FMS Publications

Mitigation of Wiring Interference on Computer Systems

ABSTRACT

An organization which required more operating space chose to extensively remodel an existing building, including completely replacing the electric service. In the process, errors occurred which created elevated magnetic fields, which in turn caused computer monitor disturbances and certain unsafe electrical conditions. Conventional investigation by the various building contractors was unsuccessful in locating or resolving the problems. Detailed investigation by Field Management Services (FMS), an engineering services company specializing in magnetic field issues, promptly identified and resolved the problems.

1. INTRODUCTION

A leading graphic software developer needed larger offices for their software development activities and chose an existing building in an historic area for their needs. The existing building had been a car dealership and garage in its earlier days and required extensive remodeling for this new function. The garage area was reconfigured to a generally open office area with modular offices and several dispersed, enclosed offices to break up the traffic patterns. The former office and car display areas were similarly remodeled into areas of modular offices interspersed around enclosed offices on two levels, to break up noise and traffic flow. The entire building was decorated with bright colors, plants, fountains, artwork and contemporary wall treatments (Figure 1). The entire facility was rewired and a new electric service installed from the underground facilities in front of the building.

As the building became occupied it quickly became apparent that something was wrong. In many of the office areas, the images on the large graphic monitors and many of the smaller monitors were unstable. The electric utility was contacted and provided a survey of the power frequency magnetic fields within the facility. Elevated field values were reported throughout the area, particularly around the electrical room. The building electricians and electrical contractor were unable to cure the problem and FMS was asked to investigate the situation and recommend a solution. The inquiry focused on the office adjacent to the electrical room and the two office areas beyond that adjacent office.

2. INVESTIGATION

FMS conducted a detailed investigation of the facility and confirmed that the fields were elevated in the office adjacent to the electrical room, but, interestingly, that the fields diminished at a very slow rate and consequently extended over a considerable area. Figure 2 presents the initial conditions adjacent to the electrical room. Even in areas where no problems had been noted, the magnetic fields were higher than is typical for a facility of this type. Since experience teaches that this condition (elevated magnetic fields across a wide area of a building) is often the result of wiring errors (wiring conditions which violate the National Electric Code), additional measurements were taken which could specifically confirm the presence or absence of abnormal wiring conditions. These measurements strongly suggested wiring errors.

Magnetic fields from these kinds of wiring problems (net-current fields) are very difficult to resolve. First, the errors must be located and corrected. Then, if necessary, shielding can be effective. Accordingly, the FMS survey report reflected the need to identify and correct the wiring problems present in the building and proposed that this would most likely resolve the majority of the problems present. The report further stated that although wiring corrections would substantially reduce the overall field levels, it was likely that fields in the office adjacent to the electrical room would still exceed the threshold for monitor interference. However, these fields could be further reduced by a custom shielding scheme.

3. CORRECTIONS

Following another period of unsuccessful investigation and effort on the part of the contractors, FMS was requested to assist the electrical contractor in the identification and resolution of the problems. Comparison with values from earlier measurements confirmed that the conditions were essentially unchanged. Measurements in the electrical room suggested a total net-current present of approximately 60 amps, which was distributed among the feed conduits to the five sub-panels.

Measurements throughout the building disclosed two locations in which receptacles were touching the plaster ring, but correction had little noticeable impact on the overall net-current. The existence of elevated fields in portions of the parking lot and along the front sidewalk and street raised the possibility that the source was external to the building.

Although reluctant to allow the service to be turned off (computer file servers in this building service offices throughout the world) a short time slot to test the results of turning the entire service off was agreed upon. Switching the individual sub-panels off in succession at first showed little impact on the overall net-current, but the second, third, fourth and fifth panels yielded large and nearly identical incremental reductions until all of the net-current was removed with the last panel. This test clearly demonstrated that the source of the problem was internal to the building.

The utility granted a request to measure net-current, phase and neutral currents in the pad mount service transformer located in the parking lot (see Figure 3) which was, at first, more confusing than revealing. The net-current value at the transformer was identical to that in the service panel, as expected, but there was little neutral current in either location, even though there was a meaningful phase imbalance. There was also a net-current on the transformer primary conductors very similar to that on the secondary side. This condition was most unusual.

All of the above conditions led to a stop to further efforts to find the source of the problem within the building while the meaning of the high net-currents coming into the area were digested. The utility agreed to cooperate in assessing the situation, as their service conductors were obviously involved. As service panels were being re-covered, it seemed appropriate to check the voltage between the neutral conductors feeding the meter panel and the adjacent distribution section. The result was clear and startling. Rather than the few millivolts that should have been present, there was a voltage of several volts.

The utility was requested to unseal and open the metering section to allow inspection of conditions of the various busses passing through the area. The results of the inspection were dramatic. The bolts on various bus splices between the metering and distribution sections clearly showed evidence of heating, as did the bus splices themselves. Furthermore, an adapter plate for the service cables had two bolts in place (as required) but only one nut. The owner was immediately informed that the source of the problems had been located and that an outage of several hours duration would be needed to tear down, clean, and properly reassemble the main metering panel. It was also recommended that the equipment manufacturer's requirements for this procedure should be determined to maintain warranty coverage.

That which had been confusing was now understandable. Faulty assembly of the panel during initial installation had caused corrosion of the bus splices between the metering and distribution sections, particularly on the neutral splice, until very little of the neutral current was able to pass via this intended path back to the source transformer (see Figure 4). Instead, the neutral current was passing from the neutral bus section in the main distribution section of the service panel to the bond between the service panel case and ground, which was properly connected to the neutral at this service point. The current then was passing via the conduits, water pipes and building steel into the earth at every opportunity. Since, for the circuit to be complete, the current must return to the source transformer, it found sufficient opportunity, through all of these alternative paths, to get back to that location. The best and lowest impedance path for that to occur was via the utility primary cable, which included bare concentric conductors (bare wires spiraled along the outside surface of the insulated conductors), and which paralleled the front of the building within approximately 15 to 20 feet.

At the transformer, the current from the building passed from the grounded primary concentric conductors to the grounded secondary neutral, thus completing the circuit. The net-current measured on the service cables had been a deficiency of current, in that the current provided by the phases was not matched by the neutral, but was instead made up by the return through the ground to the utility concentrics, from there to the transformer ground and finally to the secondary neutral. Thus the net-current which had been measured on the utility primary cables was an excess of current.

4. RESULTS

Following overhaul of the service panel by a representative of the manufacturer, FMS returned to the site to measure the magnetic field levels. As expected, fields in the parking lot and along the front sidewalk were at nominal values. Within the building, the fields in most areas had been reduced to values of 2 milligauss (mG) or less, with the exception of those normally elevated levels in proximity to sub-panels or other high-current electrical equipment. In the office adjacent to the electrical room, the fields on the wall immediately adjacent to the panel were approximately half of their former values and other values were small fractions of the former readings (see Figure 5). In all cases, the fields now diminished at the expected rate for fields from conduits and electrical panels, and thus were inconsequential except in the office immediately adjacent to the electrical room itself.

Since the office near the electrical room had been made into a conference room during the period when computers could not function reasonably in this environment, and since there was no plan to return it to regular office use, it was decided not to proceed with shielding at this time. It would not be difficult or unreasonably expensive to shield this room to allow for full utilization in the future.

SUBSEQUENT ACTIONS (Epilog)

Subsequent to the original investigations and corrections discussed above, the building owner had a further increase in staff and needed to use the room adjacent to the electrical room as an office instead of a conference room. The field levels within much of the office area restricted the area in which a computer monitor could be located free from interference. In view of this situation it was decided to consider shielding for this office.

A simple shield was installed in the office adjacent to the electrical panel and full use of the space was restored. The conditions after installation of the shield are depicted by Figure 6.