While conducting an investigation recently to determine why a piece of equipment in a telecom center was resetting occasionally from a perceived low-voltage condition, I actually was able to review the facility’s electrical drawings. Sometimes, the operations manager has no clue where they are. Other times, they are so marked up that I can’t decipher the latest changes. After the preliminary walk-through, the facility manager found them in an unaltered condition from when the facility was constructed about 10 years prior to my visit. Unfortunately, that is where the good news ends.
The drawings called for a ring ground around the exterior of the building that connected to a halo ground of the building’s metal joists and roof. The ring ground also connected to unused satellite dishes, an antenna tower and the chain link fence around the backup generator area. The drawings indicated that there should be six inspection wells where ground rod connections would be visible. The design called for the ring to be connected to the main ground bus inside the building at the same place where the utility ground connection was to be bonded. This main ground bus also served as the single point connection to the internal grounding system, which included ground wires connecting all of the power distribution equipment (uninterruptible power supply, transformers, static transfer switch, surge suppressor, etc.), as well as to each bay within the 10 rows or racks of equipment and to the metal supports of the raised floor using 3-inch-wide copper ground strips.
In most systems, if the ground potential at any point is raised significantly higher than another point that is interconnected, damaging currents can flow, and communication errors can occur. On paper, these designs seemed like a reasonable way to protect the equipment inside the building from external disturbances, such as lightning, and provide an equipotential ground plane within the building. That way, equipment that is metallically interconnected (such as with RS-232 cables) should operate properly in the presence of ground-induced transients. These designs also can minimize the impact from equipment that shunts current from their power supply circuitry into the grounding conductor. Though this equipment-grounding conductor current should amount to only a few milliamps, some currents exceeded an amp.
A more detailed inspection of the facility determined that any resemblance of how the building and equipment grounding was actually done was purely coincidental to what the drawings showed. Here are just a few of the discrepancies I found:
• Four of the six inspection wells for the ring ground had no ground rods, and the pigtails of the grounding conductor were just sticking up through the gravel.
• I never found the inspection well that the drawing showed would have a wire connecting the ring ground to the internal main grounding bus bar. It is possible that it was under the generator pad concrete, which was a new addition. However, measurements indicated that it wasn’t a solid ground with low impedance.
• I determined the original generator location when I found several pig-tails of the ring ground wire sticking up through the concrete, but they were not connected to any equipment.
• The main service didn’t connect to the utility grounding electrode but rather to the “grounded” or neutral wire from the pole to the main grounding bus bar.
• The power sources and equipment that were having the problem were not grounded back directly to the main grounding bus with a wire like the rest of the equipment in the facility. Instead, the wires went to a small grounding bus bar on the opposite end of the building from the main ground bus bar. From there, a conductor ran under the floor and then up to the building steel at one of the tie points for the joists.
• The bay where the problematic equipment was located didn’t have its own vertical -inch copper bus bar where all the equipment-grounding connections were made in the other bays, but it had wires connected to the painted racks themselves. In some cases, there was some effort made to remove some paint, but it was marginally effective.
The net result of all these discrepancies was, when lightning struck either the antenna, guy wires, overhead distribution wires, antenna dishes or the building roof, or even if the electric field was coupled into the partial external grounding system, the troubled equipment would have its ground potential at a significantly different value than the rest of the equipment, which would be at the internal ground potential. Such voltage differences are sources of temporary disruptions to equipment operation, and if the potential is large enough, damages may result.
The electrical drawings showed a telecom operation that would have been fault-free. However, the owner had not built or maintained the building according to that design, which led to a building riddled with power quality issues.
BINGHAM, a contributing editor for power quality, can be reached at 732.287.3680.