From the files of a power quality detective, this case examines how a warehouse endured continuous problems with high-pressure sodium lights.

As the man walked up to me, I knew he wanted something. He had that look of desperation I've seen so many times in the power quality business. It comes from living with a PQ problem you just can't quite put your finger on. I've worn this face myself on a few occasions.

After completing a power quality seminar in Boston, I looked forward to getting home. As I packed my materials, a man from the audience asked to speak with me. I found out Jim was the plant manager for a very prestigious manufacturer of medical implants. He signed up for my seminar in hopes of finding a solution to a devastating problem he had with high-pressure sodium lights in his warehouse. For no apparent reason, the lights flickered most of the time. On several occasions, his department turned off the lights and had to relamp. These two to three minutes of darkness were a safety problem as well as a production killer. I agreed to come up a day earlier on my next trip and see if I could help solve the case of the flickering warehouse.

As I drove up the beautifully landscaped property toward the single-story building, I could see the owners spared no expense in the design and architecture of this facility. Hard to believe something this cutting edge suffered from power problems.

During the interview process, Jim impressed me. He really knew his stuff. From the surface, it looked like everything in the warehouse was designed to be low maintenance and high reliability. This meant two feeders to the property, double-ended stations, and standby generators - what a set up! In fact, it looked too good to me. That's when I got the feeling this facility was actually over-designed.

When I asked about the grounding system, Jim got excited and explained how careful he and his staff were to put a ground ring around the entire building. They even used exothermic connections to connect the ground ring to the building frame.

As we walked through the facility, Jim showed me the state-of-the-art manufacturing equipment. The CNC area had about 30 machines - all powered from overhead bus duct. This area was on the way to the warehouse, so I asked if they had ever experienced problems with the CNC machines. "Just the normal occurrences," Jim replied. "You know, memory loss of programs, a board here or there - it's expected with these machines."

Once we got to the warehouse, I immediately saw the flickering lights. Although I thought it might be a loose connection, I knew it wasn't. As I walked around, I kept thinking about those big CNC machines we had passed on our way in. When I asked Jim about the machines again, he told me he had called the manufacturers in for routine changes. They didn't think there was any relation to the problem. I still had a hunch, so I took out my handheld ammeter and asked Jim to have one of the mechanics open up one of the cabinets on one of the CNC machines. Surprisingly, I found three ground wires in the cabinet: one to the source; one to a building support; and one to a ground rod at the machine. Jim said this was the manufacturer's suggested installation method to keep noise out of the machines.

As I took current readings on each of the ground wires, I was surprised to read 4A on the one attached to the building support. I asked Jim to go to the warehouse and call the mechanic on the radio when he got there. Then, I took the mechanic's side cutters, and on Jim's signal, cut the ground wire. He then radioed back and said he had noticed the lights flicker right after I made the cut.

By 3:00 p.m., we had cut all of the frame ground wires. Based on our work, the flickering had stopped. You see, the harmonics were originally traveling through the bus duct, through the lighting circuits, and to ground at some other location. The machines were over-grounded and had created ground loops of a sort.

When troubleshooting PQ problems, it's always helpful to use the pyramid approach. Look for the most common type of problem first (i.e., those that are ground related) and work your way up.

If the building ground is engineered and installed correctly from the beginning, this minimizes many other potential problems. Industry experts estimate that designing and installing a good grounding system can eliminate up to 50% or more of power quality problems. That's why you should always start from the ground up when troubleshooting.