CIA402 motor switched out from Operational enabled state, why?

[WillyTuring]

Here is how the status word changes of the motor:
"###### slave 0 : 208
Motor 1 is in fault state. Clearing error...

slave 0 : 208

Motor 1 is in fault state. Clearing error...

slave 0 : 208

Motor 1 is in fault state. Clearing error...

slave 0 : 208

Motor 1 is in fault state. Clearing error...

slave 0 : 208

Motor 1 is in fault state. Clearing error...

slave 0 : 208

Motor 1 is in fault state. Clearing error...

slave 0 : 208

Motor 1 is in fault state. Clearing error...
EtherCAT OP

slave 0 : 208

Motor 1 is in fault state. Clearing error...

slave 0 : 231

Motor 1 is ready to switch on. Switching on...

slave 0 : 231

Motor 1 is ready to switch on. Switching on...

slave 0 : 231

Motor 1 is ready to switch on. Switching on...

slave 0 : 233

Motor 1 is ready. Enabling operation...

slave 0 : 233

Motor 1 is ready. Enabling operation...

slave 0 : 233

Motor 1 is ready. Enabling operation...

slave 0 : 237

Motor 1 is enabled.
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
ab7
ab7
ab7
ab7
ab7
ab7
ab7
ab7
ab7
ab7
ab7
ab7
ab7
ab7
208
208
208
208
208
208
208
208
208
208
208
208
208
208
208
208
208
208
208"
The slave is in the OP mode consistently.
We checked on the wireshark capture data, the eabling data in control word and the command torque is shown on the packet in wireshark.
What is the problem here?

[WillyTuring]

PS: i was running the code just fine in Linux machine, the motor works totally fine; but now i am switching to the Windows OS, this problem appears. I wonder why.

[ArthurKetels]

It would help yourself to print the statusword as binary and convert to readable text.
Then 0xab7 would become: Operation Enabled - Warning - Quick stop - Internal limit active
So you have to figure out in the drive documentation what the internal limit active means.

It could well be that the drive detects too much jitter in the PDO timing. Windows is not known for its real time properties.

On the other hand, you should be the one that determines if the timing constraints are acceptable, and not the drive. That is why I am an advocate for the slave principle "be strict in what you send out, be permissive in what you get in". But try to tell this to drive manufacturers.

And the switching to Windows is in general a very bad idea. For real time control it is like going back to the stone age.

[WillyTuring]

Hi Arthur,

The information you gave is very informative and helpful, i indeed wondering the problem might be related to jitter on PDO timing. But when i try to read the 0x092C, the really weird thing i found out is that the time difference on both Linux and Windows are less than ideal, it is way larger than 100us, but still under Linux OS, it works just fine; but not under Windows OS.

Now also you mentioned "And the switching to Windows is in general a very bad idea. For real time control it is like going back to the stone age.", the situation is that we are developing the application on Windows(the UI for the robot control), at the moment we are just remotely launch the EtherCAT master on Linux with ssh then have our Windows UI application talks to the EtherCAT master on Linux via zmq, that is why i want to migrate the EtherCAT master from Linux to Windows, is there any way to do so? Or is this just mission impossible, and remote launching is the only way to go?