Effect of TX Pulse Mismatch

1. Project: IEEE P802.15 Working Group for Wireless Personal Area Networks (WPANs) Submission Title: [Effect of TX pulse mismatch] Date Submitted: [27 February, 2006] Source: [P. Orlik] Company [Mitsubishi Electric.] Address [201 Broadway; 8th floor; Cambridge, Massachusetts 02139; USA] Voice:[(617) 621-7570], FAX:[], E-Mail: [porlik@merl.com] Re: [802.15.4a.] Abstract: [A summary of teleconference activity between Garden Grove and Vancouver.] Purpose: [To resolve modulation accuracy and EVM comments on draft D1.] Notice: This document has been prepared to assist the IEEE P802.15. 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 P802.15. P. Orlik, MERL

2. Background • Current draft only specifies that a pulse shape have a cross correlation coefficient greater that 0.7 when correlated with a root raised cosine pulse. • This requirement maybe too loose • Several comments on a definition of modulation accuracy and Error Vector Magnitude (EVM) have been received P. Orlik, MERL

3. Draft definition of “Golden Pulse” • Golden pulse = root raised cosine (RRCOS) • Rolloff factor b = 0.6 • Pulse width = 2 ns • 500 MHz @3dB • Rolloff factor effects the slope of the spectrum from passband to stopband P. Orlik, MERL

4. Peak Correlation Tests • How do pulses that do not match the golden pulse fair against the 0.7 cross correlation metric? • Tested RRCOS pulses with various rolloff and 3dB bandwidths • Bandwidth 400 – 500MHz • Rolloff 0 – 1.0 • Tested Butterworth pulses with various filter order and 3dB bandwidths • Bandwidth 400 – 500MHz • Order [4 8 12 20] – higher order implies steeper slope in transition region P. Orlik, MERL

5. Peak Correlation Coefficients (RRCOS,RRCOS) P. Orlik, MERL

6. Peak Correlation Coefficients (RRCOS, Butterworth) P. Orlik, MERL

7. Peak Correlation test Results • Summary – All tested pulse shapes easily achieve the 0.7 cross correlation metric P. Orlik, MERL

8. Cross Correlation Shape Test • What about the cross correlation function itself? • Has implications on the synchronization requirements • High sidelobes lead to sync errors • Very steep fall off from peak lead to tracking difficulty. • Tested both RRCOS and butterworth pulse shapes at 400MHz 3dB Bandwidth • 400MHz is the worst case peak correlation point from earlier test. P. Orlik, MERL

9. Cross Correlation (RRCOS, RRCOS) 500MHz pulse correlated with a 400MHz RRCOS pulse Rolloff factor (Beta) varying Sidelobes < 0.2 Correlation Function is above 0.7 for about 1 ns P. Orlik, MERL

10. Cross Correlation (RRCOS, Butterworth) 500MHz pulse correlated with a 400MHz Butterworth pulse Filter Order varying Sidelobes [0.2 – 0.4] Correlation Function is above 0.7 for about 0.6 - 1 ns P. Orlik, MERL

11. Summary • Width of the correlation peak ranges from 0.6 to 1ns • Sizes of the sidelobes may be an issue with some pulse shapes • Suggest further specifying transmitted pulse shape by • Requiring main lobe of cross correlation function remain above 0.7 for sum length of time (0.6 – 1 ns seems easy to achieve) • Could also place limit on peak sidelobe: around 0.2 – 0.4 • A value of 0.2 still admits a low order butterworth pulse shape but prohibits higher orders. • The additional requirements should serve to close out comments: 153, 471, 874, 142, 145, • Along with corrected formula for RRCOS comments: 436, 643 693 go away as well. P. Orlik, MERL