Re: Shielding Crosstalk EM Issues in wiring
Posted by
Jim Buslepp
on 2003-05-28 16:07:23 UTC
Dave,
The currents on the stepper windings are large and fast switching. This
makes them large EMI aggressors. The home/limit sensors and other digital
or control signals are typically low voltage and high impedance, making
them very susceptible to EMI. Therefore, I would not run motor windings in
the same bundle as control signals if I could avoid it.
More important than shielding may be the pairing of wires. When a current
moves in a wire, it creates electric and magnetic fields based on the
direction and rate of change of the current. This is what creates radiated
EMI, and radio. If the currents in two wires run close together are of
equal magnitudes but in opposite directions, their fields tend to cancel
each other out. If the wires were infinitely small and infinitely close
together, the fields would cancel completely. As a practical matter, the
closest we can get to this ideal is to use a twisted pair.
I would recommend the wiring each stepper motor winding as a twisted pair.
Twisted-shielded pair is better, but is expensive and bulky. If the wires
will be flexed a lot during operation of the machine, don't use foil
shields. They're good for stationary applications but will disintegrate
under repeated flexure. These pairs can be bundled together if desired.
Make sure you're using big enough wire, too. Wire ratings are usually based
on how much current it can carry before it heats up too much. IR drops can
degrade a system long before the wires get hot, so take this into
consideration.
I would run motor windings, power and control signals in separate bundles.
Route these bundles as far apart as it's reasonable to do. Where they
cross, cross at as close to right angles as you can. Where the must come
together, separate them as soon as you can.
I would shield the control signals. The shield should be tied to chassis
ground, not the signal ground. Ideally, the signal ground and chassis
ground should be connected at a single point only, usually at the main
power supply. Unfortunately, if you're using a personal computer in your
system it will tie these two grounds together. Where grounds are properly
separated terminate both ends of the shield. Where grounds are mixed,
terminate only one end to avoid creating ground loops. Keep the wires
connecting the shield to ground as short as possible. The inductance of
these wires can rapidly degrade the effectiveness of the shield.
Absent a fault there should be no current in the shield.
If you have the option to run control signals as differential pairs on
twisted shielded pair wiring, this will further improve their immunity to
EMI. Unfortunately, this is can get expensive.
Shielding the power cables is not usually necessary, but can be cheap
insurance if it doesn't run up cost excessively.
EMI generated faults are notoriously difficult to troubleshoot. I find it's
a lot easier to use these techniques up front and not have to worry about
it later.
Jim Buslepp
The currents on the stepper windings are large and fast switching. This
makes them large EMI aggressors. The home/limit sensors and other digital
or control signals are typically low voltage and high impedance, making
them very susceptible to EMI. Therefore, I would not run motor windings in
the same bundle as control signals if I could avoid it.
More important than shielding may be the pairing of wires. When a current
moves in a wire, it creates electric and magnetic fields based on the
direction and rate of change of the current. This is what creates radiated
EMI, and radio. If the currents in two wires run close together are of
equal magnitudes but in opposite directions, their fields tend to cancel
each other out. If the wires were infinitely small and infinitely close
together, the fields would cancel completely. As a practical matter, the
closest we can get to this ideal is to use a twisted pair.
I would recommend the wiring each stepper motor winding as a twisted pair.
Twisted-shielded pair is better, but is expensive and bulky. If the wires
will be flexed a lot during operation of the machine, don't use foil
shields. They're good for stationary applications but will disintegrate
under repeated flexure. These pairs can be bundled together if desired.
Make sure you're using big enough wire, too. Wire ratings are usually based
on how much current it can carry before it heats up too much. IR drops can
degrade a system long before the wires get hot, so take this into
consideration.
I would run motor windings, power and control signals in separate bundles.
Route these bundles as far apart as it's reasonable to do. Where they
cross, cross at as close to right angles as you can. Where the must come
together, separate them as soon as you can.
I would shield the control signals. The shield should be tied to chassis
ground, not the signal ground. Ideally, the signal ground and chassis
ground should be connected at a single point only, usually at the main
power supply. Unfortunately, if you're using a personal computer in your
system it will tie these two grounds together. Where grounds are properly
separated terminate both ends of the shield. Where grounds are mixed,
terminate only one end to avoid creating ground loops. Keep the wires
connecting the shield to ground as short as possible. The inductance of
these wires can rapidly degrade the effectiveness of the shield.
Absent a fault there should be no current in the shield.
If you have the option to run control signals as differential pairs on
twisted shielded pair wiring, this will further improve their immunity to
EMI. Unfortunately, this is can get expensive.
Shielding the power cables is not usually necessary, but can be cheap
insurance if it doesn't run up cost excessively.
EMI generated faults are notoriously difficult to troubleshoot. I find it's
a lot easier to use these techniques up front and not have to worry about
it later.
Jim Buslepp
> Date: Wed, 28 May 2003 19:22:50 -0000[Non-text portions of this message have been removed]
> From: "turbulatordude" <davemucha@...>
>Subject: wire question
>
>Hi all,
>
>I'm looking at some wire and am wondering if I should get unshielded
>for the steppers (16 condutor) and shielded home/limit/low votage and
>then unshielded for the 110VAC for the routers.
>
>I'm wondering what the best combination might be.
>
>I can get 30 conductor shielded and put the steppers and low voltage
>in one shielded and run 110VAC shielded.
>
>or ..... starting to loose if I keep playing with the
>combinations....
>
>any advice ?
>
>Dave