Even the newly coined FCC definitions of "UWB bandwidth" and "Ultrawideband (UWB) transmitter" pose some interesting technical issues. From the R&O, "the UWB bandwidth is the frequency band bounded by the points that are 10 dB below the highest radiated emission." While this is an appropriate definition for more conventional systems, where the spectral density is determined by the modulation type and rate (e.g., direct sequence spread spectrum, FM, etc.); this definition is of dubious utility for many short pulse, or impulse-excited systems.

The bandwidth
For example, in many proposed UWB systems, the bandwidth is determined solely by the electromagnetic characteristics of the antenna, which are often all but well behaved. While the 10-dB bandwidth may indeed exceed hundreds of MHz or GHz, the 3-dB bandwidth (the bandwidth where greater than 90 percent of the total energy is concentrated) may be significantly narrower. This, combined with high pulse-rate operation, can result in the generation of a small number of strong spectral lines containing the vast majority of the UWB signal energy, further exacerbating problems with interference.

The FCC further defines a UWB transmitter as "[a]n intentional radiator that, at any point in time, has a fractional bandwidth equal to or greater than 0.20 or has a UWB bandwidth equal to or greater than 500 MHz, regardless of the fractional bandwidth." The new definition opens the way for a myriad of wideband modulation schemes, many yet to be invented. Yet, curiously, the FCC, in its UWB Notice for Proposed Rule Making stated that "[w]e recognize that other types of modulation could be employed to produce UWB equipment. However, we do not believe that we have sufficient information to propose limits and measurement procedures for such systems." Indeed, no comments on other than short-pulse techniques for UWB were received in the nearly 4-year proceeding (ET Docket 98-153). Has the FCC opened Pandora's box?

However, one of the most curious results from the current UWB R&O relates to peak power limitations. According to the new regulations, UWB peak power can reach 0 dBm or 1 milliwatt EIRP within any 50- MHz bandwidth. For a 500- MHz bandwidth waveform, this translates into a maximum peak EIRP of +20 dBm or 100 milliwatts full bandwidth peak power. (Note that peak power increases as 20 log bandwidth.) However, according to a May 14, 2002, interpretation of existing Part 15 regulations, a representative from the FCC's Office of Engineering and Technology stated that full bandwidth peak power for non-UWB Part 15 devices was -21.25 dBm, or 7.5 microwatts, more than 13,000 times less than for UWB devices under the new regulations. Are UWB devices really allowed to be that much stronger, particularly given all the concerns about interference? What happened?

The problem actually stems from a little understood provision in Part 15 dating back to 1987. In developing what is now 15.35 of the commission's rules, the FCC addressed the issue that power spectral density, or power-per-unit bandwidth, was a "better indication of the interference potential" (FCC NPRM 87-300).

Power spectral density, both peak and average, was measured in a specified resolution bandwidth (at least 1 MHz). Thus, historically, at least, the 21.25-dBm peak power limitation of Part 15 devices (20 dB higher than the average limit) was to be measured in a minimum 1-MHz bandwidth, and not to represent full bandwidth power. With this interpretation, the new UWB rules (for a 500-MHz bandwidth waveform) are actually 12.75 dB lower than previously existing Part 15 limits, consistent with the FCC's stated conservative approach. Academic? Perhaps. But, to date, at least two UWB systems have been stopped in their tracks because of this confusing issue.

Robert Fontana is president of Multispectral Solutions Inc.