POWER QUALITY Q&A

Our Expert Offers Answers to Frequently Asked Questions About Power Quality

Question:
I've heard a number of different theories on grounding the shield on a signal conductor. What is the best way to avoid noise problems?

Answer:
The best way to approach this subject is to remember that one of the major causes of electrical noise is the electromagnetic field set up when current flows through a conductor. Therefore, when we ground a cable shield at both ends, we create a current path between two points of different potentials. (This will be the case, of course, unless we take extraordinary measures to ensure that both ends of the cable have the same ground potential.)

When the current flows, noise will develop.

The only way to ensure that this does not happen is to isolate one end of the cable shield from ground. This open circuits the shield, thereby eliminating the current flow, and eliminating the source of noise.

Of course, the cable shields should be PVC (or other insulator) covered to avoid contact with ground or other cables. Running a cable with a bare shield is not only dangerous but it will create loops in many places.

By convention, it is usually the source (or supply) end of the cable shield that is grounded. By doing it the same way for all cables, there will be less confusion and less chance that a ground loop is inadvertently produced. It has been pointed out to me by some that it might be more difficult to isolate the ground in the field on some devices or that perhaps the difference in potential may create a hazard in some areas. In these cases, a convention of grounding at the device end should be adopted.

The other issue that must be addressed is wiring terminals in the field. The best choice is to avoid them. However, it is sometimes necessary to break the signal conductors at terminal blocks. In this case, it is imperative that the shield of the cable is continued using its own separate terminal block, isolated from ground and the shields of other cables. Failure to do this will again set up ground loops and the resultant noise problems.

Question:
We have installed a control system in our plant. The manufacturer is recommending a dedicated ground connection to keep the resistance to ground very low. How do I go about doing this?

Answer:
Some manufacturers of sensitive electronic equipment will require a signal ground of less than 1 ohm resistance to ensure that their equipment will function reliably. This is good practice for any of the electronic equipment that is installed in our facilities. We often find that the grounding methods employed within our factories give us ground resistances of up to 10 ohms or more.

To achieve our goal, it is usually suggested that a dedicated ground electrode be installed near the equipment to give this low resistance.

This is fine except for a couple of things. One, one ground electrode is not likely to give such a low reading and two, the Electrical Safety Code requires us to bond the dedicated ground electrode to the main system ground electrodes.

This may seem counterproductive but it is a necessary and logical step. All grounds of all equipment should be able to trace a copper path back to the main system ground grid. It is only in this way that there is a chance for the grounds to be close in potential.

If a dedicated ground electrode is separate, there will be a potential difference between ground systems and a shock potential set up.

The other situation that might occur is that the main ground grid may experience a voltage rise because of a fault or lightning strike. If the dedicated ground does not rise with it, a severe potential difference may exist between the electronics and the system ground. I have seen this cause significant damage to electronic equipment.

Properly connected, any noise from the power system will go to the main system ground and not back up into the electronics.

David is the President of Wintek Engineering. You can forward your questions or comments to him at wintek@wintek-eng.com. ET