Shielded Cable Drain Connection

Status
Not open for further replies.
L

LesliemanLM

New User
Location
Orlando
Occupation
Electrcian
On a multiple twisted pair shielded cables
Should the drain wires connected at both ends one in the field device and Panel side where the Instrument is mounted
 
If I remember correctly (IIRC), one device is designated the 'hub,' and all wires radiating from the hub have their drain/shields connected at the hub. Basically, the hub extends its shield outward, wrapping around all the wires radiating outward.
 
Only time there would be an exception to this rule is when the shield is being used to ground the remote device. This is why it's best to consult the manufacturer's literature before making any generalizations.

-Hal
 
This is very true:
hbiss said:
This is why it's best to consult the manufacturer's literature before making any generalizations.
Click to expand...

Whether or not it creates a ground loop really depends on the devices in question and what their shield terminal is connected to. Think of just connecting the EGC buses or metal chasis between two pieces of equipment. That doesn't hurt anything or cause ground loops. And I recall reading in a certain instrumentation manufacturer's doc that they actually recommended connecting on both ends. The claim was that connecting on one end makes the shield act a little bit like an antenna.
 
The shield (or "drain") should almost always be connected to the EGC bus at the control end ONLY, unless, as noted above, the manufacturer says otherwise.
( The manufacturer of the device, that is.)
 
Last edited:
The general rule I was taught and follow was "at the sensitive end", almost always the receiving end.
 
Just FYI though, with shielded POWER conductors, like those used on VFD outputs, the shield would be grounded at BOTH ends. Different purpose for the shielding. In control circuits the shield is protecting the control wires from outside interference. In POWER conductors the shielding is keeping the noise INSIDE of the conductors. So connectingthe shields at both ends creates a "Faraday cage" around the conductors.
 
Like was said, follow the manufacturer's instructions if provided.
A twisted pair substantially reduces any interference from magnetic coupling when used with differential mode drivers and receivers. A shield reduces capacitive coupling from electric fields but not magnetic coupling. So typically it's best to tie the shield to the "chassis" or enclosure ground on the control end as mentioned above. But don't tie it directly to a grounded conductor or other circuit ground. Think of the shield as an extension of the panel or enclosure. As bwat mentioned the shield can act as an antenna and pick up high frequency interference. For that reason you want to tie the shield directly to an enclosure so the high frequency currents will flow to the outside of the enclosure and not into sensitive circuitry.

I just noticed Jraef's response and I agree. For VFD's the shield needs to serve as a ground return for the very noisy common-mode output current of the VFD. To do that the shield needs to be connected at both ends of the cable.
 
I would bond the shield at the end that got power, that was just my convention. I probably terminated thousands of shields, with 18-2 STP I would use a 1/4" x 1 heat shrink, cut of the shield and then shrink the tube. For end where the shield would be terminated, I used the 1/4 x 1 heat shrink and then 1/16" clear heat shrink over the bare ground. I had a kit and the terminations went fast as there was usually 10-20 in a cabinet.
Often I would see 33 wrapped around the end of the STP, its ok but not very neat looking.
Here is a great publication from AB, its from 1998 but the requirements are still the same
 
Jraef said:
In control circuits the shield is protecting the control wires from outside interference. In POWER conductors the shielding is keeping the noise INSIDE of the conductors.
Click to expand...

I've been thinking about this. Can you explain why it makes sense to connect at both ends for keeping noise in, but not for keeping noise out?

I'm definitely not saying you're wrong. It's just not obvious to me. It seems like the best shielding should be the best shielding; regardless of which direction the EM waves would otherwise go (in or out).
 
If connected at both ends and there is another parallel conductive path, any current that does flow on the other path will try to flow on your shield as well. Say you have 18 AWG shield wire (plus maybe foil shield) in parallel with something that has a 400 amp overcurrent device. Ground fault on the 400 amp circuit may put some relatively high current on your shield. This somewhat ok but does need some consideration in some cases, that 400 amp overcurrent device probably selected to trip in relatively short time - maybe even milliseconds.

But now comes a maintenance guy with a welder - he better know how to limit stray current by placing his work lead near what he is working on or in right conditions he overheats your small shield wire because some of his welding current is carried over it.
 
bwat said:
I've been thinking about this. Can you explain why it makes sense to connect at both ends for keeping noise in, but not for keeping noise out?

I'm definitely not saying you're wrong. It's just not obvious to me. It seems like the best shielding should be the best shielding; regardless of which direction the EM waves would otherwise go (in or out).
Click to expand...

You connect the shield at both ends when you _want_ current to flow on the shield. For example, if you have 'common mode' current flowing on the contained wires, and you want to protect other things from those currents, then you connect the shield at both ends so that 'balance' current returns on the shield.

The specific example of a motor on a VFD is something that I am dealing with right now. The VFD 'switches' the output terminals between the + and - DC rails at a couple of kHz. The motor windings 'see' the voltage difference between the terminals, this is the desired voltage that runs the motor. But at the same time the entire motor winding is jumping around at say 700V peak to peak and 4 kHz relative to the grounded motor frame. This can inject amps of current into grounded metal via capacitive coupling. You want this 'common mode' current to return along the same path as the VFD drive cables back to the VFD.

-Jon
 
winnie said:
You connect the shield at both ends when you _want_ current to flow on the shield. For example, if you have 'common mode' current flowing on the contained wires, and you want to protect other things from those currents, then you connect the shield at both ends so that 'balance' current returns on the shield.

The specific example of a motor on a VFD is something that I am dealing with right now. The VFD 'switches' the output terminals between the + and - DC rails at a couple of kHz. The motor windings 'see' the voltage difference between the terminals, this is the desired voltage that runs the motor. But at the same time the entire motor winding is jumping around at say 700V peak to peak and 4 kHz relative to the grounded motor frame. This can inject amps of current into grounded metal via capacitive coupling. You want this 'common mode' current to return along the same path as the VFD drive cables back to the VFD.

-Jon
Click to expand...
I appreciate the response, but isn't this a different reason than eliminating EMI caused by the conductors? Or maybe I missed your point. This sounds like the reasoning similar to shaft grounding rings which is preventing stray currents through your bearings.
 
winnie said:
You connect the shield at both ends when you _want_ current to flow on the shield. For example, if you have 'common mode' current flowing on the contained wires, and you want to protect other things from those currents, then you connect the shield at both ends so that 'balance' current returns on the shield.

The specific example of a motor on a VFD is something that I am dealing with right now. The VFD 'switches' the output terminals between the + and - DC rails at a couple of kHz. The motor windings 'see' the voltage difference between the terminals, this is the desired voltage that runs the motor. But at the same time the entire motor winding is jumping around at say 700V peak to peak and 4 kHz relative to the grounded motor frame. This can inject amps of current into grounded metal via capacitive coupling. You want this 'common mode' current to return along the same path as the VFD drive cables back to the VFD.

-Jon
Click to expand...
Agree. And if you don't connect the shield at the motor end of the cable, then the common-mode current will find another path forming a much larger "loop antenna" that will efficiently radiate the interference.
bwat said:
I've been thinking about this. Can you explain why it makes sense to connect at both ends for keeping noise in, but not for keeping noise out?

I'm definitely not saying you're wrong. It's just not obvious to me. It seems like the best shielding should be the best shielding; regardless of which direction the EM waves would otherwise go (in or out).
Click to expand...
Yes, reciprocity applies and so the amount of coupling will be the same in either direction (transmitting or receiving interference). But a STP control cable (with balanced signal pair currents) and a VFD cable (with substantial unbalanced high frequency currents) will require different shield connections to minimize interference. A twisted pair with differential signalling already has high rejection to coupled magnetic fields. To attenuate electric field coupling the shield on a STP only needs to be grounded on one side. If the frequency of interference is high enough that the cable is longer than 1/20th of a wavelength then it may become advantageous to ground the STP at both ends (because stray capacitance of the shield is completing a ground loop through the shield anyway).
 
Status
Not open for further replies.
Top