研究目的
Investigating and mitigating injection pulling issues in multicore RF-SoC through digital fractional division.
研究成果
The paper demonstrates a successful mitigation of injection pulling in multicore RF-SoCs through digital fractional division. The proposed method allows oscillators to resonate at frequencies far away from each other, virtually eliminating the injection-pulling effect. The experimental verification in a two-channel system shows that the injection-pulling spurs are significantly reduced, making the method attractive for wireless TXs.
研究不足
The study is limited to a two-channel system in 65-nm digital CMOS. The effectiveness of the proposed method in systems with more channels or different technologies is not explored. Additionally, the physical separation and shielding methods may not be feasible in all integration scenarios due to cost and area constraints.
1:Experimental Design and Method Selection:
The study involves designing a two-channel system in 65-nm digital CMOS, each channel comprising a high-swing class-C oscillator, frequency divider, and phase rotator. The methodology includes breaking the integer harmonic frequency relationship of victims/aggressors within and between the RF transmission channels using a digital fractional divider based on phase rotation.
2:Sample Selection and Data Sources:
The experiment uses a two-channel system with ~8-GHz oscillators realized in 65-nm digital CMOS. The oscillators are placed 200 μm apart on the same CMOS die.
3:List of Experimental Equipment and Materials:
The setup includes LC-tank oscillators, frequency dividers, phase rotators, and a 65-nm digital CMOS process for IC fabrication.
4:Experimental Procedures and Operational Workflow:
The procedure involves measuring the injection-pulling spurs at the oscillator output versus separation of the two carriers for three different test cases (common substrate, diced substrate, and diced substrate with metal shield).
5:Data Analysis Methods:
The analysis includes calculating the injection current strength for different measurements and validating the results against a behavioral model.
独家科研数据包,助您复现前沿成果,加速创新突破
获取完整内容