Instrument and Cable Grounding 1

[KenAlmon]

Where an instrument is supplied by a 24 Vdc, 2 wire cable, and information from the device is sent back over those two wires (as in a 4-20 mA loop for a temperature transmitter, or perhaps a digital "Profibus" signal), is there a code requirement for an equipment grounding conductor? If so, could you point me to the right section in the NEC? Is it voltage driven (24 Vdc OK but 48 Vdc and above not) etc.

My understanding from a power perspective, is that if the device could potentially "short out", then the case could become live (albeit at "only" 24-28 Vdc). If the case was plastic, then you could argue a ground is not required. However, if a junction box is used to make the terminations in, and the box was metal, you could probably argue that the box should be bonded back to the supply.

Complicating the issue is the shield, which, depending on practices, may or may not be present. However, the standard practice when shields are available are to float them at the field connection, which does not allow for bonding the box or instrument back to the supply.

Any NEC/IEEE/IEE/IAS comments appreciated.

Thanks

Ken Almon

 

[electrifire]

NEC Article 250-121 to 250-125, which probably does not apply to the system described. See also Article 504 if in haradous atmosphes. Article 250-5a for grounding of systems less than 50 volts and Article 725 for remote-control and signaling circuits.

Grounding requirements are based on power supply input voltage and va rating of power supply. But in general the power/signal conductors are not bonded until the power supply has over 1000va capacity. Article 725 is probably the one you need to reference if I understand your descritpion correctly and it does not discuss equipment bonding.

The shield is not a bonding conductor and is for shielding from unwanted EMF interference only.

 

[KenAlmon]

Electrifire:

Thanks for your comments. Our application involves an offshore gas platform, that is (obviously) covered by Article 500/505. Some of the instruments being discussed are intrinsically safe, while others are not. Reviewing Honeywell's instruction manual reveals that their standard is to provide an "internal ground terminal" although all of their drawings show only the two wires connected to the "+" and "-" while the ground terminal remains identified but with no wire. They also state that "A ground
terminal can be optionally added to the outside of the electronics housing."

Article 504 points to 250.100, which in turn points to 250.92(B) which says devices are to be bonded. One argument is that on a metal platform where everything is welded to one another, you could argue that the bond could be established locally, by connecting a wire from the instrument to the deck, or any other welded connection of convenience, providing there was no bolted connection in the path to the deck.

To avoid installing an extra equipment grounding conductor in each cable would be a tremendous savings. Hoping to hear from others on this issue...

 

[jbartos]

Suggestion to KenAlmon (Electrical) Nov 6, 2002 marked ///\\Where an instrument is supplied by a 24 Vdc, 2 wire cable, and information from the device is sent back over those two wires (as in a 4-20 mA loop for a temperature transmitter, or perhaps a digital "Profibus" signal), is there a code requirement for an equipment grounding conductor? If so, could you point me to the right section in the NEC? Is it voltage driven (24 Vdc OK but 48 Vdc and above not) etc.
///NFPA 70-2002 NEC 250.4 General Requirements for Grounding and
Bonding. This applies in general. There is no specific low voltage mentioned. The following general requirements identify what grounding and bonding of electrical systems are required to accomplish. The prescriptive methods contained
in Article 250 shall be followed to comply with the performance requirements of this section.
(A) Grounded Systems.
.....
This applies to your metallic apparatus equipment grounding.\\My understanding from a power perspective, is that if the device could potentially "short out", then the case could become live (albeit at "only" 24-28 Vdc). If the case was plastic, then you could argue a ground is not required. However, if a junction box is used to make the terminations in, and the box was metal, you could probably argue that the box should be bonded back to the supply.

Complicating the issue is the shield, which, depending on practices, may or may not be present. However, the standard practice when shields are available are to float them at the field connection, which does not allow for bonding the box or instrument back to the supply.
///This is the standard practice up to 1MHz signal frequency. Above it, the shield is required to be grounded at more locations including the instrument cable shield end.\\
Any NEC/IEEE/IEE/IAS comments appreciated.
///ANSI/IEEE Std 142-1991 Green Book\\

 

[WESTHAM]

The grounding is for equpotential bonding purposes and as such should be tied down to the power system earth. There should be a screen in the instrument cable which requires to be connected at the field end or at the main control panel end, but NOT AT BOTH, if this happens to can produce circulating currents and +ve.

If the circuit is "IS" i suggest you contact the IS Barrier Vendor and confirm what type of barrier is in use for earthing purposes, there are many right answers to this type of problem, but always remember that there should be an IS earth and power system earth arrangement, the former being insulated from the latter and connected to the power system earth at one point only with a lable stating "IS EARTH"

 

[KenAlmon]

jbartos & Westham
Thanks for your interesting comments.
The solution, as I see it, will be to cable to the instruments with shielded twisted pair, floating the screen at the field end, and locally bonding the instrument to ground/earth with an external grounding lug, specified to be on every bought instrument.

Ken Almon