doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 1 Project: IEEE P Working Group for Wireless Personal Area Networks (WPANs) Submission Title: [Peak Power Margin for UWB waveforms] Date Submitted: [ 11 May, 2005] Source: [C. Razzell] Company [Philips] Address [1151 McKay Drive, San Jose, CA 95131] Voice:[ ], FAX: [ ], + [R. Aiello] Company [Staccato Communications] Address [5893 Oberlin Drive, Suite 108, San Diego, California 92121] Voice:[ ], FAX: ], + [D. Leeper] Company [Intel Corporation] Address [CH6-460, 5000 W Chandler Blvd., Chandler, AZ, 85226] Voice:[ ], FAX: [], Re: [Previous a panel session discussion of FCC Waiver implications] Abstract:[Analysis and measurement of peak power headroom under FCC rules.] Purpose:[Consider carefully when evaluating claims of performance enhancements under FCC waiver] Notice:This document has been prepared to assist the IEEE P It is offered as a basis for discussion and is not binding on the contributing individual(s) or organization(s). The material in this document is subject to change in form and content after further study. The contributor(s) reserve(s) the right to add, amend or withdraw material contained herein. Release:The contributor acknowledges and accepts that this contribution becomes the property of IEEE and may be made publicly available by P

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 2 Outline Under the recent FCC waiver, gating may be applied to UWB waveforms to allow higher signal level in bursts shorter than 1ms The only power limit, in addition to the average -41.2dBm/MHz, is peak power limit of 0dBm in a 50MHz bandwidth –However the standard measurement procedure is at 3MHz DS-UWB authors have stated that waiver provides 4x advantage over MB-OFDM, in terms of increase in throughput, increase in transmit power, decrease in power consumption (*). This implies that: –DS-UWB signals have at least 6dB peak to average margin advantage with respect to MB-OFDM –Higher signal level, lower duty cycles signals have advantage in terms of increase in throughput, increase in transmit power, decrease in power consumption This work seeks to determine the peak to average margin, that limits the headroom available for increasing the signal level power in bursts shorter than 1ms –Simulation for MB-OFDM, AWGN and DS-UWB –Conducted (not radiated) measurements for MB-OFDM and AWGN (*) Source: FCC Waiver Ruling, March 10, 2005, Technical Overview Martin Rofheart, Director of Ultra-Wideband Operations, Freescale Semiconductor

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 3 Simulation Results

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 4 FFT Simulation Methodology ( Mimics Modern Digital Spectrum Analyzer) Baseband simulation with a sampling rate of 2640MHz A single MB-OFDM sub-band is simulated with 5x oversampling –A 1 in 3 duty cycle is applied to simulate frequency sequencing A DS-UWB impulse waveform with a 220MHz PRF and a chip rate of 1320Mcps is simulated with 2x oversampling 1ms of time domain data was generated for FFT analysis. The number of points in the FFT is given by NFFT=Fs/RBW, where Fs is the sampling frequency and RBW is the desired resolution bandwidth. N non-overlapping FFTs were taken, where N=TS*RBW, where TS is the simulation time. The mean and the max of the N values in each FFT bin were taken as the estimated spectra

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 5 3MHz bandwidth, DS-UWB case

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 6 3MHz bandwidth, MB-OFDM case

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 7 3MHz bandwidth, AWGN case

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 8 Summary of FFT Simulation Results Resolution Bandwidth (MHz) Power (dBm)

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 9 Simulation Results at 50 MHz & 3 MHz Observed Peak Power Headroom 50 MHz limit –DS-UWB: 17dB –AWGN: 13 dB –MB-OFDM: 8 dB 3 MHz measurement –DS-UWB: 0 dB margin –AWGN: 1 dB margin –MB-OFDM: 1 dB margin Power (dBm)

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 10 Measurements

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 11 Measurement Setup Equipment Used –Staccato Communications Ultrawideband full-rate transmitter: Model number SC1010DB2 –Spectrum Analyzer: Rohde and Schwarz R&S FSP (9KHz – 13.6GHz) –Noise Source: NoiseCom (NC346 Series) 50MHz – 24GHz) Calibration: –RMS power levels (MB-OFDM and Thermal Noise) observed by activating the RMS detector on SA. Corresponding power offsets (simple delta) introduced to achieve the -41dBm/MHz FCC mandate Remainder of setup notes in backup foils

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 12

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 13 Summary measurement results 10 MHz measurement –AWGN: 7 dB –MB-OFDM TFC 1: 9 dB –MB-OFDM TFC 6: 16dB 3 MHz measurement –AWGN:.7 dB margin –MB-OFDM TFC 1: 3.7 dB –MB-OFDM TFC 6: 10dB

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 14 Discussion and Conclusions

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 15 Discussion Measurement of peak power based on a 50MHz measurement bandwidth cannot be performed with normal commercial EMC test equipment –3 MHz RBW adopted by FCC for their own labs –Allowance is made for wider bandwidths, but the measurement technique will be scrutinized closely Using a 3MHz bandwidth, there is virtually no headroom for gating –DS-UWB and AWGN have 0-1dB margin at 3MHz (simulation and measurement are in agreement for AWGN) –MB-OFDM has 1-3.7dB margin at 3MHz (simulation and measurement agree within 2dB) At 50MHz bandwidth there are measurement challenges, but if overcome and accepted by the FCC: –> 8dB headroom for MB-OFDM (simulation and measurement) –15dB for DS-UWB (simulation)

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 16 Conclusions - recommendations Estimates of peak power headroom are dependent on the measurement or simulation methods chosen –Simulation shows similar margins at 3MHz, larger margin for DS-UWB at 50MHz –Measurement shows larger margin for MB-OFDM at 3 and 10MHz –MB-OFDM transmitters could take advantage of the waivers gating rule to increase the signal level by a factor of 6 Additional peak-to-mean ratio introduced by pulse gating will increase concern for interference potential to narrowband receivers by a similar amount. –We know that all receivers with bandwidths greater than a few kilohertz will experience proportionally increased peak interference powers –With slow gating, less chance for channel filtering to perform averaging and for FEC interleavers to perform randomization UWB transmitters need to co-exist peacefully with other services in the same frequency of operation –Slow gating waveforms with an additional 15dB of peak power have not been subject to extensive coexistence studies –We prefer to take a more cautious approach until more data is available

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 17 Backup

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 18 Measurement set-up (contd..) MB-OFDM measurements (TFC mode 1 and 6) –Center Frequency: 3096 MHz (center frequency of BAND_ID 2 of Band Group 1). –Frequency span: 500MHz –Trace detector: Max Peak Detector with MAX hold ON –Resolution BW: 1, 3, 10MHz; Video BW: 10MHz (constant) –Sweeptime: 500ms Thermal Noise measurements –Center Frequency: 4092 MHz –Frequency span: 1600MHZ –Trace detector: Max Peak Detector with MAX hold ON –Resolution BW: 1, 3, 10MHz; Video BW: 10MHz (constant) –Sweeptime: 500ms

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 19 Peak envelope value of bandpass filtered Gaussian Noise

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 20 Peak envelope value of bandpass filtered Gaussian Noise

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 21 We believe that there is a simpler method of measuring peak emission levels in a manner that also takes into account the interference potential of the equipment. In order to perform a peak measurement on a spectrum analyzer, the VBW must be at least as large as the RBW. The largest VBW on a spectrum analyzer is about 7 MHz. Thus, the widest RBW that could be employed is 3 MHz. However, there are several receivers used by the authorized radio services that employ greater bandwidths. Thus, the concern is how to ensure that peak measurements performed with a 3 MHz RBW will protect receivers that employ a wider bandwidth from harmful interference. The peak EIRP limit is 20 log (RBW/50) dBm when measured with a resolution bandwidth between 1 MHz and 50 MHz. RBW is the resolution bandwidth in megahertz actually employed. This bandwidth must be centered on the frequency at which the highest radiated emission occurs. We intend to employ at our laboratory a measurement procedure using a 3 MHz resolution bandwidth. However, we will permit responsible parties to test their UWB products using different resolution bandwidths ranging from 1 MHz to as high as 50 MHz. The use of a higher resolution bandwidth may be particularly helpful for measuring a system operating at a higher PRF. If a resolution bandwidth greater than 3 MHz is employed, the application for certification filed with the Commission must contain a detailed description of the test procedure, calibration of the test setup, and the instrumentation employed in the testing. Peak Power Measurement per FCC R&O

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 22 1MHz bandwidth, DS-UWB case

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 23 1MHz bandwidth, MB-OFDM case

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 24 1MHz bandwidth, AWGN case

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 25 Caveats on AWGN Peak Power The peak power of a Gaussian signal source is not bounded, and does not stabilize over time As observation time increases, measured peak values continue to increase without limit The maximum value is given by a single sample, which was the worst case over the entire simulation run Over several trials of the same experiment, different results can be obtained

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 26 Peak/Mean envelope value vs. Sample Size for a Complex Gaussian Source

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 27

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 28 Additional Peak Power Simulation Setup Simulating Traditional Analog (Envelope-Detector) Spectrum Analyzer Notes MB-OFDM clipped at 9 dB PAR DS-UWB BPSK Low-Band Code Length 6, Code Set 1 AWGN BW ~ 1600 MHz BPF 6-Pole Butterworth BW = 1, 3, 10, 50 MHz Scaled to dBm/MHz avg PSD for all waveforms Fixed Center Freq = GHz1 ms observation times MB-OFDM Tx DS-UWB Tx AWGN Waveforms LPF True RMS Volts True Peak Voltage

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 29 Peak & Avg Power vs Resolution Bandwidth MB-OFDM, DS-UWB, and AWGN Simulated Analog (Envelope Detector) Spectrum Analyzer Fixed RBW filter ctr freq = 4092 MHz 1 ms simulation Avg PSD = dBm/MHz for all waveforms Averages Peaks Power (dBm)

doc.: IEEE r0 Submission May 2005 C. Razzell, R. Aiello, D. LeeperSlide 30 Headroom Against FCC Peak Power Limits MB-OFDM, DS-UWB, and AWGN Simulated Analog (Envelope Detector) Spectrum Analyzer Resolution Bandwidth (MHz) Headroom (dB)