4-20mA Isolated Ground?

[Dale Hayes]

Industrial Control Panel with:

Ground for a 480VAC, 3 phase, 30Amp power supply.
Ground for a 480VAC to 120VAC 100VA Control Transformer.
Ground for a 120VAC to 24VDC 15W power supply (this power supply operates a 4-20mA signal loop). This ground is for the 120VAC primary side of the 120VAC to 24VDC 15W power supply.

We have a 120VAC PID Temperature Controller. It has no ground connection. The PID Temperature Controller has a 2-wire, 4-20mA output. We have a (2-wire with shield cable) connected to the 24VDC power supply and a temperature transmitter/thermocouple assembly.

A request has been made to ground the shield on the 4-20mA PID output. We know that the shield should only be grounded at one end. We cannot ground it to the PID controller because it has no ground. The request is that we ground the shield to the same ground used for the 480V, 3-phase, 30Amp power supply, the 480VAC to 120VAC 100VA control transformer and the 120VAC to 24VDC power supply.

Should'nt the shield wire for the 4-20mA signal loop have a separate isolated ground?

 

[tom baker]

Bond the shield at the its power supply end. Typically this is done at the terminal strip with a grounding terminal block.
You don't need an isolated ground. 4-20 mA signals are very immune to noise, due to the twisted pair. The shield just adds a bit more noise isolation.
Tom
ISA Level III Control System Technican (that means I have worked on PLCs and 4-20 for a long time)

 

[mcclary's electrical]

Industrial Control Panel with:

Ground for a 480VAC, 3 phase, 30Amp power supply.
Ground for a 480VAC to 120VAC 100VA Control Transformer.
Ground for a 120VAC to 24VDC 15W power supply (this power supply operates a 4-20mA signal loop). This ground is for the 120VAC primary side of the 120VAC to 24VDC 15W power supply.

We have a 120VAC PID Temperature Controller. It has no ground connection. The PID Temperature Controller has a 2-wire, 4-20mA output. We have a (2-wire with shield cable) connected to the 24VDC power supply and a temperature transmitter/thermocouple assembly.

A request has been made to ground the shield on the 4-20mA PID output. We know that the shield should only be grounded at one end. We cannot ground it to the PID controller because it has no ground. The request is that we ground the shield to the same ground used for the 480V, 3-phase, 30Amp power supply, the 480VAC to 120VAC 100VA control transformer and the 120VAC to 24VDC power supply.

Should'nt the shield wire for the 4-20mA signal loop have a separate isolated ground?

IMO, you would be fine using the groundfrom the cabinet. How could you TRULY ISOLATE this ground? think about it, Where you you get an isolated ground from? Run a ground back to the service? What are you gaining? Unless the cabling is having a false, induced reading, what are you trying to accomplish?

 

[gar]

091223-0740 EST

A shielded wire usually means an electrostatic shield. At low frequencies this has essentially no effect on magnetic fields, just capacitive coupled fields. The twisted pair is the primary defense against magnetic fields.

If the twisted pair is feeding a balanced input receiver, then the twisting to some extent also cancels capacitive coupling.

Consider a single signal line with a shield. The purpose of the shield is to prevent electric fields from coupling to the signal wire. Note: there is capacitance from the shield to the signal wire. If the shield is connected to a large voltage source relative to the reference point for the signal wire, then that voltage would be coupled to the signal wire.

View this problem from the viewpoint of the input of a single ended input amplifier. For example a vacuum tube amplifier with its cathode connected to a common reference, a copper bar in the amplifier box, and also connected to the chassis of the amplifier (the box). The voltage applied to the grid of this tube, relative to its cathode, is what is amplified. Therefore the shield should be connected to the same place as the cathode. Basically the common copper bar. Basically the chassis is OK. Usually this common will in turn be connected to the EGC.

Now suppose you connect the shield to the EGC at the main panel. Any noise currents flowing thru the EGC now become a noise signal coupled to the signal line via the shield.

Without my continuation of this discussion where would you connect the shield in your application to couple the least noise into the amplifier, and why?

.

 

[ELA]

Good points Gar,

Especially the idea that sometimes connecting the shield to EGC can increase noise issues! Sometimes you are better off not connecting the shield at all. Often people with little real world experience claim that connecting the shield at the power source and at one end only is the best. This usually stems from their experience/study with 60hz only and the desire to prevent ground loops.


In fact the absolute best shield is a 360 degree shield. This being each piece of interconnected equipment in a shielded box and the shielded cable between the two uses 360 degree metalic connections at each box.

In this case there is no need for a ground whatsoever.

When using shielded cable that is anything short of the full 360 degree "dumbell", terminating the shield becomes somewhat of a black art.
How you connect the shield depends upon the offending frequency.