July 17, 2000

Mr. Paul Roosa
Office of Spectrum Management
National Telecommunications and Information Administration
Room 4099 HCHB
1401 Constitution Avenue, N.W.
Washington, D.C. 20230

Re: Docket No. 000623194-0194-01; RIN 0660-XX09; Notice and Request for Comments on Ultrawideband Systems Test Plan

Dear Paul:

On behalf of Zircon Corporation, please accept the following comments in the above-captioned docket concerning the proposed test plan for Ultrawideband (UWB) Systems. Should you have any questions please do hesitate to contact Chuck Heger at Zircon Corporation(408-376-2824) or the undersigned (202-626-6421).

Comments on Master Plan

Page 1, footnote 1

"The UWB signals for the devices of concern in this plan are generated by direct current impulse responses fired into a tuned circuit. This generates a burst of energy of ideally one positive going cycle shaped by the tuned circuit to a specific portion of the spectrum. "

Zircon disagrees with this statement. One basic method of UWB generation is to utilize directly the spectrum of a very fast current edge, creating a broad spectrum consisting of harmonics of the PRF frequency and extending to frequencies on the order of 1/transition time. There is no resonant circuit and no nominal center frequency. Any spectrum shaping is largely created by superimposing the characteristics of the transmitting antenna on the frequency spectrum of the impulse.

Page 2, sec 3:

"The UWB signals probably will be most easily (and, most accurately) measured in the time domain."

Zircon disagrees with this statement. The statement assumes that either a direct wired connection is available from the device under test (DUT) or that a VERY good, low dispersion antenna is available to intercept the emissions of the DUT. The standard procedures that have been in place for years are based on frequency domain measurements. Zircon realizes that this is only an investigation and not intended for the measurement procedures that will ultimately be imposed on the user community, but the above-stipulated conditions still must be meet to ensure success.

Page 2, sec 3:

"Given this, it must be determined how to use the knowledge of the time-domain characteristics in the analyses of interference in the frequency domain. It must also be determined if measuring a UWB signal in the frequency domain is practical and useful."

Again Zircon disagrees. Fourier mathematics controls on this point. Given that the time domain signal can be expressed mathematically, the frequency domain characteristics can be accurately predicted and conversely, any frequency domain measurement can then be converted to the time domain equivalent. Any potential conventional victim receiver 'listens' in the frequency domain, not the time domain.

With regard to the statement that "it must also be determined if measuring a UWB signal in the frequency-domain is practical and useful," Zircon restates that the only practical way to deal with the ultimate requirement for standards compliance is to use the frequency domain. To require the large number of testing labs and services to add time domain capability is to significantly jeopardize the future of UWB.

Page 2, sec 3:

"Furthermore, since resonant structures are used, the pulse shape may be temperature dependent."

Zircon disagrees. The very fact that these are UWB signals implies that there are no high-Q resonant structures involved. Any reasonable Q will produce such significant 'ringing' that a majority of the radiated energy would be at a single frequency and thus violate both the intent and definition of a UWB signal. Any low Q network will have insignificant temperature effects.

Page 2, sec 3:

"Some UWB signals have been shown to produce a spectrum signature that is similar to noise (at least, when measured over some fairly narrow bandwidth)."

To measure noise, particularly pseudo-noise or so-called pink noise, a wide bandwidth must be incorporated, not a narrow one. For example, assume that a PRF were dithered with a pseudo random digital signal of some finite code length. The total signal structure would appear as noise at bandwidths greater than the frequency of the code length. But if a narrow bandwidth were used, the individual spectral lines associated with the repetition length of the code generator could be observed and would appear as single frequency 'bright lines', absolutely the opposite of noise.

Page 5, Task 8, subsection B, 3rd par:

"In the case of one UWB signal getting into one receiver, the UWB signal has a nanosecond or narrower pulse and an emission spectrum of up to several GHz. That very wide emission exists only for the duration of the pulse."

Zircon disagrees, in particular with the last sentence. Any UWB signal structure has spectral emission lines at certain frequencies as determined by the particular system characteristics. (See comments above.) This means that even though a UWB signal is generated by a very short pulse of energy, there is energy at all the spectral lines that as an aggregate, comprise the resultant 'UWB' signal. Assume a PRF of 10MHz and a pulse width of 100ps; there could be a spectral line at 100MHz (the 10th harmonic 0f 10MHz). This energy at 100MHz is a discrete CW signal that is present ALL THE TIME, even though the period of 100MHz is 10ns which is 100 times longer than the 100ps pulse. So, a victim receiver does not "see" a pulse of energy but rather only those spectral lines that can pass the receiver's tuned circuits. If the receiver has a very broadband front end, even though subsequent stages are significantly more narrowband, more aggregate energy from the pulse can get into the front and potentially cause limiting or distortion effects.

To say that the emitted energy is only present during the pulse time is totally wrong.

Task 8, subsection B, 3rd par:

"If the receiver response time is much slower than the pulse's width, the IF circuits could ring for some time after the pulse passes at frequencies related to the receiver IF bandwidths. Thus, the effects in the receiver are as much due to the receiver response as to the interfering signal characteristics and will be different for different receiver characteristics."

The major concern is that there seems to be confusion as to the inherent physics behind the principles of UWB. The use of short pulses in the time domain causes a broad frequency spectrum of conducted or radiated signals. This spectrum will always consist of a series of discrete, continuous wave signals. The frequency at which these lines can occur will be exclusively harmonics of the lowest modulation frequency in the system.

UWB spectra appear similar to noise whenever the relative spacing of the signals is small compared to the bandwidth of the receiving system, so that multiple frequencies can be simultaneously received within that bandwidth. Receiver IF response times relative to impulse times do not in any way determine the response of the victim receiver, while IF response time compared to modulation frequencies can be very important.

Comments on ITS-UWB Measurement Plan

Zircon is concerned that the ITS Measurement Plan may be missing the principal point of UWB as the plan appears to focus on UWB signals that are the result of pulsed RF rather than on UWB signals that are the result of impulse emissions. As just one example of this, Table 1 provides a formula for computing "Total Peak Power" which may be correct for pulsed RF signals but is incorrect by a factor of 2 for impulse (DC pulse) signals. Most UWB applications utilize impulse signals which have significantly different characteristics than pulsed RF signals. It is essential, therefore, that the ITS Measurement Plan use test criteria which are specifically designed to measure impulse UWB emitters accurately.

Respectfully submitted,

Terry G. Mahn 
Counsel for Zircon

TGM/bls

40032178.rtf