孟令硕,胡友德,韦援丰,等.一种应用于MIPI D-PHY的高速SLVS驱动电路[J]. 微电子学与计算机,2023,40(7):89-96. doi: 10.19304/J.ISSN1000-7180.2022.0646
引用本文: 孟令硕,胡友德,韦援丰,等.一种应用于MIPI D-PHY的高速SLVS驱动电路[J]. 微电子学与计算机,2023,40(7):89-96. doi: 10.19304/J.ISSN1000-7180.2022.0646
MENG L S,HU Y D,WEI Y F,et al. A high speed SLVS driver circuit for MIPI D-PHY applications[J]. Microelectronics & Computer,2023,40(7):89-96. doi: 10.19304/J.ISSN1000-7180.2022.0646
Citation: MENG L S,HU Y D,WEI Y F,et al. A high speed SLVS driver circuit for MIPI D-PHY applications[J]. Microelectronics & Computer,2023,40(7):89-96. doi: 10.19304/J.ISSN1000-7180.2022.0646

一种应用于MIPI D-PHY的高速SLVS驱动电路

A high speed SLVS driver circuit for MIPI D-PHY applications

  • 摘要: 介绍了一种采用SMIC 14nm FinFET工艺制造的应用于MIPI D-PHY的高速SLVS驱动电路. 分析了传统电平提升结构下拉网络延迟较大的问题,通过使用中和电容降低回馈噪声影响,同时交叉耦合结构中加入调整支路以平衡上拉网络与下拉网络,实现了输出节点更快速率的充放电.针对驱动电路转换期间寄生电容造成的信号过冲问题,通过加入耦合电容减弱转换期间流经晶体管的电流,降低输出信号超调量,减少开关管在转换期间造成的共模扰动.传统驱动方案中开关管和尾电流管源漏电压过低导致面积开销过大,通过将开关控制路径与共模反馈调节路径并联以节省晶体管面积.为了补偿信道对信号高频分量的衰减,通过检测信号跳变并生成脉冲信号驱动预加重电路来增加信号高频分量,从而提高眼图质量,实现更高频率的数据传输. 测试结果表明,在1.8 V电源电压下,输出共模电平200 mV,电压摆幅200 mV,工作速度高达4 Gbps,总功耗约4.25 mW,能量效率1.0625 mW/Gb/s. 在3.6 Gbps和4 Gbps速率下结构优化后输出眼图水平张开度分别提高了0.08 UI、0.07 UI.

     

    Abstract: This paper introduces a high-speed SLVS driver circuit for MIPI D-PHY, which is manufactured by SMIC 14nm FinFET process. The problem of the delay of pull-down network in traditional level lifting structure is analyzed. By using neutralization capacitor to reduce feedback noise, and adding adjustment branch in cross coupling structure to balance pull-up network and pull-down network, the output node can be charged and discharged more quickly. To solve the problem of signal overshoot caused by parasitic capacitance during the conversion of the drive circuit, the coupling capacitance is added to weaken the current flowing through the transistor during the conversion, reduce the output signal overshoot, and reduce the common mode disturbance caused by the switch during the conversion. In the traditional drive scheme, the source leakage voltage of the switch and tail current tube is too low, which leads to excessive area overhead. The area of the transistor is saved by paralleling the switch control path with the common mode feedback adjustment path. In order to compensate the attenuation of the channel to the high frequency component of the signal, the high frequency component of the signal is increased by detecting the signal jump and generating a pulse signal to drive the pre emphasis circuit, so as to improve the eye map quality and achieve higher frequency data transmission. The test results show that under the 1.8 V power supply voltage, the output common mode level is 200 mV, the voltage swing is 200 mV, the working speed is up to 4 Gbps, the total power consumption is about 4.25 mW, and the energy efficiency is 1.0625 mW/Gb/s. The horizontal opening of the output eye map after structure optimization at 3.6 Gbps and 4 Gbps speeds increased by 0.08 UI and 0.07 UI respectively.

     

/

返回文章
返回