Advanced Search

Depends partly on how much data each device needs to transmit/receive, and how often. Theoretically, rs485 can multi-drop a large number of devices. I have used rs485 almost strictly via modbus protocol, but I see you do not mention modbus.

Using modbus rtu (rs485) with the mb2hal component, I was able to operate 1 VFD per converter/adapter. I was reading/writing approximately 16 parameters per VFD several times per second. I could not reliably run 2 VFDs on one converter. It seemed I was pushing the limits on bandwidth. But I had it running a lot slower - I think it was 9600, I needed that for better interference tolerance.

I used the ebyte NA111 ethernet bridge. They seem ok. Maybe a bit sensitive to interference. (I was using them in a large cabinet with one 30hp, one 10hp, two 5hp, and one 3hp 480VAC VFD, so there was plenty interference available and they've been running relatively reliably since about 4 month ago when they started using it.)

I also tried using the ebyte NE2-D11 but was very disappointed with it. It was so sensitive to interference it did not work with the Allen Bradley VFD I tried to interface. Well, it worked, but froze every time, the instant the VFD was started. I want to try it sometime on a sensor, or maybe even multiple sensors, I expect it might be fine for that.

I also have on another LinuxCNC installation a cheap USB to rs232 converter reading a position sensor about 50x per second nearly continuously since about 2016.

mesaflash --device device_name [options] --sserial


mesaflash --help gives more stuff

or you can try

man mesaflash

Weird, the test/setuid in mesaflash is very basic:

    if (seteuid(0) != 0) {
        printf("You need root privileges (or setuid root) to access PCI hardware\n");
        return -1;
    }

No idea how that can not fail for a normal user...

I have three RS485 devices (each operating at 115200 baud) and one BNO08x IMU, which supports UART at 3,000,000 baud, as well as I2C and SPI.Currently, I am using three USB-to-RS485 adapters and one USB-to-UART adapter to connect these devices. This setup is bulky and inconvenient.I am looking for a more integrated solution, such as a single controller or I/O board, that can communicate with all these devices independently and reduce cabling and USB usage.Any recommendations?

zippoffs wrote:

I have most of the electronics mounted to a piece of sheet metal, which is connected to main ground by the green/yellow terminal block by the contactor. Should I not do this?

That should be fine. I mount a lot of stuff to DIN rail like that on top of metal plate, seems to me sheet metal should be about the same. I try to keep low voltage DC circuits completely separate from AC circuits. Sometimes it is inconvenient to do so then I run them together a bit like you did. And then sometimes it just doesn't work then I have to separate them.

It seems as soon as there are some electronics directly connected to a PC's parallel port or PCIe slots a bit like these BOBs work, things get a lot more sensitive. Lately I've been using mostly ethernet connected 'BOBs' Mesa electronics or Beckhoff etherCAT because I rarely have interference troubles with those. (ethernet has very good isolation)

I used a BOB some years ago on a cnc plasma cutting table retrofit, I was happy with it. I probably went to the extreme shielding that BOB. Mounted it inside a die cast aluminum box that cost more than the BOB. All shielded cords to the die cast box, and no AC power into it. You should not need to go to that extreme. Plasma cutters make more interference than most VFDs.

Thank you so much.

Result:
A12 Servo driver alarm has vanished.
I could "ServoOn", "Jog move". without any error in Servo Driver LED.

Buf I got this  message "Failed to get reference clock time: Input / Output error"  many times. (screenshot)

I attached dmesg log.

But you see automata's posts here? Do not be surprised if it still does not work.
Thankfully, automata has kindly posted the solution here.
Please report back how it goes and if there are problems with lcec
I hope automata will guide you to use his solution.
 

Thanks, I will report back as much as I can for a while.  ( I'm borrowing Yaskawa-Sigma-X motors for just a few weeks)

But I'm not familier about EtherCAT master, I'm not sure what to do next, so please guide me, everyone 

And I'd like to summerize procedures what I have done with your great help to get Yaskawa worked with LinuxCNC for those who are not so familier with EtherCAT master in detail like me.

 

A couple years ago I bought two BOBs and never did anything with them. Just now dug them out for a closer look. So I have two different BOBs, both look different from yours. It seems to me, if your BOB had a ground screw that had direct connection to the metal shell of the parallel cable, that would be a prime location to connect the shields of any cables connected to the BOB. That is what I am referring to as frame ground.

PCW wrote:

That's really odd

You are not running as root are you?

is this a normal Linux installation?

What does

uname -a

report?

This is a fresh installation from LinuxCNC website.
But I downloaded  www.linuxcnc.org/iso/linuxcnc_2.9.7-amd64.hybrid.iso instead of provided  www.linuxcnc.org/iso/linuxcnc_2.9.4-amd64.hybrid.iso  by replacing 4 with 7 in the URL.

I am NOT running as root, because other commands require sudo.

That's really odd

You are not running as root are you?

is this a normal Linux installation?

What does

uname -a

report?

Connecting all grounds to the buss bar sounds great.

One way to track down redundant grounds is to disconnect the known grounds then take an ohm reading to find out whether it is still connected to ground via an unexpected route.

Like I said, you can cheat some. I am not real strict on my projects about having no redundant grounds, but it is one of the things I work on if there are interference problems.

In your case, I would suggest concentrating on the shield grounds right at the BOB as that seems to be a pain point what with it acting up as soon as you so much as connect stepper signals and start the vfd.

Please, did you try moving those AC wires that are parallel to the BOB and within inches of it?

thank you for testing. i have been getting a very different result but will go back and see if it was something i was doing wrong.

PCW wrote:

What version of mesaflash do you have?

No warnings without sudo.

I was just working on a mdi command, so it was easy to have a look at this in FlexGui.
Starting with  G0G55X0Y0

 

Then pressing the button with MDI command

 

and it goes to    G0G54X0Y0

 


the subroutine:

o<work_zero> sub
G90 G54 G0 X0 Y0
o<work_zero> endsub [1]
M2

tommylight wrote:

5i25 is not 5I25
Linux is case sensitive.

Here is a piece of code from mesaflash:

Device names are converted to upper case automatically.

Hello all! I'm new here on the forum, I've been searching for a day or so now and can't really seem to figure this out. I've been using the machine with a rpi4, with the 7c81 running the 5abobx2 bitfile setup for a little while now and decided to expand with the 7i73. From what I've read, when using pncconf I should see the 7i73 or at least a way to set pins on the smart serial interface, right? So far, it never seems to show up. I have it connected to the 7c81 via a cat6 cable, and I have a power light on the 7i73. I've set the jumpers on the 7i73 to all the default locations (terminated serial, no lcd, no keypad). when using mesaflash to check the config of the 7c81, I can see the smart serial interface. I feel like there's something very simple I'm missing. I hope this isn't a repeat question, I couldn't find it here. I appreciate any help you can offer

  Module: SSerial
  There are 1 of SSerial in configuration
  Version: 0
  Registers: 6
  BaseAddress: 5B00
  ClockFrequency: 100.000 MHz
  Register Stride: 256 bytes
  Instance Stride: 64 bytes