Following up on a controller reboot I just reported — between the last <Idle> status report and the FluidNC boot banner there were only 27 ms of silence, no error: or ALARM: line beforehand. WiFi-over-WebSocket is a great convenience but it gives me no visibility into why the board rebooted.
Wondering if anyone is running a long wired USB connection from a stationary computer to the Maslow board, specifically to get more reliable comms and richer logs when things go wrong.
A few specific questions:
Distance. My workshop layout puts my computer ~50 ft from the machine. USB 2.0 spec maxes out at ~5 m (16 ft). For 50 ft you’d need an active extension cable, USB-over-Cat5/Cat6, or fiber-optic USB. Has anyone made this work reliably? What gear?
Does USB serial actually log more than the WebSocket? I’d expect to capture boot-time messages before WiFi is up, ESP32 panic traces / core dumps, and bus-level errors that the WS stream can never show. Is that true on FluidNC, or does it expose roughly the same thing both ways?
Anything I’d be giving up vs WiFi, beyond convenience? E.g. does the Maslow UI work the same over a USB serial connection, or is some functionality WiFi-only?
Goal: when the controller hangs or reboots, I want to see why in the log — not just a 27 ms gap followed by the board coming back up with stale belt positions.
Wondering if anyone is running a long wired USB connection from a stationary
computer to the Maslow board, specifically to get more reliable comms and
richer logs when things go wrong.
yes
A few specific questions:
Distance. My workshop layout puts my computer ~50 ft from the machine.
USB 2.0 spec maxes out at ~5 m (16 ft). For 50 ft you’d need an active
extension cable, USB-over-Cat5/Cat6, or fiber-optic USB. Has anyone made this
work reliably? What gear?
there are active usb extensions that have electronics in them to boost/clean up
the signal. I’ve used 50’ ones for exending webcams. the one I used was from
aliexpress but just look for a 50’ usb2 extension cable that has bumps in the
wire, not just cable and ends.
Does USB serial actually log more than the WebSocket? I’d expect to
capture boot-time messages before WiFi is up, ESP32 panic traces / core dumps,
and bus-level errors that the WS stream can never show. Is that true on
FluidNC, or does it expose roughly the same thing both ways?
yes. much of it is the same, but not all
there are logs at times that the web connection is not running (startup/crashes)
some of this data gets replayed into the web log, but not all
logs about wifi or browser problems/disconnects may only exist in the usb feed
there is other data that is considered too noisy to put in the web log
(esecially if you enable motor debug logs
Anything I’d be giving up vs WiFi, beyond convenience? E.g. does the
Maslow UI work the same over a USB serial connection, or is some functionality
WiFi-only?
this is not in place of wifi, it is in addition to it. The browser UI does not
work over USB (the UI exposed to the USB serial interface is pretty much
complete, but there is no nice UI that knows how to use it
but this means that it doesn’t need to be the same computer. you could have a pi
or something cheap and low power live close to the machine and be capturing the
logs all the time, but not use it for anything else (so no need to have
keyboard/mouse/screen/computer with fan near the sawdust)
Goal: when the controller hangs or reboots, I want to see why in the log —
not just a 27 ms gap followed by the board coming back up with stale belt
positions.
yep, I have a test board (no frame/motors) that I use for a lot of development
that sits with power and usb connected and then I connect to it via wifi