Need help with electrical noise issue

I finished building my cnc router and all was working well until i installed limit/home switches. i am getting constant false limit switch trips in Mach 3. I played with the debounce setting in Mach 3 and if I set it VERY high, around 40000 it will stop the false trips but adds a delay in the time between an actual switch trip and the time Mach 3 recognizes it. Even if I slow the home speed down to 10%, by the time mach 3 recognizes the home switch trip and stops the movemnet, the axis has physically hit and moved the limit/home switch out of position..

This is my first cnc build and didn't use shielded wire on the stepper motors which is the cause of the problem, or on the limit switch wires which are picking up the noise and sending false trip signals to the breakout board. I'm going to replace both sets of wires with shielded cables but I have a question about the connection between the BOB and the individual driver boards. Should those wires also be shielded?


Here is my setup

Mach 3 usb BOB using separate 5V power supply instead of the usb power, if that makes any difference.
2M542 microstep drivers (4) 1-X, 2-Y, 1-Z, with plans to add 1-4th axis down the road.
48vdc power supply
5vdc power supply
115vac cooling fan
All mounted in an old aluminum MAC computer case.

http://www.cnczone.com/forums/phase_...se_issues.html
Al.

Al,

Thanks for that link. I have a question about the common ground for the power supplies. I have the 115VAC grounded to the enclosure. If I understand what I read in that other post, I should also run my -48VDC and -5VDC to the enclosure?

I usually do if the equipment allows it. Most do.
Al.

Al,

I tied all the power supply grounds together in the enclosure and checked that pins 18-25 of the parallel port all have low resistance to the 115VAC ground plug on the computer side. I tested and there was no change. I have noticed that if I connect a short amount of wire (4-5 feet), I don't get any false trips. If I increase the wire length by a couple of feet, I start getting the false trips. I'm assuming that the wire is acting like an antenna and the longer wire is picking up enough noise to cause the trips. I should have the shielded wire by the weekend and I hope that will fix things. I've read elsewhere that someone recommended putting a .01 uf capacitor on the limit switch inputs. Do you think that would help? I did a little checking and there are hundreds of .01 uf capacitors out there, is there a type that would work best? I would appreciate any other suggestions.

Byron

It is difficult without knowing the total setup.
Personally I prefer not to resort to capacitors, but almost any modern type will be OK for that application.
Al.

You want to use shielded wire where ever practicle! However a lot can be accomplished by careful wiring. For example never run signal wire parallel to and near power wiring. Also make use of twisted pairs where possible. Back in the day I did my own twisted pairs for power wiring with a vise, a pistol grip drill and a bit of working room. It is probably easier to buy twisted pairs these days. I'd suggest searching the net for recommended machine tool wiring practices as I can't cover everything here.

Quote:

---

Originally Posted by b42000

Al,

I tied all the power supply grounds together in the enclosure and checked that pins 18-25 of the parallel port all have low resistance to the 115VAC ground plug on the computer side. I tested and there was no change. I have noticed that if I connect a short amount of wire (4-5 feet), I don't get any false trips. If I increase the wire length by a couple of feet, I start getting the false trips. I'm assuming that the wire is acting like an antenna and the longer wire is picking up enough noise to cause the trips.

---

It may very well be. However there is another bit of ugliness that can pop up where if the electrical length of the wire is right you can get reflections reinforcing themselves.

Quote:

---

I should have the shielded wire by the weekend and I hope that will fix things. I've read elsewhere that someone recommended putting a .01 uf capacitor on the limit switch inputs. Do you think that would help? I did a little checking and there are hundreds of .01 uf capacitors out there, is there a type that would work best? I would appreciate any other suggestions.

Byron

---

Caps can solve problems. It is very hard to say exactly what your problems are though so I wouldn't bother until you have done a nice rewriting of your machine.

Thanks for all the info. I have a new question about the grounds for the limit/home switches. My BOB has a +5v and ground terminal for each axis limit switch. I originally ran the wires from the +5v through the switches and back to the ground terminal. When I rewire the enclosure with shielded wires, I'm planning to run a wire from the ground terminals on the BOB for each axis to a common ground point in the enclosure and then run the return wires from the limit switches to that same ground point? Now I'm wondering if it would be better to ground the frame of the machine to the common ground and connect the ground side of the limit switches directly to the frame where the switches are mounted? I would appreciate any thoughts on which is the best method.

I prefer to use a more accurate description for the -ve side of supplies as 'Common(s)' instead of the current misuse of the term 'Ground' for every common. It can get very misleading.
You should have a star earth ground point set up and any and all commons that need to be grounded connected to this point.
Your PC power supply common is already at ground potential in the PC.
I usually follow a practice of earth grounding all commons.
But ground itself should not be relied on, or used as a circuit conductor.
Al.

Thanks for all the information. I got the shielded cable today and tested it by connecting all 30 feet of wire to one limit switch terminal and connected the other ends together. It's been running about 15 min without a false trip, it would trip every few seconds before so I think this may work. I do have a couple of last questions for when I permanently re-wire the enclosure. I've looked at several sample wiring diagrams and they don't show whether the metal cases of the DC power supplies should be insulated from the metal enclosure which will have a single point common connected to the enclosure, wouldn't that create a separate path to "ground" and possible ground loop? Should I insulate them or is it ok for them to be "grounded" to the enclosure when mounted? Also, I know a single star common connection is preferred, but I'm going to have about 15 common leads to connect. Would a common buss bar mounted to the metal enclosure be a bad idea or should I mount an insulated buss bar to connect all the commons leads to and then connect that bus bar to one point with a single heavier wire? Thanks again for all the help.

I usually set up a copper bus tapped with #10-32 to take ring terminals etc for all the grounds.
This would be mounted to the metal enclosure.
This is a common way it is done in commercial systems.
The service supply ground should also be attached to this point.
Al.