forum.cncdrive.com

I just ran into an issue with my machine. Every time that I try to turn the spindle on whether through gcode or manual override, the machine trips out on me, but it happens so fast that I cannot see what limit is being tripped. I thought that we had narrowed it down to a C10S board that went bad, but I replaced it tonight and the exact same issue pops up on me. The only time I can get the spindle to run is if I physically unplug the servo motors. I attached a picture of my control box for reference, not sure that any diagnosis could be made from that, but it hopefully will give you an idea of what I am working with . I know there probably isn't a way anyone can diagnose the issue with this vague description, but I guess until I diagnose the issue more is there a way to see exactly which switch tripped the reset button?

Edit, couldn't attach it, but here is my control box: https://www.amazon.com/photos/shared/dg ... PGRelDwLDH

If the issue happens every time you turning the spindle on then it is very simple to diagnose which limit input or inputs getting the noise and trigger.
You just have to set all limit inputs ports and pins to 0 keeping valid only one of them and then turn the spindle on to see if it triggers or not.
Then go through all limit inputs the same way.

Those chinese VFDs unfortunately does not have basically any mains filtering is probably why the issue happens.
Also check your PE groundings of your power supplies and the VFD and make sure they are star grounded with thick and short wires to the same point, because that will reduce EMI noise.

Also in the latest few development releases of the UCCNC there is an input debounce parameter which filters (ignores) the inputs for the set amount of time, so it acts as a digital noise filter. You can try that out too...

I am going to try that method tomorrow night. I thought we had it licked but apparently were way off.

I do feel like it is the VFD. I didn't have any issues with the spindle until I upgraded the steppers to the servos with a feedback loop.

I couldn't find the debounce parameter setting in the manual. Is it called something else in there?

The debounce parameter is in the development releases only, the latest few ones. It is a relatively new parameter and is not available in the current official release.

Gotcha. I see the thread now. Would you recommend grabbing the latest release or 1.2101, the first that has the debounce feature?

Always use the latest from the development releases.

Just wanted to follow up with this. I tested out individual inputs and nearly all of them tripped out the spindle instantly. I updated the software and upped the debounce to 3ms, and it seems to have fixed the problem. It's really weird. I had been running the setup this way for over a year, and just one day it decided to start acting up. Seems like its working find now though, thanks!

Because you have this system running for a year already without problem, this indicates that something changed.
My bet is that the main electrolit capacitor in the VFD is drying out and it's ESR is now high enough that the EMI noise level the VFD generates is significantly higher than before and that triggers the inputs in the motion controller.
Hopefully your system will still work for some time before that capacitor will dry out enough that the VFD will die due to overvoltages on the DC rail. I mean that is why usually VFDs (IGBTs in them) die after some time.

There was an event of me letting a buddy use their own Gcode and watch the machine where the bit dove through the MDF piece and MDF spoilboard in one pass and cut at about an inch depth at 200IPM where my normal DOC is 1/8". After I hit the E stop resetting everything and then checking his gcode, I had a tough time getting the spindle to turn on then but I managed to get it to finish the project. After that night, it wouldn't turn on at all though. The motors and spindle pushed through it, and the bit was surprisingly still fine, but it was enough force on the machine to break a skateboard bearing on one of the slides. Would overloading the spindle like that help push a capacitor to drain faster? Also, would moving the VFD out of the case of the rest of the circuit boards possibly help reduce this interference, or is the EMI through the wiring?