Newbie- General Wiring Advice(Noise Imminent)

[holyjewsus]

Hello again,

I'm about to designing an electronics enclosure and wiring a fairly complex system(to me), and wanted to ask some questions regarding the various components that are still mysteries to me after searching for a bit. Basically I'm trying to figure out where my wires need to be twisted or shielded and where the shields should be attached...

1) signal lines inside the enclosure: we're using a pmdx system and geckodrive 201xs. Should all signal lines inside the enclosure be shielded cable where the shield attaches to ground on the controller side?... and this is done by soldering or screwing down a length of shield?

2)I remember reading about twisted pairs, but I couldn't find what was supposed to be twisted together, and I can't remember!

3) Motor power wires to 5 steppers, what kind of cable to be used here, and should it be shielded? steppers are nema 34 an 23 running at 56 volts. 3 - 6 amps.

4) VFD and 1.5hp spindle 220v 3ph, okay, I know nothing about this, what I've read indicates very noisy, should be in an enclosure(VFD): what kind of wire should we use, part of the cable routing we are following has the motor lines from the VFD going alongside 4 stepper power cables and limit switch lines... will this be a problem?

5) shielded cable for limit switches?

Thanks for the help

[Gootee]

You can't go wrong by tightly twisting all "natural pairs" of wires, i.e. a signal or power wire and its ground-return wire or its opposite polarity wire. And wherever you would want to twist a pair, it would be even better to use shielded twisted pair, with the shield grounded on one end only (and not connected to anything on the other end).

The basic problem is that any conductive loop can either transmit or receive electromagnetic fields, through the air. Any loop that is in the presence of a time-varying EM field will have a corresponding time-varying current induced in its conductors. And any loop that contains a time-varying current will induce an EM field in the air.

Depending on the purpose of the wiring in question, some loops will cause more problems as "receivers" and some will cause more problems as "transmitters". Minimizing the enclosed loop areas will alleviate both cases.

The larger the enclosed loop area, the "better" the loop is at doing receiving or radiating. So you want to minimize all of the enclosed loop areas. Twisting the wires together does that. Note that it can sometimes take a lot of planning (analysis and design) to not let "enclosed loop area" happen.

Also, it helps to keep any small-signal conductors as far away as possible from conductors that carry high level or highly-dynamic currents. And if they must approach each other, they should be orthogonal to each other.

Another significant problem is the sharing of ground-return conductors. All conductors have resistance and inductance. In a ground return, any current will induce a voltage back at the "non-ground" end of the conductor. Time-varying currents are potentially worse since the voltage across the conductor's inductance is proportional to the rate-of-change of the current, rather than to the amplitude of the current as it is for the resistive component of the conductor. So even a small-amplitude current could induce relatively large voltages, if the current were fast-changing.

So if a large dynamic ground-return current shared a length of conductor with, for example, the ground reference for something sensitive, such as an amplifier input's ground reference pin, then the "bouncing" induced ground-conductor voltage would be arithmetically summed with the amplifier's input signal, which would be "a BAD thing".

So run separate ground-return conductors all the way back to the main power supply smoothing capacitors, for each subsystem and circuit, and usually even from within different parts of a single circuit, depending on what types of currents the ground-return conductors will be carrying.

Tom

[dsnellen]

This is a good explanation but I am still in the dark. I have an electrical noise problem I can't sort out. I can't listen to a radio when the electronics are powered up. The RF noise overpowers a FM radio signal. It has always been this way but after a couple of crashes, I want to install limit switches, an E-Stop switch and use the probe.

All of the electronics are housed in a steel container with a plexiglass door. Cat-5 twisted pair is used for the connections. The Cat 5 from the electronics is terminated with a soldered Dim-9 that goes into a plastic breakout box where wiring is divided into that which goes to the X-axis (E-stop, X limit) and that which goes to the Y-axis (Y limit, probe, tool zero). From the plastic breakout box I also use Cat-5 cable (Dim-9) and tried to keep the amount of untwisted wire to a minimum at the limit switches.

It's wired correctly, but the probe will only work (g31) about 10% of the time. Same with the tool zero. I can't run a program with the E-Stop or limit switches activated. They will randomly activate. The problem also seems to be somewhat random as sometimes the probe will work great 3-4 times in a roll and then fail 10 times in a roll.

Any help would be appreciated. Thanks.

Dave

[Al_The_Man]

A couple to start with..

http://www.cnczone.com/forums/phase_...se_issues.html
http://www.ormec.com/Portals/0/files..._Grounding.pdf

Al.

[dsnellen]

Originally Posted by Al_The_Man A couple to start with.. http://www.cnczone.com/forums/phase_...se_issues.html http://www.ormec.com/Portals/0/files..._Grounding.pdf Al.

Thanks Al,
I will review the grounding configuration. As I sit here thinking about it, not sure if the power supply output is grounded to the medal enclosure. It has two wires running to the controller but I don't recall a third wire. Anyway, I will look closely tomorrow for grounding issues so I can ask more specific questions. Thanks for your reply.

[dsnellen]

Al, So much for my recall....

Here is the wiring setup.

110 3 wire into the power supply.

The power supply output is +28v, 18 amps & -V to the drive controllers.

The breakout box is powered by a 2 wire transformer out putting 5V @ 2A
The COM pin on the breakout box is set for +5V output to the drives

From a casual observer point of view, I don't see common ground. Now that I know exactly what I have, I can reread the links & they will make more sense. From your point of view, what should I do to ensure I have a good ground? Thanks for your input.

Dave

[Al_The_Man]

For a start, ideally you want the supply or service ground going to a common star ground point, to this you should make sure that all motors have a ground conductor to their frame, these together with the PC case or P.P. common should also be tied to the common earth star point, as well as any cable shields.
This would be a start, if problems continue then more may be required.
Al.

[dsnellen]

Originally Posted by Al_The_Man For a start, ideally you want the supply or service ground going to a common star ground point, to this you should make sure that all motors have a ground conductor to their frame, these together with the PC case or P.P. common should also be tied to the common earth star point, as well as any cable shields. This would be a start, if problems continue then more may be required. Al.

Thanks Al,
Here is what I understood you to say:

Connect the Power Supply input ground (the 110 V green wire) to a terminal strip (common star point). From this terminal strip, run a wire to the CNC aluminum frame, run a wire to each of the motor frames and for good measure, run a wire to the PC case. I can't figure out what a "P.P. common" would refer to. Only the probe has a shield so that would also be attached to the terminal strip.

Is this what you said?

Thanks
Dave

[Al_The_Man]

P.P. is parallel port, the port common is already at earth ground through the mother board screws that earth ground the internal power supply, if this is a desk top or tower PC.
Al.

[dsnellen]

Thanks for the fast reply. I'll try it and let you know. Dave

[dsnellen]

Al,
Here is what I did. The electronics are housed in a medal container so I ran a wire from the power supply input ground to the medal container. That is now my ground star point. This also grounds all the drive controllers. Also ran ground wires to the machine 8020 frame. Using an ohm meter, I have verified I have a continuous path and all of the motors now have a frame ground. The only thing I forgot was the connection to the PP.

The probing is better. I have a simple program that makes 20 probes at various point around the table. The success rate has increased to about 70% valid probes from less than 10%. I still get random E-stops although at a lesser rate.

What else should I do? What about external factors like a window fan or florescent lights?

Thanks for your help
Dave

[Al_The_Man]

I don't think the fan or F.L. will have any influence.
How is you probe connected? does it go straight into the P.P. or is does it go through a BOB or have a separate supply?
Al.