Advanced Search

Konstantin wrote:

Hello,

I have made the decision to try LinuxCNC for my plasma machine that has been running with Mach3 so far.

Thebplan is to use EtherCAT servos for the motion of the machine. I installed LinuxCNC from the official site Downloads section, the latest stable version - 2.9.8. Then followed MrRodW's video about installing the EtherCAT master and CiA402. This is the - Video.

The computer I try to run LinuxCNC EtherCAT is a mini conputer a few years old with an HDD of 1TB.

The problem is that when I reched the step that requires using the halcompile command, it executes without any error but there is no cia402 folder in /linuxcnc/configs. Without having the folder, I cannot continue the procedure for CiA40w installation.

What could be the reason that prevents creation of the cia402 folder?

Best Regards
Konstantin Kolev
 

I just quickly checked the video you mentioned and I think you know have a working Ethercat/LinuxCNC. The next step will be to create your machine configuration: you'll need at least a "chooseAName.ini" file, a "chooseAName.hal" file and a "ethercat-conf.xml" file. Put them in a folder that you create in your /home/linuxcnc/configs/ folder. As 0x2102 said, there is an example for the .hal and .xml files on dbraun1981's github. For the ini file, if it is your first time with LinuxCNC, just start one of the simulation that has similar properties as your machine (3 axis? Gantry?) and have a look in the files that are generated automatically in the "linuxcnc/configs" folder. 

The documentation explains every thing in details and I can only recommend you to read it, but long story short, the ini file will describe your machine, the hal file will load components, create links between them and attach them to threads, and the xml file will describe the element in your cia402/ethercat component.

Hope this helps.
Cheers, Luca

Hi everyone, i have an industrial cnc plasma table made by CR Electronic on wich the numerical contol has stopped working and i would like to try and retrofit it with linux cnc and QTPlasmaC interface..
it has 3 axis for the usual X,Y,Z movements and then an additional axis named "K" that is basically a slave to the X axis. it is a tray connected to the air extraction/filtration system that travels underneath the table to catch all the sparks and fumes made when cutting.
it has servomotors on all axis with analog/digital drives controlled with  +/-10V

i have a few question that i'd like to ask before getting started with the retrofit:

What board should buy?
since it has the analog drives i think the ideal board for this retrofit would be the the Mesa 7i97T  combined with a THCAD2 for the THC control.
the plasma currently has a proprietary board wich handled the thc control and the Z axis motion. also it has ohmic sensing and torch breakaway detection. the voltage signal comes from the plasma generator wich is a Hypertherm HD3070 and outputs direct plasma voltage to the thc control, so i'll have to put a high voltage resistor prior to the THCAD2 board.

Do i need 1 or 2 THCAD2 board?
i've read multiple topics about retrofits on cnc plasma tables on the forum but can't figure out if i need to get 1 or 2 boards to combine with the 7i97T Board. can the 7i97T take the input from the plasma voltage to control the THC? or do i need to use the THCAD2 board an then connect the frequency output to the 7i97T? also since it has ohmic sensing and torch breakaway detection wich are 2 wires that trigger two different things, when the torch breakaway circuit opens it senses that the torch disconnected from it's magnetic base and for the ohmic sensing it had another wire on wich when the torch touched the metal plate it closed a circuit and sent the signal to the old thc board that the metal plate was found and the zero was set. can the torch breakaway wire and the ohmic sensig wire be connected directly to the 7i97T board or would i need another THCAD2 Board to have these 2 functions?

Z axis motor:
the old cnc control THC board controlled the Z axis motor wich is powered by 24VDC and since i will be removing that board i won't have a driver for that motor to use, i wanted to replace it with a new servomotor and i've seen the "stepperonline" 100w 230Vac servomotor with his dedicated analog drive,  would that be compatible to be controlled by the Mesa 7i97T? i'll add a picture and the manual to that. 

How to wire the analog servo drives to the 7i97T?
i've read multiple times the manual for the 7i97T board but i'm struggling a little bit to wrap my head around how i should connect the motor drives to the I/O board. i have all the wiring diagrams of the machines wich are very detailed so i've been able to locate all the encoder wires plus the +/- 10v signal and 0V. i still have to understand how to connect the drives to the board correcty.. it seems to have a circuit on the X and Y drive wich is 24VDC generated by the drives for the drive enable signal. i'll add pictures at the end of this topic.

Are all of the current motor drivers on the machine compatible with the 7i97T?
the X and Y axis use the same model of motor driver wich is a TDE-Macno DMBL7 i'll add the manual for that at the end of the topic.
the K axis has a Axor industries mini magnum  MM230T 6/12 drive, i'll add pictures of that too.
this is the new Z axis motor and driver that i'm looking to get. Motor: A6M40-100L2A1-M17 Driver: A6M40-100L2B1-M17 i'll attach the manual and a few pictures of this too. 

Is a Raspberry PI5 suitable for the application?
i wanted to make sure before buying one that the Raspberry PI5 is suitable for my application or if it would be better to get a normal pc.

Thank you in advance for the help.



 

Once your EtherCAT master is working you will need:

1.) LinuxCNC EtherCAT:
github.com/linuxcnc-ethercat/linuxcnc-ethercat

2.) For CiA402 drives:
github.com/dbraun1981/hal-cia402

Some servo drives have native support in linuxcnc-ethercat.

Take a look at the linuxcnc-ethercat documentation - which also includes example ethercat-config.xml files.

I was suggesting wiring the encoder directly to the 7C81 pins
on a different 7C81 connector than the BOB.

(with series 100 Ohm resistors for protection from negative voltage spikes)

Thanks for your reply. But the cheap Chinese boards require 24v power supply to energize its inputs. So inputs are all not responding to 5v input of the encoder. Kindly help as how to wire the 5v inputs with 7c81.

How can i Check this fault status ?
this way ?

What is the omronmx2 fault status?
 

You should be able to use single ended encoder pins directly with a 5V encoder.
The 5abobx3d configuration has encoders on all connectors and the 5abobx2
configuration has encoders on both connectors.

For more input protection, I  would add a say 100 ohm series resistor
to the encoder signal lines.
 

I would try using the latest mesaflash:

github.com/LinuxCNC/mesaflash

As there were some bugs with the 7I90 name

( build instructions on bottom of page )


 

 


 
 

Hakan wrote:

It can be that the motor speed needs to be ramped up, like it is by linuxcnc.
Setting 6000 rpm as a step is quite challenging.
Anyway, 3000 rpm is more than decent for a stepper motor.

Consider to use csp mode, that's what most people do.
The position feedback is lagging two servo cycles when using ethercat
and a velocity loop has a hard time.

If you use csp, the position update will stop in your scenario, and the motors will stop.
Check out the SDO listing, there is a max current setting (per motor).

 

If I got my math correct, target velocity 500000 = 3000rpm, target velocity 600000 = 3600rpm

>needs to be ramped up...
Yes, another detail I failed to remark on: I noticed when operating in opmode 9 this drive uses a sane accel/decel rate, which kind of surprised me.
I can change target velocity in big jumps and it ramps smoothly to my target velocity. But if I set controlword = 6, the addressed motor stops BAM!
If I set target velocity first, then controlword = 15, it does not move. I changed target velocity to 0 then to 10000 then it started. Didn't check whether I actually had to set target velocity = 0 first.

As far as I can tell, in opmode 8, it behaves as you described. If I set position in any size of small or big jump, it tries to move there instantaneously. Small jumps move the motor, big jumps just make a "chirp".

>If you use csp,
Good idea. I just had to play with my new toy.

rodw wrote:

Its interesting your feedback about motor performance. I di wonder about how I would go not using Lichuan motors. 3000 rpm is up where I was expecting!

 

I know the motor's phase inductance or voltage rating can make a huge difference in top speed attainable. These motors do not have either of those specs listed, making it harder to compare specs, but when I get a chance, I'll try to measure phase inductance.

Higher inductance motors need higher voltage to produce the same speed and torque. This is the first thing that comes to mind, seeing what these motors can do.

Frankly, I'm quite impressed I got this kind of speed out of them. They must have relatively low phase inductance. They're labelled hybrid stepper motors. I'll hazard a guess they're wound to similar specs as a BLDC motor, but in 2 phase configuration. Its been a while since I messed with steppers, the ones I previously used would never have got this speed on 24V, hardly even at 48V.

I was thinking that the best option would be to use an MPG WHB04B? Is the wireless or wired version better? Can everything be controlled from the MPG, including the start/stop spindle?

Hi, 

Please can you share a bit file where i can use 5v encoder inputs on my bob P7 for my 7c81 board. I donot want to use 24v but 5v as a quad encoder fast inputs. Rest pins are fine for a normal 5axis bob3 pins.

thanks

Maybe i almost got it. Now i have this error -> see picture