What wires should be shielded?

[eat]

What wires should be shielded?
the power wires, the encoder wires. All of them
I know there is different shielding
but for the most part I would like to know what
wires should be shielded and if some wires do not need shielding.
The main ones in my question are the power to the motors
The encoder wires the power to the spindle
And if the spindle is AC does that make a difference
And if the encoder wires are shielded do the power wires still need to be
And the limit switches should they be shielded
Thanks

[NC Cams]

The encoder wires PICK up induced signals, especially from high power wires adjacent that encounter rapid switching - I'd definitely shield them.

The POWER wires give off magnetic flux which can induce currents/noise in adjacent wires, again, shield them.

Limit switches don't seem to be as susceptible of stray induced voltages - shielding is probably optional.

My Bridgeport has the encoder and the high power wires BOTH shielded. The limit switch wires don't seem to be shielded.

[eat]

great thank you
thats about what I thought but its nice
to hear it from some one elce
you know how it is
if you just assume somthing you could be wrong

[gar]

070721-0617 EST USA

NC and eat:

A clarification.

An induced voltage from a varying magnetic field at low frequencies will not be reduced by a non-magnetic electrostatic shield. Virtually all so called shielded cable is shielded with copper or aluminum, which are non-magnetic materials. These shielded cables provide an electrostatic shield, but not a magnetic shield.

An electrostatic shield minimizes noise coupling from electric fields. That is changes in the potential difference between conductors.

Twisted wire pairs are used to minimize magnetically coupled noise. In addition one could put a magnetic material around the cable to minimize magnetic fields.

All signal wires should be twisted pairs with electrostatic shielding. There may be medium frequency noise on power conductors that could be capacitively coupled to signal wires, and electrostatic shielding around the power wires won't hurt but is probably un-necessary. Increased distance between power and signal wires is important to reduce both magnetic and capacitive coupling.

Ground path currents thru ground resistance are an important source of noise. Here optical isolation and/or balanced signal lines (RS422 type signals) can greatly reduce noise problems.

.

[nabuhajar]

It will be appreciated if someone gives a standard statement regulates the use of shielded and twisted pair conductors.

[gar]

070722-0619 EST USA

nabuhajar:

Everything is relative.

If noise is coupled into a circuit such as to cause a problem, then you need to take appropriate action.

Both electric and magnetic fields diminish with increasing distance from the source.

The effect of a wide area changing magnetic field can be reduced by twisted pair wires, by a magnetic material shield, or at high frequencies a conductive shield.

The effect of a wide area changing electric field can be reduced by a conductive shield.

A difference in ground voltage between two different points requires isolation in the signal path or a differential signal line with adequate common mode range.

In some cases you can use low pass filtering at the signal destination to reduce noise.

You may be able to reduce coupled noise by filtering at the source.

You can twist and shield everything and maybe not solve a problem. You really need to analyze your circuits.

If you work with 5 V TTL circuits you probably want noise voltages less than 0.5 V at the input to a gate. If you work with a straingage transducer you may want noise less than 1 microvolt at the differential output.

Generally use twisted pairs and electrostatic shielding for signal wires that extend from your control to the machine. Do not ground any signal source at the machine.

If not twisted keep power wires (high current) close together, and far from signal wires. Do not ground power wires at the machine.

Machine frame, control chassis and circuitry common, and electrical safety ground should be tied together with a large flat wire. The electrical safety ground ultimately gets back to a grounding system (at least one ground rod) at the main breaker box.

.

[NC Cams]

RFI shielding is a VERY complex science as well as an "art". The prior post by GAR are as simple and informative as I"ve seen of late.

I recall seeing a post quite some time ago where another member went into more machine related details on how to go about shielding. For the life of me, I can't find the link anymore. Perhaps another member might post a link or perhaps a site search for "shielding" or "RF shielding" might turn something up.

[pminmo]

Originally Posted by gar 070722-0619 EST USA Machine frame, control chassis and circuitry common, and electrical safety ground should be tied together with a large flat wire. The electrical safety ground ultimately gets back to a grounding system (at least one ground rod) at the main breaker box. .

I take some exception to this. Machine frame and any metal chassis to safety ground is imperitive, no issue there. Cricuit commons is case specific, and when treated improperly can cause havock. It is not something that should be just done without looking at the total system, including the utility system. It is probably the cheif cause of ground loops in systems.

[Al_The_Man]

I have always used the grounding method outlined by GAR and I agree you have to know what the correct procedure is, but just as you do if you decide to use floating commons.
It is virtually impossible to prevent various machine points from coming in contact with earth ground at various points, the important thing is to use something that Siemens call Uni-potential bonding when grounding, as outlined in their excellent article on Machine grounding.
Following this procedure means that, although you may have grounding at different points, the potential between any is as close to zero as possible.
All the many examples of noise problems by router motors etc seen here just about once a week, is caused, IMO by Lack of grounding, rather than ground loops per se.
Al.