首页> 外文会议>Solid-State Circuits Conference - Digest of Technical Papers, 2009. ISSCC 2009 >A 14mW 5Gb/s CMOS TIA with gain-reuse regulated cascode compensation for parallel optical interconnects
【24h】

A 14mW 5Gb/s CMOS TIA with gain-reuse regulated cascode compensation for parallel optical interconnects

机译:具有用于并行光互连的增益重用调节共源共栅补偿的14mW 5Gb / s CMOS TIA

获取原文

摘要

Short-distance parallel optical links are poised to replace copper interconnects in high throughput links between computing nodes. In the receiver front-end of such systems, one- and two-dimensional monolithically integrated arrays of power- and area-efficient TIAs are required. Differential signaling is preferred for superior supply rejection while power consumption per amplifier unit should be minimized to avoid heat dissipation issues and to maintain the energy efficiency (J/bit) advantage over copper interconnect. On the other hand, lower gain is acceptable due to the intended short-haul application. Multi-Gb/s CMOS TIAs suffer from a fundamental BW limit at the input. The TIA input sees significant extrinsic capacitive loading (CIN), not only from the photodiode (CPD) which is dominant, but also from the ESD structures (CESD) and the offset-cancellation circuit (COC), which is needed due to finite extinction ratio of the transmitter laser diodes. The impact is more severe in CMOS as the input impedance of the circuit (1/gm) is relatively high due to the low gm/lD of CMOS technology. The regulated cascode approach is an effective BW-extension technique that desensitizes the circuit to input capacitance by reducing the TIA input impedance through negative feedback. However, in its conventional implementation, power dissipation is increased due to the additional broadband feedback amplifier while stability can be compromised by phase shifts. In this work, the voltage gain available from the common-gate TIA front-end is reused for the compensation rather than employing a dedicated amplifier; with the aim of demonstrating a more power- and area-efficient approach to CMOS BW extension.
机译:短距离并行光链路有望取代计算节点之间高吞吐量链路中的铜互连。在这种系统的接收机前端,需要功率和面积效率高的TIA的一维和二维单片集成阵列。最好采用差分信号传输,以实现出色的电源抑制性能,同时应将每个放大器单元的功耗降至最低,以避免散热问题并保持铜互连线的能效(J / bit)优势。另一方面,由于预期的短途应用,较低的增益是可以接受的。多Gb / s CMOS TIA在输入端受到基本带宽限制。 TIA输入不仅受到主要的光电二极管(C PD )的影响,而且还受到ESD结构(C ESD )和偏移消除电路(C OC ),这是由于发射器激光二极管的消光比有限而需要的。由于较低的g m / l D <,电路的输入阻抗(1 / g m )相对较高,因此在CMOS中的影响更为严重。 / sub> CMOS技术。稳压共源共栅方法是一种有效的带宽扩展技术,该技术通过通过负反馈降低TIA输入阻抗来使电路对输入电容不敏感。然而,在其常规实现中,由于附加的宽带反馈放大器而增加了功耗,而相移会损害稳定性。在这项工作中,可从共栅TIA前端获得的电压增益重新用于补偿,而不是使用专用放大器。目的是展示一种更省电,更省电的CMOS BW扩展方法。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号