Advanced Search

Soft limit values can be changed through mcodes using the appropriate ini hal pins. For an example see the included 'millturn' sim config:
github.com/LinuxCNC/linuxcnc/tree/master...mach/millturn/mcodes

Apparently the serial is through a special cable on these, not just the normal 2x12 board pins and those cables are are and expensive. sigh.

How much worse is the usb adapter? This is not a critical application really so if it lags some milliseconds I don't think it matters. The most critical thing is being able to respond to an alarm and stop the machine.

 

the error says spidle.1.on does not exist ,

the qtplasmac_comp.hal needs those spindle pins
github.com/LinuxCNC/linuxcnc/blob/master...plasmac_comp.hal#L28

you need to set up those spindle pins to use qtplasmac
github.com/LinuxCNC/linuxcnc/blob/master...c_metric_ini.inc#L46

in gmoccapy, click on the settings button in the lower right:
 

You might have to put in the default code to enter the settings:
 

Once you are on the settings page click on the halshow button in the lower right:
 

Once the halshow tool opens you can expand the 'pins' in the left hand panel by clicking on the + beside it and check for 'pyvcp'
 

Maybe the EL5101 can say when it doesn't have a good reading, or something along those lines. Obviously check cabling.
It seems you get zero Hz as frequency, that's a unique signal, THCAD always give some hundred/thousand of Hz.
Maybe some hal trickery to hold the last good value while frequency is zero. An own component perhaps.

I have prepared and tested an installing procedure for Debian systems.
(thank you freemoore for the collaboration)
Maybe youu could be interested.

Regards

It says MX 23.5 uses Debian Stable (eg Bookworm)
So the approach in the docs should work
linuxcnc.org/docs/stable/html/getting-st...th_preempt_rt_kernel

1. Install MX
2. follow the instructions I linked to above

Old thread but just wanted to add/archive that this (xset -r) fixed jogging irregularities using a standard USB keyboard on a Raspberry PI 5 install with the PREEMPT_RT kernel.

I am gearing up for an MPG build shortly but the difference in function from an x86 machine and .iso install was concerning me that there were underlying issues with the OS.  I doubt that is the case now.
 

If I comment out those 2 lines it starts fine and works as expected but the spindle speed and at speed led do not work.
Where do I find the 'Show hal configuration' tool?
I'm not really very good with linux as the first one I used worked straight out of the box.
Thank you for helping me though, although I'm an old dog it's never too late to learn a new trick  


Back story:-
I had an older pc with built in parallel port that worked great but sadly the pc died and the new one doesn't have it built in so I use a pcie card for the parallel port. I downloaded the latest release thinking it would work the same but sadly this is not the case. 

Looking at the old config folder it seems the spindle speed display was setup in glade and this worked ok, sorry I can't remember which version of linuxcnc was installed on that machine.

Hmm, can you advise how long it drops to zero for? I did a component once that does a moving average of the torch voltage. You might be able to use it to bridge the outage...

Also you could try just using a 0-10 volt Ethercat analog input instead of the encoder input and thcad..

1. check wiring, literally remove each wire and re-seat it properly on the THCAD side and BeckHoff side
2. use a DVM with memory min/max and measure the 5V supply for the THCAD, if it goes under 4.75V change the power supply or add a 100-220Ufarad capacitor at the THCAD terminals and test again
3. set the divider higher or lower depending on how you look at it, if it is 1/32 set it to 1/64 and plot again.

Will do tomorrow and post the results!

Yes, much better! Buy a serial port to RS485 adapter rather than a USB one.

Aussie link for example
core-electronics.com.au/rs-232-to-rs-485...hSu9V9RoC41AQAvD_BwE

Yes, a rack would be easier for sure and save quite a bit of time, debugging and possibly issues.
But would I extend the work area, I still don't want to lose 10cm to the rack.
Also due to exponentially increase of the risk of ballscrew whip, I don't want to go bigger than 2000mm on the screws, which means 1840mm of work area after extending +200mm.

It's just my application, probably for most other people it would be fine to place the rack.

Construction wise, I only shifted the X beam about 5cm from the current symmetric configuration.
It wasn't easy to design but there is room to do the toolchange outside of the working area and still reach the full width with the mill and probe.
One issue I have to solve is the retractable probe is entering the work area, so I will change that later on to a different construction not protruding, with two carriers,

 
 

Would only be an issue if I put something on the table which occupies the whole space, like a huge block of foam.

Though perhaps I should have started at the beginning:
Install the correct Pi image from here:
www.linuxcnc.org/downloads/