Problem- Limit switch problems

[kennected]

Hello Gentlemen,

I've been paying attention to this forum for about 3 years, got the bug and built one. It's been a long time coming but I'm about ready to run something. X, Y, and Z are done. Xylotex board and moters. You guys have messed me up good with this "disease".
Everything seems to be ready with Mach 3 and just so I don't load some code and tear up 2 years of serious puttering, I'm using all the discipline I have to wait until I have a limit switch circuit wired and proven. There's the current rub. I'm using 18 ga. wire and normally closed micro switches (I've even wired in a NC E-stop in the same circuit) to make the loop and something isn't right. When I press reset in Mach 3, after a second or two the reset says a limit switch has been tripped when nothing has been touched. I've tried to prove the board by putting a short loop of wire on the Xylotex terminals and the reset holds. If I wire it back to my limit switch circuit and doesn't. When I press reset, sometimes it takes a second and sometimes it takes 5, but it always faults as if the circuit has tripped opened. I keep looking and I don't think the fault is in the wiring. It's pretty simple and it looks right.

I can't say I understand all the electronics, but I'm currently configured in Mach 3's Input config with the ++ and -- for each axis "Enabled" and each axis' Home "Disabled", The port number for all axises is number 1, and pin number 10, with "Active low" and "Emulated" boxes disabled with a red "x". Does this sound right?
I'm probably a bit over my head, but this has been a lot of fun and it's about to get funner.

Thanks in advance,
Ken

[oldskimmer]

I recently built a machine with the exact same problem. Any slight variation in voltage caused by wiring interference will trip Mach3. I tried everything, including Mach3 config changes, sheilded wires, grounding whole machine to controller box and anything else. Although it improved, it still kicked out at various times while running the program. I finally solved the problem by adding 4 normally closed reed relays, each controlled by a normally open limit switch. System tests out fine and haven't had miss yet during any program no matter how long it is.

[ger21]

You have noise on the limit switch lines. Try using shielded wire, grounded at one end only. You can increase the Debounce setting in Mach3 to mask the noise issue and possible get it to work as well.

[kennected]

Thanks gents. I'll be giving your suggestions a try and reporting back in to this thread.

Thanks again,
Ken

[Torchhead]

THe xylotex has no isolation on the inputs or outputs. That means the outputs from your switches are directly connected to the high impedance parallel port pins. That also means just a couple of volts of noise on the ground leads will give a false "positive". The motors are connected to the same ground and produce enough swing on the ground to cause all kinds of problems. The "relay" method acts as an isolator so the noise connection is broken. Another way is to use opto isolators. You can use all the shielded wire you want but if the noise is the common ground connection it won't help (and may make things worse).

One of the reasons good BOB's have isolated inputs is to break the noise link. It also helps that the Opto's typically used have poor upper frequency response (at the PWM frequency motors typically use) so reject a lot of that as well. Finally the opto input is really low impedance so it takes a LOT of noise to make one turn on.

Remember what my old EE prof told me many years ago: "Ground is not ground". I thought he was crazy at the time but years of hard lessons have proven him out.

Star grounding of all critical circuits will work to get rid of most noise but unfortunately you may not have access to the lowest impedance point of each circuit. Running wires across (daisy chaining) or tying cable shields on both ends just gives noise more paths to travel. Your PC ground which the parallel port uses for one side can be several volts away from the ground on the Motor DC.

TOM Caudle
www.CandCNC.com

[pminmo]

There is no technical reason you won't be able to get your board to work. You didn't mention if you had added pullup resistors or not. I don't believe Xylotex boards have any pullup resistors. A few simple recommendations:
1. The common side of your switches should be wired directly to the xylotex board gnd terminal on the connector for the unused pins of the db25.
2. You need to add pull up resistors, per the xylotex data sheet. Use 1K pull up resistors.
3. Use NC contacts on the switches.
4. Don't tie the limit wiring up near the stepper motor cables.

If you still have an issue, as ger21 suggested you can use mach3's debounce option.