I’m experiencing some frustrating behaviour and I’m not sure what is causing it.
My Root Controller and RPi terminal are mounted quite close to the VFD which is earthed. Being close the wiring is short. Photo attached. Yes I do need to get a short USB cable.
I’m using a RPi3 as a local terminal connected to the controller using a USB cable. I run both the Fluidnc web console and, if desired, the terminal on the RPi.
If I pause a job using the fluidnc.local pause button, and then stop the job I often lose the RS485 connection with the VFD. This is evident as the spindle cannot be stopped. The only option I have discovered is to remove power from the controller and VFD and restart. This loses the job state of course.
Is there an explanation for this behaviour and/or a way of restoring the RS485 connection?
I’ve had this issue before… although the solution surpasses me right now (its getting late here)…
Ah - I did have this issue with a Siemens V20 VFD drive and I think the issue was down to a RS485 timeout on the VFD drive it self. Does the VFD throw and errors when this happens?
What does the wiring for the Isolated RS485 zone look like?
Here is a photo of mine, just a simple 4 core cable is being used:
My RS485 wiring is a 4 core shielded cable about 30cm long. From the photo you can see the proximity of the controller and VFD. The main difference I see from your RS485 setup is that I’m using the 24Vdc supply from the second teir of connectors rather than the 10Vdc you have used.
I’ve also added a side view of the cabinet. Although the wiring looks messy that was a consequence of running the wires from point to point and leaving them unbundled. As a avionics trained technician I was tempted to create a wiring loom but decided to opt for a solution that would allow me to make changes easily, and possible reduce the liklihood of having problems with crosstalk.
I haven’t seen any errros from the VFD but I believe I have seen responses from FluidNC about connection with the VFD being lost.
I tried setting to debug mode. I’ve been doing some modifications today which has meant I’ve been using the terminal to control the machine over a number of hours with three axis moves at varying rates and occasional use of the spindle with the VFD powered on throughout.
I had about six occasions when FluidTerm reported that the VFD RS485 was unresponsive. I found no common action responsible for losing the RS485 connection. I tried a number of test scenarios but it made no difference. On most occasions I could regain access to the VFD by turning off the power to the VFD, waiting for it to fully shut down and then turning it on again. The controller seems to buffer the commands sent so if I had only realised that the VFD had stopped responding before commands were sent, those commands would be sent when the VFD came back online.
I went through the VFD instruction book but didn’t find anything that could be set to auto reboot the RS485 circuitry.
All very weird. If I have serious problems with the RS485 VFD control, I may switch to analogue control although I did want to add a LED Laser some time in the future.
Well done on the debugging! those posts are very clear and full of information!
Now you’ve mentioned this - this will be causing some issues. I believe the 5V and 24V outputs on these VFDs are a rather small power supply and starts to brown out the RS485 circuity. if you’ve got a multi meter - verify the voltage coming out of the VFD. you might find this is dropping down and causing the circuity to slightly turn off.
IIRC the 10V output has a larger supply and can power the Root controllers Isolated RS485 area.
Try swapping to the 10V output and try to repeat that testing
