UI Solves EMI with SUBCALC

By Eugene Kallaur, P.E.
The United Illuminating Co.

"SUBCALC provided United Illumination with a valuable tool..."
- Eugene Kallaur

A customer contacted The United Illuminating Company to request assistance locating the source of a computer interference problem. The office spaces were located on the ground floor of a relatively new (10 year old) building which faced an attractive tree-lined street. These office spaces were therefore considered to be prime real estate, for which the customer paid prime lease prices.

Two office areas were located along a windowed wall overlooking the street, with a personal computer equipped with a VGA color monitor in each office. The monitors on both computers exhibited extreme jitter making them nearly impossible to read. The customer indicated that prior to the installation of these computers, older computers with monochromatic (green on black) monitors had been used in these two locations for several years with no apparent problems. No other computers in the office spaces appeared to be malfunctioning.

A measurement of the magnetic field in the office confirmed that a fairly high level of electromagnetic interference (EMI) was present in the vicinity of the two affected computers. Measurements indicated the presence of a magnetic field of approximately 160 mG at ground level directly under the computers The field decreased to approximately 80 mG on the desktop adjacent to the monitors (with the computers turned off), but was still far in excess of the maximum acceptable level for a typical color monitor of approximately 20 mG. Older monochromatic monitors could function in a higher magnetic field, and were therefore relatively immune to EMI interference at these levels. Measurements were made on the sidewalk against the wall immediately outside of the building. The highest reading of approximately 385 mG was in a location directly adjacent to one of the affected computers.

An inspection of the immediate area revealed that the EMI appeared to be originating from an underground transformer vault which supplied electrical service to the building. The vault was located under the sidewalk and abutted the building directly in front of the customer’s office. Physical measurements of the vault indicated that the two malfunctioning computers were located almost directly over the transformer bank. Since the transformer vault configuration had never been changed, it was assumed that the magnitude of the magnetic field in the office had been virtually the same since the building was occupied. Apparently, these extremely high field levels were not noticed prior to this time because the older computers used monochromatic monitors.

The transformer bank serving the building consisted of three single-phase 333 kVA poletop distribution transformers (13.8 kV / 208-120 V) connected delta-grounded wye. An inspection of the vault revealed that the low voltage (LV) lead configuration was the most probable cause of the EMI interference. The leads (12-400 kcmil per phase) left the LV bushings of each transformer and entered a cable raceway suspended from the ceiling of the vault, eventually entering the building through RGS conduits. As long as they were in the raceway, the leads were grouped closely together and effected a significant magnetic field cancellation. However, the leads were as much as 10 feet apart in the area between the LV bushings and the raceway, which created large magnetic loops. These observations were substantiated by measurement and calculation.

It was determined that replacing the three single-phase transformers with a single three-phase transformer could cause a significant reduction in the magnetic field, since the LV leads could then be grouped very closely together. However, this option was very costly and time consuming, since the vault shaft would have to be rebuilt to accommodate a larger transformer. Construction time was a critical factor in choosing the ultimate solution since the office building was fully occupied by tenants who could not carry on their business operations without electricity. This required that UI choose a solution that could be implemented within a 24 hour outage on a weekend. Since a "second chance" was not an option due to cost and inconvenience to the customer, UI required self assurances that any modifications made would reduce the magnetic field to levels within the acceptable range for the affected equipment. In this case, it meant having to reduce by a factor of 20 or more.

SUBCALC was used to model the existing vault configuration to establish a base case reflecting the conditions at the time measurements were made. The field calculations were compared with the results of previous magnetic field and building load measurements, and matched the actual conditions very closely. Several modification alternatives were studied and categorized by the degree of reduction and estimated cost. The configuration which was ultimately chosen for implementation showed a reduction of over 10:1. However, it was believed that a more significant reduction could be achieved, since SUBCALC did not have the capability of modeling the shielding effects of RGS conduit.

The building was de-energized early on a Sunday morning to perform the modifications. Approximately 22 hours later, work was completed and the building re-energized. At the location where the magnetic field had been 385 mG, the field was reduced to 14 mG.

SUBCALC provided UI with a va1uable tool to (1) verify that the vault configuration could be modified without replacing the transformer bank with a three-phase transformer, and (2) minimize the duration of the power outage required to implement the modifications. Although this is not the first use of SUBCALC for UI, it is the first time that the program has been used to model an underground transformer vault. UI was very pleased with the results and, more importantly, the customer is satisfied.


Enertech Home Page | EMFW Main Page | EMFW Newsletter