November 2004 doc IEEE 802 15 040641 r
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Project: IEEE P 802. 15 Working Group for Wireless Personal Area Networks (WPANs) Submission Title: [MB-OFDM No Vote Responses] Date Submitted: [Nov 15, 2004] Source: [Matthew B. Shoemake and John Terry, Joy Kelly and Jim Lansford, David Leeper and Jeff Forrester, Joe Decuir, Charles Razzell] Company [Wi. Quest, Alereon, Intel, MCCI, Philips] Address [8 Prestige Circle, Suite 110, Allen, Texas 75013] Voice: [+1 214 -547 -1600], FAX: [+1 214 -547 -1606] E-Mail: [Provided throughout document] Re: [IEEE 802. 15. 3 a No Vote Responses from September 2004 Vote] Abstract: [This presentation contains no vote responses to comments submitted after the IEEE 802. 15. 3 a downselect vote in September 2004. These responses have been complied by multiple supporters of the Multiband OFDM approach as indicated throughout the document. ] Purpose: [The purpose of this presentation is to address no vote responses thereby enabling an affirmative vote during the confirmation step and thereby enabling IEEE 802. 15. 3 a to move forward into the working group balloting stage of standardization. ] Notice: This document has been prepared to assist the IEEE P 802. 15. 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 802. 15. Submission 1
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Outline 1. 2. 3. 4. 5. 6. 7. 8. Introduction – Matthew Shoemake, Wi. Quest Regulatory Compliance and Interference – Joy Kelly, Alereon MAC – Matthew Shoemake Location Awareness – Joe Decuir, MCCI Harmonization or Coexistence – David Leeper, Intel Multipath Performance – Charles Razzell, Philips Time to Market – Jim Lansford, Alereon Summary Submission 2
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Regulatory Compliance and Interference Charles Razzell, Philips, charles. razzell@philips. com Jeff Foerster, Intel, jeffrey. r. foerster@intel. com Joy Kelly, Alereon, joy. kelly@alereon. com Submission 3
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Overview of FCC Compliance / Interference Section - Many no vote comments have been elaborated on in reply comments made in response to the MBOA waiver request - Following slides review significant points raised and provide technical analysis, simulations, lab measurements, and field measurements demonstrating that MB-OFDM waveforms, as proposed in the Waiver, will not cause greater interference than waveforms already allowed by rules – This body of work is precisely what was requested by FCC/OET (i. e. to demonstrate no greater interference potential than waveforms allowed by the rules) MB-OFDM devices operating under the proposed Waiver will comply with the Part 15 f limits on peak and average power Submission 4
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Summary of main opposing comments & claims MB-OFDM will increase the potential for interference - Not true, as will be shown here Granting the Waiver will give MB-OFDM an unfair advantage (increased range) relative to other UWB technologies - Not true, even by opposer’s claims Waiver will ‘open the door’ to other systems seeking relief from the rules - Scope of Waiver is narrow and does not impact most of the FCC rules FCC should wait for more data and delay making a ruling - Reply comments provide comprehensive data; no new information will come from more tests on the 3 -band MBOFDM waveforms Waiver is not in the public interest and will negatively impact small businesses - MBOA SIG represents 170+ companies, including many small start-ups Submission 5
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Summary of main opposing comments & claims MBOA technical justification is filled with errors - Inclusion of WGN in comparisons ‘masks’ MB-OFDM interference potential Thermal noise and other interference sources are a reality - Wrong BER operating point BER criterion based on quasi-error free performance - Field measurements are invalid Same position and separation distance tests are valid and reflect real systems - Simulations results are wrong because they included noise Noise is a reality and simulation results are supported by lab and field measurements - APD analysis is erroneous Shown to be technically accurate using NTIA code Submission 6
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Summary of technical points No greater interference than systems allowed by FCC rules - All UWB signals will be well below the system noise floor of C-band satellite receivers Makes differences between waveforms negligible - MB-OFDM looks like WGN to narrow bandwidth systems (less than a few MHz), including OFDM systems with narrow tone spacing - MB-OFDM systems do not synchronize and will not increase the potential for aggregation of interference - MB-OFDM has been consistently shown to have less interference than a class of impulse radios allowed by the rules, supported by analysis, simulations, lab measurements, and field measurements Differences between all UWB signals allowed by the rules are within a few d. Bs when measured in realistic scenarios Bandwidth of information-carrying tones is 503. 25 MHz Submission 7
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Summary of technical points (cont) MB-OFDM technology advantages - Band switching (the multi-band concept) increases frequency diversity provides coarse spectrum flexibility at Tx enables efficient CMOS designs provides protection from strong interferers at Rx - OFDM efficiently captures multipath energy, shares common components with other technologies (Wi. Fi, Wi. Max, DSL) leveraging best known methods in design and manufacturing, provides fine spectrum flexibility at Tx, and enables efficient signal processing techniques for interference mitigation in Rx - Spectrum flexibility will be necessary to enable worldwide interoperability and to adapt to future spectrum allocations Submission 8
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 No greater interference: C-band satellites 802. 11 a devices Other UWB devices No risk of aggregation Submission 9
November, 2004 C-band Satellites doc. : IEEE 802. 15 -04/0641 r 0 - C-band satellite systems have little margin MBOA field measurements have confirmed this (only 2. 5 d. B margin) Table 1: Power spectral density limits for some US government bands Submission 10
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 C-band Satellites (cont) C-band satellites have low margin due to challenges of communicating over long satellite link (see Petition for reconsideration of Satellite Industry Association in docket 98 -153) - Low margin requires interference to be below systems noise floor of Cband satellite receiver What is system noise floor for C-band satellites? - From SIA, Isat is defined as the interference at a given C-band receiver caused by adjacent C-band channel interference and cross polarization noise from the satellite link - Isat/N = 1. 4 d. B (1. 38) - N_sys = N + I_sat where N is thermal noise floor FCC adopted criterion of IUWB /N < 0 d. B (docket 98 -153 March 12, 2003) - N_sys = N + I_sat = (1 + 1. 38)= 2. 38 - IUWB /N_sys = 10*log 10(1/2. 38) = -3. 8 d. B Submission 11
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 C-band Satellites (cont) - FCC criterion, when incorporating total satellite system noise experienced (as defined by SIA) is thus IUWB /N_sys = -3. 8 d. B - Further support of appropriate protection levels for C-band satellites Xtreme. Spectrum filed response to SIA Petition for Reconsideration (Sept 4, 2003 in docket 98 -153) stating that – “Using SIA’s stated operating levels, XSI [now part of Freescale] demonstrates that an I/N of – 6 d. B assigned to a UWB device is an appropriate protection level” – IUWB /N = -6 d. B IUWB /N_sys = -9. 8 d. B Submission 12
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 C-band Satellites (cont) - Simulation results from an Alion report presented to FCC (Feb 11, 2004) - Based on simulation results, coalition of C-band constituents proposed reducing current FCC limits by 21 d. B (98 -153 & 02 -380 February 18, 2004) - Motorola (in ET Docket No. 98 -153, April 9, 2004) provided analysis in response to Alion report which took into account more realistic factors into the simulations. Motorola concluded that ”Based on the revised simulations with the more realistic path loss models, building blockage effects for devices of high in the air (and near the antenna main beam), the inclusion of a realistic duty cycle (<10%) and realistic density projections, it is clear that no significant interference will result. The aggregate UWB signal power levels drop by 25 -60 d. B when more realistic assumptions are made in the simulations. ” Submission 13
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 C-band Satellites (cont) - UWB average interference power will be well below system noise floor when considering realistic deployment scenarios - When I/Nsys << 0 d. B, simulations, lab measurements, and field measurements show little difference between MB-OFDM, impulse, and DS-UWB waveforms Results also consistently show MB-OFDM causes less interference than impulse radios already allowed by the rules Submission 14
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 C-band Satellites (cont) Supporting analysis from other companies - Freescale commented that “…when you mix one part MB-OFDM energy with 20 parts Gaussian noise, the result is a composite signal that looks very much like noise. ” - Time. Derivative commented that “At very low I/N ratios the noise dominates and little can be said about the difference between various interferers. ” This is precisely the point: to simulate, analyze, and measure the interference potential of alternative UWB waveform types in a realistic operating environment Submission 15
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 MBOA C-band Satellite field test results Field test objectives - Measure interference potential to C-band TV service operating in the FSS C-band 3. 7 -4. 2 GHz - Compare White Gaussian Noise (WGN), MB-OFDM & Impulse UWB signals - Quantify relative interference potential of each UWB signal - Determine safe distance from dish antenna to avoid interference Two separate tests were conducted. First test: compared the different UWB signals in terms interference potential to the FSS receiving system. Second test: determined the safe distance from the dish that must be maintained to avoid interference to the FSS receiver Submission 16
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 MBOA C-band Satellite field test results • Same position testing • Set UWB emissions to -41. 3 d. Bm/MHz for each waveform type • Measure interference power of each UWB waveform type required to yield visible block artifacts on TV monitor • Interference power measurement to produce visible block artifacts accurate to within 0. 1 d. B • Shows relative differences between waveforms • Devices being in ‘near-field’ of antenna is irrelevant • Separation distance testing • Devices had to be very close to antenna (in the ‘near-field’) to see measurable interference levels • Results are not ‘random’ and show UWB devices must be very close before interference is measurable Submission 17
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 MBOA C-band Satellite field test results: Same Position Testing • MB-OFDM has less interference than impulse radio and within 1 -1. 6 d. B from WGN • Simulations and lab measurements support this result Emission 0. 5 d. B above sensitivity 1 d. B above sensitivity 2. 5 d. B above sensitivity 3 MHz PRF impulse 0. 0 d. B MB-OFDM 3 band 0. 8 d. B 2. 6 d. B 2. 4 d. B WGN (DSSS) 1. 9 d. B 3. 8 d. B 4. 0 d. B Submission 18
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 MBOA C-band Satellite field test results: Safe Distance Tests MB-OFDM WGN Dish Orientation Scale in feet Submission 19
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Interference to 802. 11 a interference measurements were conducted, using an IEEE 802. 11 a device Two types of the interfering signals were considered - MB-OFDM signal - AWGN Submission 20
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Test Set-up for Interference Measurements to 802. 11 a Data Source at RF Interference Source at RF Adjustable Gain Digital Scope Adjustable Gain AWGN Channel 802. 11 a Receiver (Radio + PHY) Digital Scope Submission 21 BER Tester
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Test Description An IEEE 802. 11 a device was configured with the data rate of 36 Mbps (16 QAM, R=3/4) The IEEE 802. 11 a signal power was calibrated to the sensitivity level (0 d. B at BER=10 -5) in the absence of the interference. - This defines the operation thermal noise level. IEEE 802. 11 a signal level was adjusted to different levels in order to measure the impact of the interference signal and its power. With the interference added to the calibrated thermal noise, its power level (maximum tolerable interference power (MTIP)) was measured to maintain the IEEE 802. 11 a reception at BER=10 -5. Submission 22
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Interference to 802. 11 a Measurement results Signal Power of 802. 11 a above sensitivity I/N Difference between AWGN and MBOFDM Interference 10 d. B 9. 5 d. B 0. 5 d. B 3 d. B 0. 5 d. B 2 d. B -2. 3 d. B 0 d. B 1 d. B -5. 9 d. B 0. 5 d. B -9. 1 d. B -1. 5 d. B MB-OFDM produces no more interference to IEEE 802. 11 a WLANs than AWGN - 802. 11 a is OFDM based and uses symbol periods of 4 usec This results in integration over several MB-OFDM symbols, so only average interference power matters Submission 23
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Example BER results for 802. 11 a comparing MB-OFDM & AWGN Submission 24
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 802. 11 a AGC Performance in the Presence of MB-OFDM Transmission Contention that the MB-OFDM signal would cause performance degradation of the AGC in the IEEE 802. 11 a receiver. Comparison was made in the IEEE 802. 11 a packet detection and the AGC convergence performances between the MB-OFDM signal and AWGN. Measurement was conducted with the IEEE 802. 11 a device operating at 3 d. B above sensitivity. - conducted with the MB-OFDM signal level set to the MTIP level as well as to 10 d. B higher than the MTIP level. There was no detectable impact whatsoever of the MB-OFDM signal to the IEEE 802. 11 a packet detection and AGC performance. Submission 25
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 802. 11 a Packet Detection & AGC Performance in Presence of AWGN & MB-OFDM Interference Start of Packet SIFS AGC Locked Packet Detected AWGN Interferer Submission MB-OFDM Interferer 26
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 MB-OFDM Interference Impacts to 802. 11 a Systems: Conclusions Measurement results already presented to this IEEE body confirm that: MB-OFDM signal and AWGN have similar interference impact to IEEE 802. 11 a receiver MB-OFDM signal does not adversely affect packet detection and AGC convergence performance of IEEE 802. 11 a devices Submission 27
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Opponents’ Claims regarding MB-OFDM Interference to other UWB systems Freescale states (4. 4. 2, p. 23 of “Technical Analysis …”): ‘Other UWB receivers will be injured by the MB-OFDM emissions at least as much and often more than all the other victim systems since their bandwidths are so similar. While on its face, one would expect the 6 d. B higher emission limits to single out MBOFDM devices for a 2 X range advantage, the actual outcome is even worse. The noise floor of all other UWB devices would be raised far more by MB-OFDM devices than other classes of UWB devices. ’ Pulse-Link states in Section III of their “Comments …” : ‘Granting the waiver would allow the MBOA radio to more successfully jam the DS-UWB radio since it will be allowed an increase of power in band. ’ Time. Derivative states: ‘This additional power poses a significant additional risk to other UWB communications equipment. ’ No evidence has been shown to support these claims. Submission 28
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Reality: MB-OFDM Interference to other UWB systems On the contrary, consider the following example: - Assume an MB-OFDM signal which occupies a total bandwidth of 3*528 = 1584 MHz peak power spectral density (PSD) during the OFDM symbol ‘on’ time is 5. 8 d. B above average PSD occupied bandwidth of one symbol is ~500 MHz. - Evaluate interference experienced due to this signal by an impulse radio system occupying the same total bandwidth of 1584 MHz (for comparison, Freescale’s proposed DS-UWB system defines impulse radio modulation using impulses of bandwidth 1320 MHz and a PRF of 220 MHz to deliver a data rate of 110 Mbps. ). - Assume interference much higher than system noise. Submission 29
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Reality: MB-OFDM Interference to other UWB systems - At any given instant, one 500 MHz portion of impulse radio’s occupied band is impacted by an MB-OFDM symbol - Impulse radio receiver matched filter integrates all interference power over the full bandwidth of 1584 MHz. - Thus, the total instantaneous interferer power[1] at the output of a 1584 MHz matched filter is - IMB-OFDM = (-41. 3 + 5. 8) d. Bm/MHz + 10*log 10(500) MHz = -8. 5 d. Bm [1] Instantaneous interference power refers to the maximum interference power to be expected while the interference source is active or ‘on’. Submission 30
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Reality: MB-OFDM Interference to other UWB systems total instantaneous interferer power at the matched filter output from another impulse UWB radio system occupying the same 1584 MHz bandwidth would be IDS-UWB = (-41. 3) + 10*log 10(1584) = -9. 3 d. Bm MB-OFDM system offers at worst 0. 8 d. B higher potential interference in this example Furthermore, if we consider a more realistic set of conditions, this modest impact would be reduced still further. - including system noise in addition to the interference - Accounting for target DS-UWB system FEC protection Submission 31
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 MB-OFDM systems will not increase aggregate interference levels Freescale and others claim a MBOA device “seeks out and transmits on channels momentarily left vacant by others” - This requires nanosecond time-scale synchronization between devices belonging to different networks: NOT facilitated by MBOA protocols Uncoordinated MBOA devices pick different time-frequency codes (TFC) using similar protocols as other UWB devices do in order to select logical channels corresponding to their PHYs timing offsets between uncoordinated devices are random timing drifts between uncoordinated devices further randomize emissions Unrealistic fine-scale Synchronization: NOT facilitated for MBOA devices belonging to different networks Aggregation results for MB-OFDM devices are no different than those for pulse based UWB devices Submission 32
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 No greater interference: Comparisons of various UWB waveforms impact to a generic wideband DVB receiver Submission 33
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Victim Receiver The victim receiver chosen for study was a Digital Video Broadcasting receiver with the following characteristics: - QPSK modulation with a transmission symbol rate RS of 33 Msymbols/second - Root Raised Cosine Filtering at Tx and Rx - Viterbi FEC, with rates of 1/2, 2/3, 3/4, 5/6 and 7/8. Representative of the most vulnerable class of victim receivers - Very wide bandwidth and low error rate requirement BER at output of Viterbi decoder for comparisons = 2 x 10(Quasi-error free operating point for satellite systems employing standard concatenated code) 4 Submission 34
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 BER Criterion for Digital Video Freescale claims that the MBOA analysis is based upon the wrong BER criterion and stated that the BER used in our comparisons were ‘ 7 orders of magnitude higher than the specification’. They claimed that ‘the curves and corresponding conclusions are misleading, and to a skilled communications engineer, they are fatally flawed and have no technical merit. ’ It is well-known that digital video needs a low BER. it is also well-known that a BER of 2 x 10 -4 at the output of the Viterbi decoder yields quasi-error free performance at the output of the Reed-Solomon decoder (meaning a BER of 10 -10 to 10 -11)[1]. Therefore, comparing the performance at the output of the Viterbi decoder at a BER of greater than or equal 2 x 10 -4 is completely relevant to this discussion Submission 35
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Fundamental assumptions: What is the right BER criterion? • MBOA, Freescale, and Time. Derivative each proposed different BER operating points to do comparisons • Alion studies, MBOA studies, and Motorola studies all used digital C-band satellite equipment employing the same FEC as described in ETSI EN 300 421 V 1. 1. 2 • Requirements suggest interference impacts would occur somewhere above a BER of 2 x 10 -4 after the Viterbi decoder • MBOA has provided substantial data at this operating point as well as at other operating points *The following table and text is copied from ETSI EN 300 421 V 1. 1. 2 Submission 36
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Fundamental assumptions: System Noise must be considered in the analysis MBOA and Freescale have a fundamental difference of opinion with respect to inclusion of system noise in comparisons of different waveforms MBOA position is simple: - System noise is a reality…why ignore it? System noise = thermal noise + intra-system interference • When average UWB interference power is well below the system noise floor, differences between UWB waveforms are negligible (even Freescale agrees) • Even at moderate I/N levels, differences between various UWB waveforms are small • Misleading conclusions result when not considering realistic environments • Differences between waveforms are highly exaggerated when measured in unrealistic, noise-less environments System noise must be considered in any interference analysis Submission 37
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Noiseless 7/8 -Rate Coded Results INTERFERER ONLY: NO THERMAL NOISE! Impulse radio at 1 MHz PRF 0 10 -1 7/8 -rate coded QPSK BER 10 MB-OFDM 1, 1, 3, 3, 2, 2 -2 10 -3 10 MB-OFDM 1, 2, 3, 1, 2, 3 impulse simulation -4 10 MB-OFDM 1, 2, 3, 1, 2, 3 WGN noise only MB-OFDM 1, 1, 3, 3, 2, 2, -5 10 WGN only -6 10 -2 0 2 4 Eb/Io [d. B] 6 8 10 Why was the impulse-induced error rate not completely corrected by the FEC? Easily explained by looking at free distance of code and collision properties Submission 38
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Impulsive Interference and FEC Real part of received QPSK waveform X: 1633 Y: 4. 546 4 3 2 X: 2145 Y: 0. 7067 1 At equal power levels, 0 -1 X: 1633 Y: -0. 7079 -2 Real part Interfering Impulse -3 -4 1600 1700 Submission 1800 1900 2000 X: 2145 Y: -4. 546 2100 2200 2300 2400 39 impulse amplitude is 16 d. B greater than wanted QPSK signal. Soft decisions associated with collisions will be highly confident (and wrong). Impulse interference “pollution” of the Viterbi path metrics may last for quite some time and cause an associated error burst. The duration of the negative impact of an impulse is much longer than that of the impacted bits, especially where Viterbi decoding with soft metrics is used. Soft decisions must be clipped in magnitude to prevent excessive error propagation.
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 ¾-rate Coded Results (with noise) 10 I/Nsys = -10. 0, soft metric clipping 0 At I/Nsys=-10 d. B, there is very little difference between the impact of extra Gaussian noise and any of the other types of interference studied. 0. 75 -rate coded QPSK BER impulse simulation 10 10 10 MB-OFDM 1, 2, 3, WGN noise + WGN int. -1 MB-OFDM 1, 1, 3, 3, 2, 2, -2 # XSI (now Freescale) filed comments in support of this I/N level: -3 “Using SIA’s stated operating levels, XSI demonstrates that an I/N of – 6 d. B assigned to a UWB device is an appropriate protection level. ” -4 -5 -6 -1 0 1 2 3 4 5 Eb/No [d. B] Nsys = thermal noise + intra-system interference Submission 40 6 7 8 I/N=-6 d. B corresponds to I/Nsys=-10 d. B
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 ¾-rate Coded Results (with noise) I/Nsys = -5. 0, soft metric clipping 0 0. 75 -rate coded QPSK BER 10 At I/Nsys=-5 d. B, we begin to see a clear ordering, with the impulse radio being the worst case. The spread, however, is only 1 d. B. impulse simulation MB-OFDM 1, 2, 3, WGN noise + WGN int. MB-OFDM 1, 1, 3, 3, 2, 2, -1 10 -2 10 this is quite a severe case, requiring very close proximity to the victim satellite receiver dish. -3 10 -4 10 1 2 3 4 5 6 7 Eb/No [d. B] Submission 41 8 9
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 ¾-rate Coded Results (with noise) 0. 75 -rate coded QPSK BER 10 10 I/Nsys = 0. 0, soft metric clipping 0 At I/Nsys=0 d. B, the same ordering is maintained, with the impulse radio being the (co-equal) worst case. impulse simulation MB-OFDM 1, 2, 3, WGN noise + WGN int. MB-OFDM 1, 1, 3, 3, 2, 2, -1 -2 -3 -4 -5 -6 3 4 5 6 7 8 9 10 Eb/No [d. B] Submission 42 11 12 This is a highly exaggerated and very unlikely case, considering that even the White Gaussian interferer has reduced the available link margin by 3 d. B, enough to cause link failure in many installations.
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 ½-rate Coded Results (with noise) I/N = 0. 0, soft metric clipping impulse simulation ½-rate coded QPSK BER 10 10 This is (again) a highly exaggerated and very unlikely case, considering that even the White Gaussian interferer has reduced the available link margin by 3 d. B, enough to cause link failure in many installations. MB-OFDM 1, 2, 3, 1, 2, 3 WGN noise only -1 MB-OFDM 1, 1, 3, 3, 2, 2, -2 -3 -4 1 2 3 4 5 6 7 8 Eb/No [d. B] Submission 43 9 10 Nevertheless, the spread in susceptibility to the various waveforms is <2. 1 d. B at the target BER of 2 x 10 -4.
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 ½-rate Coded Results (with noise) ½ - rate coded QPSK BER I/N = -5. 0, soft metric clipping impulse simulation MB-OFDM 1, 2, 3, 1, 2, 3 WGN noise only MB-OFDM 1, 1, 3, 3, 2, 2, -1 10 The ability of the FEC to deal with impulsive interference depends on its strength. -2 10 With a very strong FEC, such as the ½-rate code used here (dfree=10), it can happen that the impulsive interference is better tolerated than the MB-OFDM waveforms. -3 10 -4 10 1 2 3 4 5 6 Eb/No [d. B] 7 8 9 10 However, the absolute difference remains small (the spread between all waveforms is less than 1 d. B). Differences of ~ 2 d. B or less are WITHIN measurement tolerance of instrumentation Submission 44
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 LAB Measurement Results Iuwb /(N+Isat) -15 -10 -5 0 1. 00 E-02 Results show that the order of interference impact starting with most benign is: AWGN 1. 00 E-03 CP MB -OFDM ZP MB -OFDM Pulse 1 M 1. 00 E-04 Pulse 3 M Viterbi BER • AWGN • 3 MHz PRF impulses • Cyclic Prefix MB-OFDM • Zero Prefix MB-OFDM 1. 00 E-05 • 1 MHz PRF impulses Relative impact and degree of impact from lab measurements resemble those from the simulations. Submission 45
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Conclusions on Interference Impact on Wideband DVB Receiver For rate ¾ and 7/8 codes, MB-OFDM is more benign than a 1 MHz impulse radio. Our simulations and measurements focused mainly on the ¾-rate code as being a representative choice from the available rates [1/2, 2/3, 3/4, 5/6, 7/8]. Low rate codes (½-rate code, for example) are slightly more tolerant to the 1 MHz PRF impulses than to MB-OFDM waveforms - Differences are still small among all the waveforms, when realistic I/Nsys ratios are used. - Interference analysis with a low rate code is not representative of a ‘worst-case’ situation (i. e. , a UWB device needs to be much closer to a victim using a low rate code compared a victim using a high rate code) Under realistic, worst-case scenarios, MB-OFDM produces consistently less interference than a class of impulse radios already allowed by the rules Submission 46
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 No greater interference: APD Analysis Submission 47
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 APD Analysis APD plots have been used by the NTIA in the course of interference studies and have been described as “a very informative measurand. ” (See NTIA Report 01 -383. ) It is important to be aware of the limitations of APD plots, and we agree that they are not susceptibility tests and should be viewed in conjunction with detailed simulations, lab measurements, and field measurements supplied by the MBOA. However, we do still value APD plots (as does the NTIA) for their ability to describe the potency of an interfering waveform, irrespective of the particular modulation and channel coding scheme used Submission 48
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Variation with I/Nsys ratio 20 15 10 d. BV relative to mean The plotted curves all show rather high (and in some cases extremely high) I/N ratios. MB-OFDM I/N= -6. 0 d. B MB-OFDM I/N= -3. 5 d. B MB-OFDM I/N= 0. 0 d. B MB-OFDM I/N= 8. 0 d. B noise alone In order to observe differences in susceptibility of as much as 5 d. B relative to AWGN: 5 0 -5 • The I/N ratio must be at least 8 d. B AND -10 • The receiver must respond to peak events with a probability as low 10 -6 AND -15 -20 • The bandwidth of the victim must exceed 16 MHz -25 -30 0. 0001 0. 1 1 5 10 20 30 40 50 60 70 80 90 95 98 99 percent exceeding ordinate APD analysis proves large impact (5 d. B as claimed by Freescale) is only possible when a large bandwidth receiver with no FEC is in extremely close proximity to a MB-OFDM device. Joint probability of this event is vanishingly small. Submission 49
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Variation with Victim Rx BW APD plots for MB-OFDM using 1, 1, 3, 3, 2, 2 TFI code APD plots for MB-OFDM using 1, 2, 3, 1, 2, 3 TFI code Amplitude relative to mean [d. BV] 20 20 bw=2 MHz bw=4 MHz bw=8 MHz bw=16 MHz bw=32 MHz 10 10 0 0 -10 -20 -30 -40 -50 0. 0001 0. 1 1 5 10 20 30 40 50 60 70 80 90 95 98 bw=2 MHz bw=4 MHz bw=8 MHz bw=16 MHz bw=32 MHz -50 99 0. 0001 0. 1 1 5 10 20 30 40 50 60 70 80 90 95 percent exceeding ordinate • In bandwidths of 4 MHz or less, the MB-OFDM waveforms are nearly identical to an ideal AWGN source for any given probability. • For large bandwidths, the APDs are almost identical regardless of which TFI code is used (1, 2, 3, 1, 2, 3 or 1, 1, 3, 3, 2, 2). Submission 50 98 99
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Summary of APD results APD analysis is correct and verified using NTIA code APD results for various receiver bandwidths and different TFI codes provided in back-up slides for more information - In bandwidths of 4 MHz or less, the MB-OFDM waveforms are nearly identical to an ideal AWGN source for any given probability. - For large bandwidths, the APDs are almost identical regardless of which TFI code is used (1, 2, 3, 1, 2, 3 or 1, 1, 3, 3, 2, 2). Submission 51
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Summary of APD results A very specific (and unlikely) combination of circumstances must occur to support a required SIR protection difference of 5 d. B as claimed by opponents of this waiver including: - Very wide victim receiver bandwidth - Extraordinary and improbable I/N ratios not anticipated by the 2002 UWB R&O - No error correction capability for the victim receiver. Under realistic I/Nsys ratios, APD analysis shows MB-OFDM only deviates from WGN by a few d. Bs a small percentage of time Submission 52
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 MB-OFDM Power Levels Submission 53
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Claims regarding Power Level (1) From Freescale waiver comments: Somewhere in the spectrum, at every instant [emphasis added], the MBOA system would be emitting at three times the [power] level permitted to an impulsive or direct sequence system. Why this is misleading: 1. “Instantaneous power level” has meaning only in the time domain. Clearly many impulse-based systems permitted by the rules have higher instantaneous time-domain power levels than MB-OFDM does. 2. In the frequency domain, power spectral density (d. Bm/MHz) cannot be measured in an “instant”. Some averaging time must be specified, such as the 1 ms interval designated in the rules. Even under the shortest interval found on high-end analyzers (10 ms), average and peak PSD for MB-OFDM comply with the rules. See charts that follow. Submission 54
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Average-Power Compliance -41. 3 d. Bm/MHz Waiver Reply Figure 21: Example average EIRP measurement for a MB-OFDM transmitter using intended operational mode according to the waiver MB-OFDM waveform, measured under authentic operating conditions, conforms to Part 15 requirements not to exceed -41. 3 d. Bm/MHz PSD. Submission 55
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Peak Power Compliance 0 d. Bm Waiver Reply Figure 22: Example Peak EIRP measurement for a MB-OFDM transmitter [Measurements taken at a RBW of 3 MHz and compensated by 20 Log(3/50) RBW factor to compare with FCC UWB peak limit in a 50 MHz RBW] MB-OFDM waveform, measured under authentic operating conditions, conforms to Part 15 requirements not to exceed 0 d. Bm peak power. Submission 56
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Claim regarding Power Level (2) Freescale claims the waiver would open the door to waveforms with much higher emission levels than the current rules allow. Freescale gives an example of a “ 3 -hop” system that is 12. 5% of the time “gated on” and 87. 5% of the time “gated off. ” Why this is false: 1. Freescale’s example would clearly violate existing peak power limits and would not be permitted under the rules. 2. The waiver does not seek change to any existing power limits. 3. The waiver narrowly seeks FCC approval to have measurements for a 3 -band MB-OFDM system made under authentic operating conditions as it would actually be deployed. Submission 57
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 MB-OFDM Bandwidth Submission 58
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Claim Regarding Bandwidth Freescale claims the MB-OFDM waveform with “hopping” turned off may not meet the minimum 500 MHz bandwidth requirement. Why this is false: Each MB-OFDM symbol consists of 122 data-bearing sub-carriers (110 data + 12 pilot). Each sub-carrier is 4. 125 MHz from its nearest neighbor. 122 * 4. 125 MHz = 503. 25 MHz The MB-OFDM waveform meets the minimum 500 MHz bandwidth requirement at all times with or without frequency sequencing turned off. Submission 59
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 MB-OFDM Technology Advantages for Coexistence with Existing & Future Services Submission 60
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Benefits of MB-OFDM Systems Superior multipath performance of OFDM signals - Advantages of OFDM well-known to the industry and is used in Wi. Fi, Wi. Max, DSL, and other communications systems - Implementation much less complex than rake receiver required for impulse radios Innovative use of spectrum by partitioning into 528 MHz bands (band switching is fundamental to the design) - Reduces overall system complexity and power consumption due to lower bandwidth filters, analog-to-digital and digital-to-analog converters, etc. Enables CMOS friendly designs utilizing lower bandwidth baseband analog components - Interleaved 3 bands sequenced in time provides 1. 5 GHz spectrum use for improved frequency diversity - Enables suppression of strong narrowband interferers at the receiver by utilizing lower bandwidth analog baseband filters to limit compression of ADC (allows for designs with steep filter roll-offs), strong FEC coding, and digital signal processing techniques Submission 61
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Coexistence Benefits of MB-OFDM Systems (1) Spectral emissions advantages - Inherent properties of OFDM waveform produces lower out of band emissions than other types of UWB waveforms - Fine grained ability to sculpt emissions spectrum via software to meet worldwide regulatory requirements and extremely stringent coexistence requirements for some applications (operation within 1 foot of another wireless system or multiple radios in the same device) Added protection by dropping a whole band Software controlled ‘notch’ Submission 62
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Coexistence Benefits of MB-OFDM Systems (2) Why is spectrum flexibility critical? - Desire single solution to support worldwide regulations and interoperability Benefits of scaling (single SKU supports larger population of devices) Interoperability between devices in different regions (take a devices from the US to Europe and it can still work via software control mechanisms) Challenges: different frequency allocations worldwide may require different emissions limits – Indoor Wi. Max systems in Europe operate in 3. 4 -3. 6 GHz band – RAS bands in EU and Japan span large part of 3 -10 GHz bands (uncertain what regulatory bodies will require for these bands) 3260 -3267 MHz 3345. 8 -3352. 5 MHz 4950 -4990 MHz 6650 -6675. 2 MHz 3332 -3339 MHz 4825 -4835 MHz 4990 -5000 MHz 10. 6 -10. 68 GHz Submission 63
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Coexistence Benefits of MB-OFDM Systems (3) - Adapt to future allocations in the US and worldwide Don’t want to change UWB solution every time FCC allocates new spectrum which may operate in close proximity to a UWB device (impacts Tx and Rx) – New 3. 65 -3. 70 GHz NPRM could be useful for Wi. Max 4 G licensed allocations in the 3 -4 GHz band being considered in many countries – UWB solutions expected to operate in very close proximity to cell phones (within a few feet) and will likely be integrated into cell phones requiring greater protection No other UWB technology can achieve the level of spectrum flexibility provided by MB-OFDM and still meet stringent market requirements (low cost, complexity, power consumption) Submission 64
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 MAC Matthew B. Shoemake, Wi. Quest matthew. shoemake@wiquest. com Submission 65
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 No Vote Comments: MAC 1. As a (potential) IEEE standard, MB-OFDM must show that it is specified 2. 3. to work with the current IEEE 802. 15. 3 MAC, with minimum modifications. Public statements by both the companies and industry organizations referenced by the MB-OFDM proposers make it clear that they intend to develop and deploy a different MAC layer for their UWB products. This means that the companies and organizations that are referenced as backing the MB-OFDM proposal do not intend to use the 802. 15. 3 a MAC and therefore would not use the 802. 15. 3 a standard. It appears from multiple sources in the popular press as well as MBOASIG Press Releases that the MBOA MAC specification will be the only one certified by the Wi. Media to work in the MBOA ecosystem. If there is no intention to use the IEEE Std 802. 15. 3™-2003 and this is merely a blatant attempt to hijack the IEEE brand then I submit that the IEEE 802. 15 mandate withdrawal of the Merged Proposal #1 and confirm Merged Proposal #2. Submission 66
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Comment #1: Support of. 15. 3 MAC ? The Multi. Band OFDM PHY proposal is designed to work with the IEEE 802. 15. 3 MAC The proposers of the MB-OFDM solution are not aware of any issues that would prohibit operation of the MB-OFDM PHY with the IEEE 802. 15. 3 MAC If the commenter has specific technical concerns about interface of the MB-OFDM proposal to the IEEE 802. 15. 3 MAC, those detailed comments are solicited Submission 67
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Comment #2: There’s a different MAC IEEE can not control external organizations nor should that be our goal The goal of IEEE 802. 15. 3 a is to help organizations and companies by setting standards, not to force anything upon them The existence of multiple MACs should not be a distraction to the IEEE 802. 15. 3 a deliberations IEEE 802. 15. 3 a can do a service to the industry by confirming a new PHY standard Submission 68
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Comment #3: Wi. Media Certification The success of a standard often depends on interoperability testing and certification The IEEE 802 Standards body has abdicated responsibility for testing and certifying compliance of products Given that, there is no direct control over organizations such as Wi-Fi, DOCSIS, Wi. Media, Wi. MAX, UNH, etc. IEEE standards are intended to help companies and the population as a whole including organizations like Wi. Media The IEEE should be supportive and appreciative of external organizations that test and certify IEEE standards based products Submission 69
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Location Awareness Joe Decuir, MCCI joe@mcci. com Submission 70
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 No Vote Comments: Ranging and Location Awareness 1. MB-OFDM must have a clear, satisfactory solution to solve the location awareness problem. 2. MB-OFDM proposal is lacking in acceptable location awareness functionality. Submission 71
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 MB-OFDM PHY supports range measurements Ranging is one-dimensional location awareness The MB-UWB PHY supports ranging using Two Way Time Transfer algorithm (TWTT, 15 -04 -0050 -00 -003 a) PHY resources are described in 1. 7 of 15 -04 -0493 -00 -003 a - minimum resolution in the order of 60 cm - optional capabilities in the order of 7 cm The corresponding MAC resources are beyond the scope of TG 3 a. - see 15 -04 -0573 -00 -004 a-two-way-time-transfer-based ranging. ppt for an overview, as contributed to TG 4 a ranging subcommittee Applications and 2 -3 dimensional location awareness are above the MAC. - see 15 -04 -0300 -00 -004 a-ranging-rf-signature-and-adaptability. doc Submission 72
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 The MB-UWB PHY ranging support is only a part of location awareness. 802. 15 TG 3 a has seen very little location awareness work. 802. 15 TG 4 a is actively studying location awareness - see 15 -04 -0581 -05 -004 a-ranging-subcommittee-report Their consensus: location awareness transcends the PHY. - It is unrealistic for the PHY layer to construct or maintain 2 or 3 dimensional models of a device location. - Ranging and/or angle-of-arrival measurements are within the scope of the PHY (and MAC). They have studied several algorithms; no choice have been made. - TWTT uses minimal additional hardware - Angle-of-arrival requires multiple antennas Submission 73
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Coexistence David G. Leeper, Intel david. g. leeper@intel. com Submission 74
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 No Vote Comments: Coexistence 1. I suggest that the Merged Proposal #1 and Merged Proposal #2 merge 2. 3. 4. 5. 6. and become Merged Proposal #3. The clock frequencies and convolutional coder do not support a common signaling mode. I believe the common signalling mode is a way of providing interoperability and coexistence with other UWB devices. Merge Proposal #2 includes a provision for a base signaling mode that would allow multiple PHYs to coexist. In order for me to vote yes to Merge Proposal #1, there must be some type of coexistence mechanism. The MB-OFDM proposal must make accomodations for the appearance of PHYs in the same space, either by some sort of CSM or by a dual PHY. The best way to break through current dead lock is to adopt a dual PHY standard and let the market choose the better. Submission 75
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Coexistence or Harmonization (1) I suggest that the Merged Proposal #1 and Merged Proposal #2 merge and become Merged Proposal #3 - Customers have indicated preference for a single PHY standard The clock frequencies and convolutional coder do not support a common signaling mode. - CSM is not required for MB-OFDM and will add unnecessary cost and complexity. Clock frequencies & conv coder do not need to support CSM I believe the commons signaling mode is a way of providing interoperability and coexistence with other UWB devices - See above Merged Proposal @2 includes provision for a base signaling mord that would allow multiple PHYs to coexist. In order for me to vote yes on Merged Proposal #1, there must be some type of coexistence mechanism - See above Submission 76
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Coexistence or Harmonization (2) The MB-OFDM proposal must make accommodations for the appearance of PHYs in the same space, either by some sort of CSM or by a dual PHY - UWB PANs need low power, low cost, and support for Qo. S. CSM and/or dual PHYs would unnecessarily impair performance and add cost/power consumption The best way to break through the current deadlock is to adopt a dual PHY standard and let the market choose the better - Customers have indicated they prefer a single PHY to a dual PHY. The added expense and power consumption brings no benefit to the end user. In WLANs, having FH and DS PHYs only confused the market – no vendor built a viable dual-PHY standard product. Submission 77
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Multipath Performance Charles Razzell, Philips charles. razzell@philips. com Submission 78
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 No Vote Comments: Multipath Performance 1. The performance in range and survivability even in moderate multipath is absolutely dismal. 2. Parallel and serial transport of the same data rate in the same bandwidth can be equally efficient 3. 4. 5. 6. against white noise, but the performance with multipath is materially weaker. With direct-sequence spreading, the difference is even greater for multipath interference. A single carrier system with “rake” receiver processing will receive and process more power from combined propagation paths than is possible with N multiple parallel paths each carrying 1/N of the message load (before considering the benefits spectrum spreading). The [direct sequence] spreading causes the multipath to appear as an interference signal in particular chips. Errors in some individual chips reduces the power sum or Boolean sum relative to no errors, but does not prevent successful evaluation of the data value carried by that sequence. This tolerance for chip errors is a property not found in OFDM which attempts to get this benefit with FEC. Some fraction of corrupted packets might be saved by FEC, but this will be for those packets with a small number of errors. With MB-OFDM, there will be coverage holes where cancellation fades have occurred, and these will no be helped by more power or better error correction. As a rough estimate based on tests at 5 GHz, there may be 5% of locations where a satisfactory decoding cannot be achieved. At such holes, moving the antenna location a small distance may cause satisfactory signal to reappear. The recent changes to the proposal to map data bits on the guard tones have shown that adding more diversity to the bit-tone mapping could help to improve the poor multipath performance of MBOFDM. Please come up with a way that can add more diversity to these mappings (especially at higher rates, > 200 Mps) in order to compensate for the degraded performance caused by the Rayleigh-distributed multipath fading. Since the 6 d. B degradation (@480 Mbps) identified in various other document has been improved by these recent modifications, please derive the new amount of degradation (e. g. 5 d. B? ) based on the new mappings for the various data rates proposed. Submission 79
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Multipath Performance System performance has been compared since March 2003 in standardized channel models CM 1 -4 MB-OFDM has consistently performed well even in the most severe channel models Range AWGN LOS: 0 – 4 m CM 1 NLOS: 0 – 4 m CM 2 NLOS: 4 – 10 m CM 3 RMS Delay Spread: 25 ns CM 4 110 Mbps 21. 4 m 12. 0 m 11. 5 m 10. 9 m 200 Mbps 14. 6 m 7. 4 m 7. 1 m 7. 5 m 6. 6 m 480 Mbps 9. 3 m 3. 2 m 3. 0 m N/A The results speak for themselves: link distance at 110 and 200 Mbps is hardly impacted by multipath Submission 80
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Multipath Performance cont. So, why the counter claims? - Frequency diversity is not inherent in OFDM - If all available sub-carriers are used to transmit independent information (without redundancy), each bit of information is subject to Rayleighdistributed narrow-band fading - In order to overcome this frequency-domain spreading is needed, which usually implies some redundancy Why is this not a big issue? - Frequency domain spreading in MB-OFDM is provided by a mixture of repetition coding and convolutional coding - Even at 480 Mbps (the worst case), sufficient spreading gain is available to obtain 3 m link distances in NLOS channels. - At 110 Mbps, the scope for spreading has dramatically improved to approximately 6, partitioned as a factor 3 for FEC and 2 for repetition coding. Submission 81
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Time-to-Market Jim Lansford, Alereon jim. lansford@alereon. com Submission 82
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 No Vote Comments: Time-to-Market 1. The earliest availability of silicon for this proposal is 2005. An 2. 3. 4. alternative proposal has ICs available today, which have the ability to be adapted to the precise protocols laid down by the standard, within a very short time of the standard being issued. The DSUWB solution has been shown to work and is commercially available. I will not vote for Merge Proposal #1 unless it can be demonstrated, with real silicon, that it meets the PAR. At least one competing standard has recently geared up it process and is threatening to snatch away a sizeable chunk from this standard’s targeted market. MB-OFDM projects its chipset availability in 2005 (at the earliest). Its counterpart has already cranked out silicon out of the development cycle. It would be interesting to see some, ANY, working hardware demonstrating feasibility of the solution -- even a breadboard, because at this late date I can no longer accept Power. Point Engineering. Submission 83
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 TTM Issues TTM difference between proposals is months, not years. - MB-OFDM chipsets operating at 480 Mb/s have been demonstrated over the air, so in some ways, MB-OFDM is ahead of DS-UWB, not behind MB-OFDM silicon will be available in the market before a draft could complete the balloting process - Demonstration silicon is available now - Early products will be available in months - History shows early chipsets have to be re-spun for standards compliance anyway PHY Submission 84 MAC
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Summary of comments related to market timing (1) The earliest availability of silicon for this proposal is 2005. An alternative proposal has ICs available today, which have the ability to be adapted to the precise protocols laid down by the standard, within a very short time of the standard being issued. - The time difference in availability of silicon is much smaller than has been stated by MBOA opponents. Commercial availability of chipsets will differ only by a few months, not years. MB-OFDM chipsets will be in the market well before the draft specification could be completed. Note that the IEEE 802 -SEC has taken the position of being against companies releasing chipsets and calling them "pre-compliant" when a standard has not completed letter ballot and sponsor ballot. The DSUWB solution has been shown to work and is commercially available. I will not vote for Merge Proposal #1 unless it can be demonstrated, with real silicon, that it meets the PAR. - At least one MB-OFDM company has demonstrated MB-OFDM proposal compliant silicon operating at 480 Mb/s over the air. This chipset meets the PAR. May we count on you switching your vote? Submission 85
November, 2004 doc. : IEEE 802. 15 -04/0641 r 0 Summary of comments related to market timing (2) At least one competing standard has recently geared up it process and is threatening to snatch away a sizeable chunk from this standard’s targeted market. MB-OFDM projects its chipset availability in 2005 (at the earliest). Its counterpart has already cranked out silicon out of the development cycle. - Again, the time difference in availability of silicon is much smaller than has been stated by MBOA opponents. Commercial availability of chipsets will differ only by a few months, not years. MB-OFDM chipsets will be in the market well before the draft specification could be completed. It would be interesting to see some, ANY, working hardware demonstrating feasibility of the solution -- even a breadboard, because at this late date I can no longer accept Power. Point Engineering. - We agree. Several MB-OFDM companies now have working silicon. As mentioned in a previous response, working silicon operating over the air at 480 Mb/s has been demonstrated which meets the PAR. We are awaiting something other than Powerpoint for a DS-UWB demonstration Submission 86
- Slides: 86