Grounding Problem

What type of loop? 4-20Ma, 0-10V........ Do you own a loop calibrator? Are you talking to a PLC or a display of some sort? More info please.

We are having Rosemount Pressure Transmitter(Range -2 mBar to +2 mBar) in the field we are having the Transmitter & the transmitter have its own display which shows correct reading proportional to field pressure but we had taken 4-20 mA through signal cable to PLC around 70 meters from field initially its shields are grounded both side while we faced problem recently we removed the grounding at one end(near transmitter).even though we are having reading fluctuation in PLC.Please give your valuables suggestions

You must ground once, it sounds as though you are not grounded at all. There must be a return path for the signal.

As I cannot "see" your cable arrangements, it would be dangerous to speculate further.....

Are you using shielded cable? Every transducer I've ever used needed a shielded cable, otherwise you can pick up stray noise from other sources, which will cause erratic readings.

If you are using a shielded cable are you grounding the shield at one end or both ends? Based on your description of the problem, it sounds like you are picking up line "noise".

Only one cable should be grounded at both ends between units, all the others at the "more-nearer" ground end only!!!

Often safety ground and logic/signal ground need to be kept fully seperate......

As if more than one cable is grounded at both ends you make ground loops......

Its usually best to cut the ground NOT in the cable, but in the socket(s) in the end equipment, so that "normal" cables can all be connected and you only need one type of cable.....avoids mix ups with cable types.....

If I am not fully clear enough for you, please ask me further for more infos/diagrams.

Signals generally need to run along shielded cables for proper operation......do not change that.....

Good call Andy, grounding both side of cable will case the problem. These are other problems that I have run into when I worked in the test industry. 1. If you build you own interfaces make sure you bring all you internal grounds to a single point in the box the box, then ground all your equipment to enclosure. 2. If you equipment is installed in an enclosure (Bud, Hoffman) make sure each piece of equipment has a good ground to the enclosure. 3. Run separate ground to both system chassis at least 4 awg. 4. Any coil switched devices (relays, solenoids) that are being drive by logic will cause large short burst spiking in the control lines these can be generated into the controls power supply. Make sure you have diode protection across the coil. 4. If you still have problems try installing a ground rod into ground next to high current equipment. 5. I had a problem at one facility where the power coming into the company was generation a 100khz signal on their lines. It's reaching, but also a possibility. Bottom line is noise is an ongoing battle.

Use calibator to input the signal and read the instrument output to see how to fluctuate.