Crashing, homing macro bugs, G93 bug

[MRob]

Running AXBB in windows 10, very clean install, only opera, mach, and onedrive are installed, with UCCNC and plugins etc to let me control Huanyang spindle motor via USB.

So, I have three issues, I'll list in order of priority:

1. Started getting a "cannot detect motion controller" type error, which effectively crashes me out (after giving me the option to save offsets). Frustrating, as I run long carving programs (3-5 hours) sometimes using expensive woods, and dont have a homing system set up on my A axis (guess I should prioritise that). It is very intermittent, I might do multiple runs fine and without issue, but then it will just randomly crash. It has only started in the last few months, and is definitely a problem with UCCNC, as I also run mach with the AXBB plugin and never had an issue. Tried reinstalling UCCNC, then copying my old macro / setup directories over, but no luck. Also tried switching off wifi, but doesnt help. Might be able to find the error details in event viewer if you like. I wonder if a windows 10 update caused the problem. Actually I thought reinstalling UCCNC had fixed it, as I hadnt seen the error in a week or so, until just now...

2. My macros with G31 probing commands go wrong at least 1 in 3 times. When looking into this I saw that you had previous had issues with feedrates in macros, which you fixed - perhaps the same bug is being activated with G31's for some reason. There are two things that can wrong - either the feedrate is incorrect (e.g. G0 moves being stupid slow), or when logging the position after the G31 move, it returns a 0. Other macros that dont use G31 commands, work fine, never had an issue. Sometimes I'll have to repeat the homing procedure 3+ times to make it log a correct result, and not move at a stupidly low feedrate which will take like 5 minutes just to return Z to machine 0. Just occurred to me, but I think the feedrate issue is only ever after the G31 command has run, however I've not been watching for this so cannot say with certainty.

3. Mentioned this in another thread some time ago, but going to put it here again anyway. When running in G93 mode, if you try to decrease the feedrate with feedrate override it gets very screwy. Like, crazy fast moves between slow sections. Also, unsure but I think it might be bugged at normal G93 rates, in that sometimes at the start of movements it'll shoot from one A axis position to another. Although this could just be because I am using roll over on 360, and it's moving in the wrong direction or something.

Would love to get some help on the major crashes. I can live with #2, although it is pretty annoying I've set the macros to throw an error when 0 position is given. The third, I'm mentioning again, just because it is clearly bugged, but hey.

[cncdrive]

1.) Just to be correct what you saying is not a crash, a crash would be if the application closing without any notice, e.g. with an "application stopped working" message. That is a lost connection issue which is probably because your computer cannot process the communication fast enough.
You could adjust the communication buffer time length, it is set to a low value 100msec and with the MAch3 plugin it is set to 500msec, because Mach3 doing things with much worse timing, so with MAch3 a 500msec is required to make it stabile while with the UCCNC usually even 50msec is enough, but with slow computers it can be too low.

2.) The mentioned G31 feedrate issue only existed in a development release which never went to official release. It was when we did lots of developments in a row, too many of them and messed up the G31. But in the latest official release and even in later development releases nobody reported such problems.
So, if you using macros to call G31 then make sure that you have a while ismoving in place after it, because it might be the root of the problem that the probing did not stop yet and a new motion command was added in the motion buffer when the endpoint of the probing is yet unknown. This could cause issues.

3.) I don't remember now if we test this yet, so Dezsoe please test it.
And if you can post some example codes and could give some instructions about how to reproduce the issue would be helpful in debugging.

[MRob]

Thanks for the very fast reply, I'll test with a comms buffer pushed up to 500ms and see if that fixes it. And understood on it not being technically a "crash", correction accepted!

With G31 issues, I have while ismoving commands on every move. Dezsoe emailed me though, so I'll mail him back now.