Advanced Search

@langdons and unknown
I'm getting tired of this. I just deleted several of your posts that had absolutely nothing to do with the topic.

Yes, the sample file is set for a 10.10.10.10 IP address
if you are using 192.168.1.121, you will get that error
 

Is the networking set up for Mesa?
Is the IP address on the board set same as in the config?

Stick to 2.9.4 for the time being, some issues have arisen, no doubt they will be rectified in a timely manner.

Elaborating on the Lenovo thing, whilst a parallel port in provided, it requires a Lenovo header cable. From what I can gather the cable uses some pin or pins to enable the parallel on what was onced referred to the SuperIO chip. Without this cable the parallel port is neither seem in the bios setup nor the Linux hardware probe. With the cable everything is fine.
Some basically testing on an old parallel port 3rd gen I5 proved it to be a suitable candidate.
But it all comes down to local availability, resellers or what you can find on the street.

The mesa website has a manual for each and every board, so that should give you enough with what you can do with mesa ethernet, a base card is very easily extended via smart serial cards. The access at register level is documented in each manual.
While obtaining mesa hardware outside of the US due to postage, it can almost double the price for items coming Down Under they are solid and reliable and very well supported

As long as network latency isn't impacted theoretically it may work.
The RPi will suffer somewhat if you require the use of a base thread for some input requirement. Personally I think an SPI for simplicity a 7c81, which is simple to replicate using off the shelf Spartan 6 Dev boards, maybe a better option for IO.

Houston !!! i have a problem!!! 

Thank you very much!!!!!

JacobRush wrote:

tagged on the PC ethernet port with subinterfaces (or however linux does it these days its been a minute) and exposing those segments as untagged to the devices.
 

That's what I figured.

If I were to run LinuxCNC on a raspberry pi (whatever variant), I could potentially handle the UI (buttons, maybe some LEDs, the MPG) through the GPIO pins, and move the digital/analog I/O to ethercat devices, which would completely eliminate the Mesa hardware. That would still require ethercat on one VLAN and general purpose ethernet on another.

I might try it each way and see what works.

I have not actually tested it but I would think it would work just fine doing tagged on the PC ethernet port with subinterfaces (or however linux does it these days its been a minute) and exposing those segments as untagged to the devices. I don't think there is a way for the devices to even know its a vlan somewhere when hanging on an untagged port.

You'll need not-overlapping subnets on the vlans if you need IP connectivity over and above just ethernet frames.