Reset Log file?

After years and years of running with no problems, I started having random reset issues.

I've already checked all the usual suspects - electrical noise, loose connections etc. but still can't find the issue.

I was pretty convinced it is electrical noise on the limit switches but now I'm not so sure. What I've done is added some Input triggers to zero the A-axis when the limit input goes active, and another when the E-stop line goes active to zero the B-axis. I just type some BS values in there and in testing, it works. Trip a limit, the A axis zeros and same with the e-stop button. However in practice it's not working.

I just ran something and it reset during the cut, but neither the A nor B axes have zeroed. Now it could be that there's some sort of debounce on the input triggers and that's why they're not being activated. e.g. a glitch on a line causes an e-stop but the glitch is too short to be picked up by the input trigger.

Is there any sort of log file that's created during a run? It would be ideal if it would log the cause of the reset.

I'm currently running V 1.2008 but can switch to something newer so long as one doesn't have to mess with it too much. I've got a ton of jobs to do and can't be messing with a new install right now unless it's drop it replacement.

Any ideas?

UCCNC does not have a debounce or noise filtering utility. The way they see it is that the noise needs to be filtered in the hardware. Some tricks for filtering noise are:

- Make sure the stuff is properly grounded, you may need to check the documentation for noise generating drives, etc…
- You may want to use a good breakout board and possibly optoisolated if needed.
- Make sure the negative of your logic is not common with the negative of main power supply if you have a DC power supply, this video explains how to check that: https://www.youtube.com/watch?v=r_BNxS02nKg
- Use ferrite beads on the power connections on the breakout board and other devices handling the logic signals.

Arturo Duncan
https://cnc4pc.com

Quote:

---

UCCNC does not have a debounce or noise filtering utility.

---

Actually, debounce was added to the latest test releases.

Quote:

---

I was pretty convinced it is electrical noise on the limit switches but now I'm not so sure.

---

Can you disable the switches, or are they also your home switches.

If they are your home switches, home the machine, then disable the switches, then run your program. See if it resets.

I'll give you a call later today.

Hey Ger,

The limit and home switches are the same, I'll have to check if I can disable them. That is slightly scary as I have had things go wrong, lost steps and had the limit switches save my spindle but....

So, the machine has run four hours and 59 min this morning without any resets and I'm kinda thinking it might be a bum wire. This machine has been running a lot (don't know total hours) over the last 5 years and I didn't use flex cable (I don't think I did anyway) and perhaps that's the problem.

Yesterday I moved my ferrite beads around and generally speaking did a more proper winding in there, checked all the wiring, jiggled the handle, gave it a few whacks etc. and it seemed to help for almost 5 hours today. I had to order some flex cable for my other machine and ordered up some flex wire for the limit switches as well so if this thing starts giving me fits again like yesterday, I could potentially re-wire my limit switches but that's really something I'd rather not do today but, I've got a good 15 hours of machine time left right now so I might have to.

For those playing along at home - I'm pretty sure it's a bum wire.

Yesterday, the machine ran for about 10 or so hours grand total and had about 6 or so reset events. One time when I was resetting the machine to continue the cut, the thing reset and when I looked at the sensor which has an LED to show active, I caught it flickering and my input triggers were triggering. Jiggled the wires and it went away. This leads me to believe that either my Y or Z axis limit switch (which are all plumbed to the same input) has a broken wire that's mostly ok but under the right circumstances results in an open circuit to that sensor. What I don't know for sure is whether an open line on one of the wires can lead to an "active" state.

The sensors are Hall Effect (the kind that use external magnets, not like a prox switch) that I picked up so long ago I'm not sure I'll be able to find the spec sheet. I still have too much work in the queue to re-wire it in the next few days so I guess I'll have to just stick it out and deal with it. Luckily the resets don't damage the part too much (lose maybe a couple steps in Z when it happens) and I've been able to pick up where I left off. The Z drop tends to leave a little blemish but it's very shallow.

In case anyone's still wondering - I did trace it down to a short in the wiring. I couldn't actually find the short, but I did find that shorting one of the 3 wires (didn't write down which one unfortunately) to chassis ground did indeed cause the limit input line to go 'active' and light the LED on the sensor. I've replaced all the wires with IGUS flex wire, did a nicer job with ferrite beads on the controller side and haven't had a problem since. Ran about 3 hours of testing with the spindle not running and a couple hours with and so far so good! I was even smart enough to pitch the old wire so I'm not tempted to use it again!