Was ist das überhaupt für ein Treiber?
Wär vielleicht nicht schlecht wenn du uns etwas mehr Info zu deinem Aufbau geben könntest.
Wir sind keine Hellseher.

I'm not sure that looking at the source code is the most efficient way of getting your spindle to turn. In your current situation it would be more straightforward to figure out how the hal language works because that is where the relevant connections are made to get signals to your parallel port. So I would strongly suggest to focus on your hal signal paths using the 'show hal configuration' tool.

My most immediate question would be:
What is the state of 'parport.0.pin-01-out'.
If that goes true and your spindle does not turn then your problem is likely in the wiring. If it does not go true then you would want to find out why by looking at 'halui.spindle.0.is-on'.

Sorry, about the bad attachment. I'll try again but I think I found a problem with the differential pairs having the polarity reversed. I can't test it now but will swap the wires as soon as I get time.

Unfortunately your schematic did not get attached properly.

Your example hal file shows the wrong index enable being used

All day I tried different options. I'm pretty sure I made this change before I left the shop...it was connected to the correct pin.
 

Yes I know that I was using it between a center punch routine and normal cutting so I could change the amps on the plasma unit. I have since found it is not needed as with the tool I use in SheetCam it fires the torch so short it just makes a dimple in the surface. It is much easier to locate the drill hole in the DXF file then measuring after the part is cut.
 

acourtjester wrote:

I sent an email to Les at SheetCam about that. He has modified the PlasmaC post processor so I could use a delay (M01 code snippet) in it.

M01 would be a pause.
linuxcnc.org/docs/devel/html/gcode/m-code.html#mcode:m0-m1

Thanks a lot ,I'm going to translate the string in the qtdragon.ui.

You may ask the expert also the group manager named 梦是乱的静 ,he/she is also able to communicate with English ,he/she is a very kind guy that has helped me a lot of times using remote desktop control,helping me config linuxcnc and making my servo and stepper motor works,saving me a lot of time .

I sent an email to Les at SheetCam about that. He has modified the PlasmaC post processor so I could use a delay (M01 code snippet) in it. So I hope he will modify it again for the Cut Rule that LinuxCNC will work with. Not sure if that will help with the problem I am having, but it was suggested by a member at Plasma Spider.

I have to go now. I will check in later.

acourtjester wrote:

Tried to use a Cut Rule in SheetCam to drop the THC about 0.250" before end but QTPlasmaC did not like the G-code see attached.

This is telling you that in line 37 of the G-code (or nearby) there is a M998 command but LinuxCNC will only accept M codes up to M199.