I had not powered up my R7 for a couple of weeks or so, then needed it to do an inlay for a box lid. I did the usual setup, loaded the gcode, and hit start. It ran great for 2-3 seconds then the spindle stopped. Everything else kept running so I hit the estop to prevent a broken bit.
Tried again, of course, (without a bit) and got same result.
Typed M3 in command line, spindle started and ran until I sent M5.
Listed the gcode and searched for M3 command., There is only one and it is at proper location. Same for M5.
I checked all visible wiring and found no problems.
Then I repeated all the above several times over the last two days, but the problem persists. If I try a job the spindle starts, gets to speed, then stops, all in about 4 seconds.
Anyone out there ever seen this? I’m thinking grbl or bCNC problems but have not yet opened the controller box (yeah, I’m chicken).
Want to try cutting air with a different sender? CNCjs, UGS, Candle?
How do we do that?
Download it and try using it in place of bCNC as a sender and see if it has the same problem?
OK Michael, thanks. I found all three of those and am going through them to figure out if I want to try that this soon or take (waste?) more time searching for an easy solution.
I did take a wild shot at just removing the M5 command from the gcode but it made no difference. I also changed out the spindle, but again no change.
The fact that it stops even when air cutting but not when you run by hand makes me doubt that your spindle power supply is failing; it seems more likely to be logic-related; either a change affecting the sender (bCNC) or the control board. It’s more likely that an update to your system changed something than that the control board developed such a weird fault. That makes the sender the first place I’d look.
If your R7 shipped with the MKS DLC board, I have been using Candle as the sender for my OX which currently has the MKS DLC board that I got from Brandon. I’m running it on Linux but it supports both Linux and Windows.
I bought the all-in-one controller with my R7 kit and it has a Raspberry Pi mounted inside it where bCNC runs. I do not know what substitute for bCNC is out there that will run on the RPi. I took a quick look at the three you mentioned above and did not think I could manage it with any of them.
I have taken the micro sd card from the RPi (not a simple task) and will reinstall bCNC on it using my other RPi, then put it back and see if that solves my problem. If it does not I’m seriously considering just throwing the R7 in a box and making my 6 yr old Next Wave Piranha my sole CNC once again - - at least I can depend on it being there when I need it.
Thanks for your time Michael.
Seems unlikely that there was a software update either. I didn’t realize this was all on an RPi in an all-in-one; I have run Candle on a laptop with a USB connection to the MKS DLC board.
So that means that it’s not a change in how it is interpreting the gcode. I’m sorry I missed that before. 
For diagnosing this, do you have a multimeter? If you can measure the voltage on the connection from the MKS DLC to the spindle power supply, we can find out whether the problem is on the RPi side or the spindle power supply side. Basically, if the MKS DLC is sending the control signal telling the spindle power supply to turn the spindle but the spindle quits turning, the most likely thing is that the spindle power supply is failing. If it the signal stops unexpectedly, then the problem is on the rPI side.
If you forced me to put money on it, I’d guess the power supply; I’ve seen plenty of reports of failing DC spindle controller supplies. But that’s just a guess.
I have not taken your suggested steps yet but I have everything taken apart so that I will be able to look at them shortly. In the meantime I switched the VFD from machine control to manual control via the potentiometer. I can start the spindle and vary speeds easily using the manual control so I now have a work-around to get some production jobs done.
I have a suspicion that the fact that this works should tell me something important, but have no idea what that is.
What it suggests to me is wiring fault that wasn’t obvious on visual inspection.