Advanced Search

Testing on current 2.9.8 with the G0 move commented give me this, which is not quite correct but it's not as wrong as what you got:

 

yes you need to write an XML file now

has anybody build hardware to mesa 7i92 board, how has made all input optocoupler board and output safety wiring to stepper drivers, i think if can build own output safe board to connection port 1 and all input pin optocoupler board connection 2, and were found bit file were have all output at connector 1 and all input have at connector 2, then have logical perfect all in/out pins configured.

how can add my linuxcnc 2.8.4 V whit mini usb wireles keyboard controll select continue or 1mm, 0,1mm not need go to display select this whit mouse, lot easy if can remote controll this scale or continue selection on to table i setting bit zero point.

I don't think many people use a PLC in conjunction with Linuxcnc. There is an Ethercat master we have a driver for so getting IO and the like is pretty easy. Stick to Beckhoff for best support but making a custom XML driver for other IO and drives is pretty simple. Whilst people have mentioned Ladder Logic which is available, the secret magic nobody seems to mention is you can write hal components which when built and installed with a 1 line command : sudo halcompile --install mycomp.comp. Once these are built, they are treated exactly as if they are part of the core program and run in real time.

I do know a guy who has retrofitted a lot of big machines and recently on one big 5 axis machine with 100 tools and multiple pallet changers, the ladder wasn't doing it for him so he learnt a smattering of C and knocked out his working high speed tool changer component over a weekend.  

you have it. Best to build from source. then  Just do a git pull and make again  for the absolute latest I did a video recently on my @MrRodW youube channel

Marcos wrote:

This is similar to how many industrial robotic cells are structured, where a dedicated motion controller handles paths, and a PLC controls the cell logic and operator interaction.

So the question is not whether LinuxCNC can do everything, but how others have structured a clean responsibility boundary between a real PLC and LinuxCNC in production machines.

Any practical examples along those lines would be appreciated.

I think of this: LinuxCNC has capability to be operated by the user, entirely using keyboard+mouse+monitor, at opposite extreme it can have external I/O wired to physical buttons, switches and dials for operator control without any keyboard+mouse+monitor if you wish.

I'm thinking wire connections from your PLC I/O to LinuxCNC I/O to "drive by wire"? Cycle start, Feed hold, etc. maybe also jogging, whatever you need can be brought out on a wire and connected to your PLC.

The Fanuc 6-axis robot bag palletizer I've been reprogramming lately is connected to an Allen Bradley PLC via ethernet/IP. I have never heard the words "LinuxCNC" and "ethernet/IP" in the same sentence, but I've been meaning to explore the possibilities. I have about $1500 usd worth of new Beckhoff ethernet/IP remote I/O hardware on the shelf for nearly a year now, that I plan to connect with LinuxCNC. My big hurdle has been my nil experience with ethernet/IP. Now on this Fanuc I got an idea what ethernet/IP configuration looks like...

spumco wrote:

Admittedly, the Click PLCs are fairly 'low-end', so higher-end PLC's likely have more sophisticated onboard motion control features.

I've always had good results with AutomationDirect's Click PLC's. But I've been disappointed with the reliability of their Productivity 2000 and also their BRX models. Maybe it was just bad luck. I've finally retrofitted most of my higher end AutomationDirect PLC's but those Clicks sure don't seem to make much trouble. Still I don't care for ladder logic.

The very first project I sold containing a Productivity 2000 quit about a week after it was out the door. I tried reprogramming it, also tried updating its firmware, no go. I was able to read out some error logs. Automation tech support told me it is bricked, and that they have been having that trouble with that one firmware version. I have to return it for warranty. They gave me a choice of sending back the bricked plc then after they have it back, they send me a replacement. Or I can buy a new plc and they refund me when they get the broken one back. So I bought the new one and got the refund later because the customer was in a hurry. This was a $1200 USD PLC about 7yr. ago. (CPU + base + PSU + motion I/O) Even the base was bricked, that contained a CPLD on its motherboard.

Marcos wrote:

Just to be clear, in your setup you are using Beckhoff EtherCAT I/O as remote I/O, but without a Beckhoff PLC runtime (e.g. TwinCAT PLC), correct?

 

Yes, I'm using Beckhoff I/O strictly as remote I/O, all my logic exists in the pc running LinuxCNC

Sorry i had no time to watch this, will do very soon.
Thank you.
OK, watched it, left a comment, nicely done and very detailed.

Magnificent or wunderbar in German, danke schon.
For context this is using a touchscreen to move the machine and watch the video from a camera, as MPG with vision.
Anyone have a better tittle, feel free to chime in, i feel this tittle does not do it justice.
Thank you.

Savaal teshekyr (if i recall correctly, been 30 years since i learned that).
One small complaint, turn the camera on it's side, it is hard to watch vertical video on a ultra wide monitor!

What pins are you monitoring?

I've totally rewritten it anyway. I was able to move the state machine into a seperate procedure which has substantially streamlined the code. Now I just have to get it to compile 

A lot of that does not need to be in the component because you are setting the pin externally. But we may need to scale or otherwise calculate data for them in which case they need to pass through the component. Get the basics first....

write all becomes something like 
homing feedback is in the read command

And here is the short video of the build process of that same PC above: