Need Help!- Home/Limit Resets without CNC Moving

[Mr.Chips]

Mach3:
I have configured my X, Y and Z limit and home switches all in one serial configuration. Using shielded wire and the shielding is grounded to the control box right before they connect to the BOB pin connections. The shielding is not grounded at the Limit/Home end of the wires.

The problem is within 30 to 75 seconds I will get a Reset condition without the machine even moving.

I have moved all the power cords (Control Box, PC, and Monitor) at lease 2’ away from the Limit/Home cable. This did not have any effect.

Inside the control box, the AC and DC lines are separated, not running parallel to each other.

What are the next steps to try.

Thanks

[Cartierusm]

First off it's best to know that there may be electrical noise in your system that would be hard to get rid of.

Next would be to diagnose your limit switches, are they cheap, how old are they?

Have you ever had a success full run?

I have a problem when it runs cutting aluminum all the vibration or extra electrical noise from metal to metal contact.

Also GOTO > Config > General Config > and in the upper right hand corner is a box call "debounce interval" change that to 2000, you can go higher if you need until the problem is gone. I forget what this does but it helps with the limit switches tripping and won't hurt anything or change the machine in anyway.

[Mr.Chips]

Originally Posted by Cartierusm First off it's best to know that there may be electrical noise in your system that would be hard to get rid of. Next would be to diagnose your limit switches, are they cheap, how old are they? Have you ever had a success full run? I have a problem when it runs cutting aluminum all the vibration or extra electrical noise from metal to metal contact. Also GOTO > Config > General Config > and in the upper right hand corner is a box call "debounce interval" change that to 2000, you can go higher if you need until the problem is gone. I forget what this does but it helps with the limit switches tripping and won't hurt anything or change the machine in anyway.

I have been running without limit switches and thought I would try and set them up as it is the correct thing to do, just in case.

They are new and an industrial grade, of good quality.

My "debounce" is set to 2000. I am having the reset condition when the machine is sitting idle after 30 to 75 seconds. I didn't think "debounce" would be an issue if the switches were not being contacted. Was trying to take this one step at a time.

Currently my PS is inside the control box with the other componets, most of the systems I have see have them this way.

Could someone give me a few things to try before I give up and just disable them, I would like to have them if at all possible.

Thanks

[dertsap]

i had the same issue , set the debounce higher and the problem should go away

[Mr.Chips]

Originally Posted by dertsap i had the same issue , set the debounce higher and the problem should go away

Changed "Debounce" to 50,000.
After 25 seconds I had a reset.
Watched it sitting idle for 10 minutes, no reset. Went into the house and came back in 25 minutes and it had reset, don't know when.
Ran a deep air cut 8" circle, after 5 minutes I had a reset.

Hmmmm, Should I go higher with the Debounce?

What exactly does Debounce do?

Any more causes of reset?

Thanks

[ger21]

50,000 is way too high to use. Debounce waits before detecting the switch in 40us increments. Setting the debounce to 50,000 will result in a 2 second delay if the machine actually hits the switch.

It seems like I've seen a dozen people with this problem in the last few weeks. You might want to try connecting them one at a time, testing each time you add one. This might help you track down where the problem lies.

[Mr.Chips]

Originally Posted by ger21 50,000 is way too high to use. Debounce waits before detecting the switch in 40us increments. Setting the debounce to 50,000 will result in a 2 second delay if the machine actually hits the switch. It seems like I've seen a dozen people with this problem in the last few weeks. You might want to try connecting them one at a time, testing each time you add one. This might help you track down where the problem lies.

Thanks for the information. I'll change the debounce setting back to 20,000.
Since all of my switches are in one series I'll disconnect the 1st switch and use a jumper to make a connection. Test that and if it doesn't fail leave that jumper on and jump the second, repeat until all 6 switches are tested.

On second thought I am using pin 12 on the BOB, I might first disconect the wire from pin 12 and jumper it to ground, and see if I get a reset, I should not because the jumper wire would be less than 2" long.

If no failure there, I'll unplug the DB9 that goes into the chassis and jumper it there, if there was a failure, it would be some kind of electrical problem within the control box. If there isn't a failure there then I'll start eleminating the limit/home switches.

One thought is that the E-Stop wires which aren't isolated are running in the same cable, might that be a factor???

Thanks

[Mr.Chips]

I was doing some searching about noise and false triggers on the reset and it seems that adding a Ferrite Bead can reduce or eleminate the problem.

I googled this and see the bead installed in two ways, one where the wire simply goes through the bead and another where the wire is run through the bead and looped around and back through it. Is going through the bead twice better?

Should the bead fit tightly on the cable? Several of the beads I have seen have a pretty large hole through them, or does it make any difference?

Thanks

[Jay C]

when I added a probe input to my C10 BOB I started getting spurious triggers which made the probe useless. My solution, I added .1uF capacitor across the +5 and GND. This decouples the AC noise to ground. Not had an issue since. Make sure you have a pull up resistor on your input and let the switches ground when activated. My wires are not shielded but are routed away from the servo lines. Or as in the case of C10 it uses pull-down resistors.

The Debounce is there because physical switches bounce when they make contact. Rather than seeing 1 contact the PC would see 50 or more. By telling the PC to make sure the contact is solid for x number of ms to avoid these multiple "pushes". It is not a fix for noise.

FWIW,
Jay

[Mr.Chips]

Originally Posted by Jay C when I added a probe input to my C10 BOB I started getting spurious triggers which made the probe useless. My solution, I added .1uF capacitor across the +5 and GND. This decouples the AC noise to ground. Not had an issue since. Make sure you have a pull up resistor on your input and let the switches ground when activated. My wires are not shielded but are routed away from the servo lines. Or as in the case of C10 it uses pull-down resistors. The Debounce is there because physical switches bounce when they make contact. Rather than seeing 1 contact the PC would see 50 or more. By telling the PC to make sure the contact is solid for x number of ms to avoid these multiple "pushes". It is not a fix for noise. FWIW, Jay

Jay,

I went out and picked up a couple of ferrite beads and placed one on the lead containing the limit/home, probe, and E Stop wires.

Currently I have all inputs disconnected from the BOB board and installed a jumper between GND and pin 13 (XYZ Limit Home in series) this was to rule out anything on the BOB causing the reset problem. It has ran for an hour without any false triggers.

My next step is to connect all the inputs back up (with the ferrite installed on lead) and let it run for an hour and see if there are any problems. I could not find any information regarding whether to just run the cable through the ferrite just once or loop and go through twice, so I went through twice.

If I have any triggers I'll will go with your suggestion and add a cap between pin 13 and ground. You do mean GND on the BOB board and not chassis ground don't you??? Pin 13 is my XYZ Limit Home in series.

Thanks

[Jay C]

Yes, I mean logic ground not chassis ground (and without getting into semantics they could be at the same potential). The ferrite might help but I think it's more for fast signals (i.e. high frequency like USB, RF).

From my experience it's the "spikes" on the power rails that cause the interference, not outside EMI.

Jay

[Mr.Chips]

Originally Posted by Jay C Yes, I mean logic ground not chassis ground (and without getting into semantics they could be at the same potential). The ferrite might help but I think it's more for fast signals (i.e. high frequency like USB, RF). From my experience it's the "spikes" on the power rails that cause the interference, not outside EMI. Jay

Jay,

I installed the ferrite bead, and it didn't have any effect in eleminating false triggers. Frequently while watching the diagnostic screen I would get a false trigger but the led's wouldn't light. ??? Possibly too fast a spike??

Per your suggestion I then installed a .1uf Cap between the GND and pin 13 (serial input of home/limits) of the BOB. After watching it for 20 minutes I didn't get any false triggers, it is still idling so will monitor it for at least an hour before I start celebrating. But it does look promising.

If this works, I may have to add a .01uf cap to pin 14, for the probe.

Thanks.

Seems the EStop which is on pin 15 and GND, both of which are on the nonisolated side of the board doesn't have any false trigger issues.