Robert Jördens

Doesn't look like an obvious networking issue, more like a crash. Are you still using the exact same experiment that you mentioned at the top and ARTIQ versions on device...

That looks a bit like [Jonas' issue](https://github.com/m-labs/artiq/issues/1487#issuecomment-660995196) which seems to be resolved but unknown (to me) how. Could you quickly try hyperterm/putty/whatever from another (without zadig modifications) computer?

Use the device manager to figure out what port number windows assigned to interface 2. Connect to that. There is only logging output if something is happening (booting, running new...

And this is with the correct baudrate/stopbits/parity settings?

Just to check: are you trying to get logging output while the device is in this "crashed" state? Does the kernel actually run? Is there output when the device boots?

With "console output around the issue" (above) I mean a time interval from well before to well after the issue occurs. With "There is only logging output if something is...

But to confirm, you can connect to the console on that other computer now and you see the boot log output?

Do not power-cycle and then reconnect. Just remain connected and power cycle. Do one thing at a time. Let's get some messages when booting first. The log should look as...

I'm not often on mattermost but usually at https://matrix.to/#/#quartiq:matrix.org Let's go there.

Looks vaguely related to #1655. Wrong RPC/Mailbox data scopes?