NTIA - UWB Test Plan Comments by Time Domain Corporation

Time Domain Corporation
7057 Old Madison Pike
Huntsville, AL 35806
256-922-9229

July 17, 2000

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

Re: Ultra-Wideband Testing by NTIA

Dear Mr. Roosa:

Time Domain Corporation submits these comments in response to the invitation issued by the National Telecommunications and Information Administration (NTIA), 65 FR 40614 (June 30, 2000). As an initial matter, Time Domain appreciates that both the Office of Spectrum Management of NTIA and NTIA's Institute of Telecommunications Sciences are staffed by highly competent and credible engineers. Time Domain personnel have had the opportunity to meet with and to work with many of these individuals in the course of the proceeding on Time Domain's Limited Waiver before the FCC and in the discourse that led to the FCC's issuance of its Notice of Proposed Rule Making, ET Docket No. 98-153, released May 11, 2000. Given its expertise, Time Domain encourages NTIA to discuss in detail this proposed test plan with FCC personnel. Such discussions could lead to a much more useful final product insofar as the Commission's efforts in the rule making are concerned.

In general the test plan documents made available for comment represent an initial step at describing what will be examined rather than detailing how it will be examined. As such, any interested party faces many unknowns in reviewing the documents entitled Ultra-Wideband Signals for Sensing and Communication: A Master Plan for Developing Measurement Methods, Characterizing the Signals and Estimating Their Effects on Exiting Systems (June 15, 2000) (Master Plan) and ITS Ultra-Wideband Measurement Plan (Master Plan Task 1.2) (ITS Measurement Plan). Moreover, the plans reflect a desire to test against a thermal noise limited environment largely on a conducted basis rather than involving real world testing that recognizes the noise that currently characterizes the allocations in which most of the systems to be studied operate. Thus, the plans make no effort to look at existing Part 15 devices and other systems that generate noise in the relevant spectrum and determine whether, and if so, to what extent, UWB devices would pose any greater difficulties than those now faced.

NTIA Master Plan

The NTIA's effort should focus on determining which limits will allow UWB to operate without causing "harmful interference." Such an approach would allow for the introduction of new technologies as is a stated objective of the Congress. The test plans refer to various unspecified protection criteria against which the NTIA will make its determination. Without knowing what these criteria are, it is simply not possible to state with any confidence that the criteria are appropriate in the face of emissions from UWB devices. A better approach would be to be a parametric analysis that measures the impact of UWB emissions from which a separate analytic process could determine the point at which harmful interference might occur.

Although the bulk of the Master Plan addresses what should be measured with respect to UWB emissions, it places very little attention on the actual process of measuring susceptibility. Moreover, the susceptibility testing is flawed by the lack of any plan for measuring the impact of emissions on performance of protected systems. Rather, the focus is on measuring intermediate signal levels with predetermined criteria as indicators of the impact on performance. The description in the plans for determining impact is to measure the amount of UWB energy reaching the IF stage of the protected receiver. This methodology can be misleading as modern receivers often have significant digital signal processing occurring after the IF stages that allows the receivers to compensate adequately for and continue to operate in the presence of noise. Moreover, the Master Plan is not clear as to how interference criteria will be determined, save to say that the investigators will "obtain or develop interference protection criteria". The devil-in-the-details issue here is that it is unclear if these criteria are based on reality or theory (e.g., GPS interference criteria found in books suggest that Time Domain's UWB signals might interfere at much greater distances than they do in reality).

In short, the goal of this testing should be to determine the susceptibility of critical federal systems to in-band UWB emissions. Spectrum managers cannot make policy decisions on the basis of a method of characterizing UWB emissions; rather they must understand the performance impacts of such emissions. Characterizing UWB signals in the time and frequency domains (and the relationship between the two domains) is a necessary, but not sufficient, step to providing spectrum managers with the required information.

The Master Plan describes a general analysis that is primarily about peak and average power. It is critical that spectrum managers understand that power will be just one factor in determining potential interference to a protected receiver. It hardly reveals the whole story, however, because, as noted above, many modern receivers have forward error correction, digital signal processing and other techniques that eliminate interference if the source has certain time domain properties (e.g. bursting).

With respect to the potential effect from aggregate emitters, the Master Plan calls for testing involving three or fewer as opposed to a more valid number such as ten or more. The rationale for using more emitters in such a case is that it should yield a better fit of any curve to the data produced if there are more points of data.

Finally, the Master Plan places a great burden on a very limited number of engineer-public servants. See the Gantt chart attached to the Master Plan. Although Time Domain recognizes that this may be due to limited resources, the approach can place unrealistic demands on a limited number of individuals so that disruptions for other duties or the vicissitudes of life can have a tremendous adverse effect on the timing of the efforts.

ITS Measurement Plan

In general, the plan as made available for comment has not proposed a susceptibility test for critical government systems. An effort to examine the susceptibility of these systems to various UWB emissions should determine when operations are first affected and then monitor the effect from initial perception all the way to harmful interference. Once such information is gathered, spectrum managers can then make informed decisions on what the protection criteria should be.

The ITS UWB test plan relies on use of UWB simulators that are directly coupled to victim receivers and ancillary measurement equipment. See ITS UWB Test Plan, Section 4.4.1 Closed System Measurements, ("Connect the UWB simulator to the RF input of the receiver being tested.") Such measurements will differ greatly from open air measurements, and will likely show increased interference with much lower power levels than would be found with open-air measurements. As a result, it would not be appropriate to use such measurements in future testing and modeling as representative of radiated UWB systems.

The model for measuring compatibility between the victim receiver and the UWB transmitter uses the noise floor of the victim receiver to determine the distance at which the two systems can coexist with no interference. See ITS UWB Test Plan, Section 4.4.1 Closed System Measurements, ("P_r is an average UWB signal level at the receiver RF input which corresponds to a peak I/N of 0 dB at the receiver IF input.") Given the low sensitivity levels with which many receivers currently operate, such measurements will lead to a very high and hence, unrealistic, results.

This impact is exacerbated by the aggregate effects measurements. See ITS UWB Test Plan, Section 4.5. Here ITS plans to correct the distance by using the data analysis from the simulators to account for interference effects. The plan to use "gated" simulator systems in Table 5 of the ITS UWB Plan will result in unrealistic measurements. Although the term "gating" is not defined, we assume that the test plan contemplates synchronizing the simulators to generate pulses synchronized in time. (Our assumption is supported by the last item of measured data to be taken, as outlined in Section 4.5: "Compare APDs [i.e., Amplitude Probability Distributions] of the measured aggregates to results obtained by mathematically combining individual APDs.") These gated pulses would impact on a victim receiver with three copies of a single energy pulse. Even where a system is synchronized, this testing configuration is not accurate. Because of the path differentials from multiple UWB transmitters, a victim receiver would at no time be presented with three synchronized-in-time pulses.

Section 4.1, the plan proposes in item three to use an instrument with a 50 MHz IF bandwidth. If this is the same or a similar instrument to the one supplied by ITS for trial use during the evaluation of the UWB equipment subject to the waiver requests, the device would not be appropriate. The instrument did not act like a spectrum analyzer for resolution bandwidths greater than 3 MHz, and had very poor dynamic range. Moreover, the specified bandwidth was not accurate (while the stated bandwidth was 50 MHz it was in fact more like 30 MHz). In addition, this device hardly qualifies as commercial-off-the-shelf ("COTS") equipment of the kind that manufacturers and test labs are likely to have. In sum, the test plan should state with greater particularity (e.g. make, model, and where appropriate, software version) the equipment proposed for critical measurement tasks.

Where the plan does contemplate radiated measurements, it does not specify whether the tests will be in an anechoic chamber or on an open air test site (an "OATS site"). The results could vary by 6 dB in field strength between these sites. Any limit developed on one site should recognize the difference due to the lack of an additive reflective component in the case of the OATS site measurement.

Equation 4.4.1 presents an overly simplified view. It fails to take into effect antenna beam pattern (both static and dynamic). This is important because many of the antennas for systems that are potentially susceptible to interference rotate and scan from the horizon skyward. As a result, there should be less interference from UWB emitters than a model that failed to account for such operational characteristics would yield.

The plan refers to "total peak power" in Table 1. Yet, the plan does not explain whether this is RMS peak, peak as measured in the frequency domain, or peak as measured in the time domain.

The plan commendably proposes to make measurements in the time domain. Nevertheless, if ITS is going to try to derive the actual radiated time domain waveforms (as opposed to the received waveforms) by performing the time domain measurements on the UWB signals, it will need to back out the losses, gains or phase distortions of the receive antennas, cables, and any amplifiers, if used. This will be very difficult because these devices are characterized in the frequency domain, but the measurement is made in the time domain. In sum, to back out to the actual radiated time domain waveform will be a very complicated mathematical activity. Time Domain would offer further comments on this aspect if the measurement antenna to be used were known. Here, too, however, the plan fails to reveal details needed to provide insight as to the effort. Any measurements to be performed in this manner must have well defined antenna and cables.

Other Matters:

The following reflect errors that either may be editorial or typographical in nature or comments that do not go to the crux of the plans:

Table 4(b) sets forth incorrect units for E_o. The units of measurement should be dBuV/m (dB microvolts per meter).

Use of a thermistor based power meter as proposed in the plan makes sense. Diode power meters will often give a misleading or incorrect answer for fast pulse systems.

At Figs. 2 and 3, ITS proposes to use a digital sampling scope. This instrument should be capable of at least 20 Gigasamples per second in order to perform satisfactorily.

The last bullet point in Appendix A should specify E as dBuV/m instead of dBmV/m.

* * *

To be of predictive value in the effort to develop UWB regulations that provide the benefits of ultra-wideband to the public without causing harmful interference, the testing should not start with any preconceived notions of compatibility, but instead seek to measure adequately and quantify the effect of UWB signals on existing systems. As part of this effort, criteria can then be developed to define harmful interference in the context of UWB emissions. From these criteria, proper operating limits for UWB systems can be developed. To accomplish this goal the testing effort should examine real world environments and not only the theoretical noise limited environment.

The testing effort should reflect the ongoing dialog amongst those knowledgeable of both protected systems and ultra-wideband technology. To this end, Time Domain appreciates the opportunity to comment and hopes that this dialog can continue.

Respectfully,

Time Domain Corporation

/s/ Paul Withington

Paul Withington

Vice-President

paul.withington@tdsi.com