Fast SAR Methods for EM Exposure Evaluation of Wireless Devices

Transcription

1 of Wireless Devices Mark Douglas, Chung-Huan Li, Niels Kuster IT IS Foundation Sami Gabriel Vodafone Group C. Bucher, D. Iliev, J. Kastrati, C. Leubler, M. Meili, K. Pokovic Schmid & Partner Engineering AG

2 Background Detailed SAR test procedures for wireless devices are defined in international standards - IEEE 1528 and IEC Strong emphasis is placed on: - high repeatability - high reproducibility - conservative results with respect to real human exposure Details of test procedures provided for - Probe: geometry, orientation, isotropy - Phantom: shape, dielectric parameters, tolerances - Uncertainty, validation, daily checks Evaluation time is long due to long test time and many tests

3 Fast SAR Methods New SAR methods have emerged - Use special techniques or algorithms to reduce test time - SAR test time reduced by orders of magnitude - Uncertainty can be higher than full SAR methods International standards committees integrating fast SAR methods into standards Principles for using fast SAR methods: 1. Use fast SAR methods to quickly find highest SAR test cases - e.g., right vs left side of head, position of mobile phone, mode (GSM, UMTS, ) 2. Use full SAR methods for final compliance evaluation - Keep high reliability of full SAR methods

4 Fast SAR Systems: relative assessment Fast SAR systems do not comply with all requirements for full SAR systems - e.g., measurement resolution Therefore, absolute SAR values are not used. Full SAR systems must be used for final compliance assessment Fast SAR systems are used to find high SAR test configurations Relative uncertainty is applied for fast SAR systems - Reduced compared with absolute uncertainty

5 How to ensure that max. SAR is tested? Full SAR method must be used to test highest SAR configuration Due to fast SAR measurement uncertainty, highest measured value may not be highest true SAR value Want confidence that maximum SAR configuration is tested with full SAR method

6 Example Two fast SAR measurements are made: - SAR 1,fast = 1.25 W/kg - SAR 2,fast = 1.1 W/kg - U fast = 10% Is SAR 1,fast > SAR 2,fast? - We are > 50% confident, but are we 95% confident?

7 Solving the Example Difference between two Gaussian random variables is a Gaussian random variable. - D = SAR 1,fast SAR 2,fast = U D = [U 1,fast 2 + U 2,fast2 ] = [(1.25*10%) 2 + (1.1*10%) 2 ] = What is the confidence for D > 0? - z a = 0.15 / = (1 a) = 82% We do not have 95% confidence - Need full SAR measurements for both

8 Solution For those test configurations where full SAR testing is required (at 95% confidence level), the fast SAR value, SAR i,fast, satisfies: i.e., D -1.64U D 0 ( SAR 2 2 1, fast - SARi, fast ) -1.64U rel, fast SAR1, fast + SARi, fast 0 Solving for SAR i,fast : SAR ( ) 2 B - 1 i, fast SAR1, fast B - where B = 1-1 ( 1.64 ) 2 U rel, fast

9 Recommended Procedure 1. Do fast SAR tests on different test configurations. Group them by mode = frequency band + modulation (e.g., GSM900, GSM1800) 2. For each mode, do full SAR measurement for highest SAR configuration only 3. For the mode with the highest full SAR (e.g., GSM1800), do full SAR measurements if: SAR ( ) 2 B - 1 i, fast SAR1, fast B - 4. For the other modes (e.g, GSM900), use similar statistical methods to check if step 3 is needed

10 Comparison of 3 systems Full DASY measurement: - Area scan + zoom scan: minutes Fast DASY measurement: - Area scan only: 5 minutes - 1g/10g average SAR estimated from area scan [1] isar: - Array of probes, test time: < 1 second - Measurements on a conformal surface - 1g/10g average SAR calculated from area scan [1] [1] M.Y. Kanda, et al., Faster Determination of Mass-Averaged SAR from 2-D Area Scans, IEEE Transactions on Microwave Theory and Techniques, vol. 52, no. 8, pp , August, 2004.

11 Results: 900 & 1800 MHz for fast DASY52 Mobile phone with 8 different configurations - Carry cases, wall charger, car charger Comparison of DASY52 and fast DASY52 Standard deviation of difference: db at 900 MHz db at 1800 MHz Full SAR testing needed for - 2 cases at 900 MHz - 1 case at 1800 MHz Total time savings: 56%

12 Results: 900 & 1800 MHz for isar Mobile phone with 8 different configurations - Carry cases, wall charger, car charger Comparison of DASY52 and isar Standard deviation of difference: db at 900 MHz db at 1800 MHz Full SAR testing needed for - 2 cases at 900 MHz (75% savings) - 1 case at 1800 MHz (87.5% savings) Total time savings: 81%

13 Conclusions The benefits of using fast SAR methods has been clearly demonstrated using three measurement systems. Standard deviation of differences for one mobile phone were within 0.2 db for all eight configurations and two frequency bands tested. Improvements in the measurement time of up to 80% were achieved with the use of fast SAR methods. The methodology provides a high degree of confidence that the use of fast SAR methods results in the same reported compliance values as the conventional approach.

14 Acknowledgment The authors would like to thank the Mobile Manufacturers Forum for their generous support of the project.