一站式电子元器件采购平台

华强商城公众号

一站式电子元器件采购平台

口袋里的移动库存

华强商城移动APP

口袋里的移动库存

元器件移动商城,随时随地采购

华强商城M站

元器件移动商城,随时随地采购

半导体行业小灵通

半导体八阿哥公众号

半导体行业小灵通

专注电子产业链,坚持深度原创

华强微电子公众号

专注电子产业链,
坚持深度原创

电子元器件原材料采购信息平台

华强电子网公众号

电子元器件原材料采购
信息平台

HFTA-07.0:精密参考时钟使用的时钟和数据恢复电路-H

来源:http://www.elecfans.com/article/83/116/2009/2009042252436.html 发布时间:2017-07-27

摘要: [导读] Applications for clock data recovery (CDR) abound in telecommunications, optical transceivers, data and storage area networks, and wireless products. The benefits of CDR technology are increasingly important as designs require much greater


Applications for clock data recovery (CDR) abound in telecommunications, optical transceivers, data and storage area networks, and wireless products. The benefits of CDR technology are increasingly important as designs require much greater bandwidth, and as allocation and spectral bandwidth usage increase. Moreover, vendors and their products are migrating from parallel to serial interfaces for system- and board-level interfaces.

In recent years, the increased use of CDR technology has brimmed from the need to handle wider parallel bus widths across backplanes while managing clock and data skew at the receiver. Additionally, routing these signals can be difficult because they consume board space and power, and require multilayer routing schemes to manage signals and line termination. EMI generation from the use of high bit-width data buses is also a concern.

CDRs are extremely important due to the advent of new communication technologies, improvements in electrical signal processing, and the need to send multigigabit electrical signals across FR-4 and backplanes, optical, and wireless media. Communication techniques that combine clock and data prior to transmission are not new. The combination of clock and data ensure that the clock and data signals always arrive at the same time. The trick, however, is the separation of the clock and data at the receiver. This is accomplished by the CDR circuitry. Products that take data from a parallel to a serial format or vice versa are called serializers/deserializers (or "SerDes" for short). These products generally have CDR blocks to deserialize the serial data stream.

This article examines the component blocks of CDR needed to implement a successful CDR in high-speed serial-communication link applications. An overview of a typical high-speed serial-communication link is provided with respect to how data is transformed and recaptured over the link. Different CDR schemes are discussed with respect to the general CDR topology. Also, special attention is given to the reference oscillator's role on both the transmit and receive sides of the link.

Clock and Data Recovery in High-Speed Serial Communications

Figure 1 provides a basic diagram of a high-speed serial-communications link. Parallel data (bits b1, b2, b3,... bn) arrives at the transmission serializer at frequency ft. Within the serializer, data is converted from the parallel format to a serial format. A serial-bit stream is developed with a minimum bit rate equal to n x ft, where n is the total number of parallel data bits. The resulting frequency (bit rate) can be higher than ft depending on whether the data is encoded to meet channel requirements for bit error rate (BER) performance or to provide rich transition content at the receive-side CDR. Reed-Solomon forward-error correction (FEC) and 8B10B encoding are, respectively, each examples of channel encoding or creating rich transition content at the receive CDR. This serial data is then readied for transmission to the channel and sent through to the receiver, finally reaching the deserializer. This basic communications block applies to data whether it is transmitted through fiber, air, or across backplanes.

Figure 1. Clock and data recovery form the basis for high-speed serial communications links.
Figure 1. Clock and data recovery form the basis for high-speed serial communications links.

Timing (clocking) is critical in CDR applications. During system design, the designer determines how to drive data from a parallel format to a serial format for transmission and reception through the channel with pending channel distortion of the transmitted signal. Minimizing the effects of the design's degradation on the data signal is important for protecting signal-to-noise ratio and maintaining BER performance. For example, in digital-transmission schemes across backplanes, the system's jitter performance is extremely important as high-speed electrical signals traverse various lengths (FR-4 and the backplane), thus causing signal degradation in terms of signal-level and time-variant distortions.

At the core of clock-data recovery is a phase-locked loop (PLL)-based circuit that in some cases could be digitally based. Figure 2 is the basic PLL block diagram that could be used in the serializer or transmission side of the communication link shown in Figure 1. The PLL block includes a phase-frequency detector (PD), filter (LPF), voltage-controlled oscillator (VCO), and a divide chain (1/n). The divide chain is used to provide comparable frequency inputs to the PD. In this manner, the VCO's output is phase-aligned to the very stable reference input, VREF. This PLL block's purpose is to multiply the reference frequency by a fixed amount (n), which would be the VCO's natural frequency. In most cases, VREF would be quartz based, providing a high degree of stability and accuracy with great phase-noise characteristics. Additionally, this reference may be temperature compensated or voltage compensated, depending on the required application or system requirements. In SONET-based applications, this reference may meet a certain stratum level (i.e., stratum levels 3, 3E, or 4).

Figure 2. Clock multiplication applications are driven by PLLs.
Figure 2. Clock multiplication applications are driven by PLLs.

At the receive side, the CDR PLL block takes on a slightly different look to address the need to retrieve both clock and data. As shown in Figure 3, the combined clock/data signal comes into the PLL block through a buffer that feeds two distinct paths. One path feeds the data-decision (DEC) block while the second path feeds the clock-recovery blocks. The clock-recovery blocks look very much like Figure 2's PLL block minus the 1/n block. The recovered clock from the VCO is used as a sampling input to the DEC, a feedback to the phase-frequency detector, and also to feed system-timing requirements downstream. In the case of Figure 1, this recovered clock is divided down to the parallel-clock frequency to drive the deserializer block.

Figure 3. Modifications to the basic PLL block are used to implement CDR circuits.
Figure 3. Modifications to the basic PLL block are used to implement CDR circuits.

Reference Oscillators in Clock/Data Recovery

The reference oscillators shown in the diagrams illustrate a VCO being applied to the input of the oscillator. This voltage control is established by the LPF stage. Typically, VCOs or voltage-controlled crystal oscillators (VCXOs) can be used as the loop oscillator, as shown in Figure 3. Primary to its role, the loop oscillator is required to track the frequency deviation of the incoming clock/data. Additionally, it provides this clock to other components downstream from the CDR (deserializer). This is accomplished by the output of LPF, which drives the voltage-control input of the VCO or VCXO.

In CDR applications for telecom, wireless, and datacom, the incoming data signal plus clock should have relatively stable frequency characteristics. This assumes the transmit clock meets certain accuracy and stability specifications. On the receive side, the design precludes for the minimum and maximum accuracy/stability. In the event the transmit clock frequency is expected to be with ±50ppm of a stated frequency, the receive clock would have frequency-adjustment capability of a minimum ±50ppm. For design purposes, however, consider a frequency adjustment capability somewhat larger than ±50ppm. This extended frequency-adjustment capability accommodates any additional signal-frequency distortion by the channel or disruptions in communications.

Although the PLL attempts to drive toward a static condition, meaning that frequency lock has been established, there could be conditions where the voltage-controlled input may move at higher rates than desired. The LPF bandwidth dictates the maximum rate at which the PLL can maintain lock. Ultimately, the role of the receive VCO (or VCXO) is to track and reproduce the recovered clock.

In the absence of a data/clock input into the CDR, the CDR is required for a specified time to provide a reference signal for any downstream communications requirements (i.e., the deserializer)

In some applications, a combination VCO/VCXO would be used. In Figure 4, VCO/VCXO allow for at least two benefits to the common CDR configuration. First, the addition of the VCXO allows fast regulation of the VCO frequency to match that of the expected clock/data signal. The VCXO frequency is chosen to match the expected clock frequency range. For example, a wideband VCO can require thousands of samples to lock onto the incoming data stream. The additions of the VCXO and lock-detection circuit ensure that the VCO maintains a certain operating frequency, and help provide a more predictable lock time in startup conditions. Second, the addition of the VCXO is helpful if the clock/data input has been lost over an extended period of time. In the absence of a clock/data signal, the system would refer to the very stable quartz-based oscillator (VREF) to provide holdover until the clock/data signal has been recovered from a loss of signal (LOS). Holdover is a specification that applies to the reference clock's ability to hold a certain accuracy over a certain period of time (for example, ±4.6ppm over 24 hours).

Figure 4. A modification on the basic CDR block provides easy locking on the incoming clock/data stream.
Figure 4. A modification on the basic CDR block provides easy locking on the incoming clock/data stream.

Conclusion

A variety of solutions address clock/data recovery and retiming, serializers and deserializers, clock generators, and TCXOs for communication applications. These devices allow designers to develop circuits for frequencies ranging from 10MHz up to 10GHz, and support applications ranging from GSM to OC-192 and above. As designs increasingly require greater bandwidth, CDR technology is ideal for telecommunications, optical transceivers, data and storage area networks, and wireless applications.

声明:本文观点仅代表作者本人,不代表华强商城的观点和立场。如有侵权或者其他问题,请联系本站修改或删除。

微信扫一扫,一键转发

关注“华强商城“微信公众号

相关阅读
  • 加深对电路中高输入阻抗和低输出阻抗的理解

    1、 输入阻抗:输入阻抗是指电路输入端的等效阻抗。将电压源u添加到输入端并测量输入端的电流I,则输入阻抗Rin为u/I。您可以将输入端视为电阻器的两端。输入阻抗器的电阻为输入阻抗。输入阻抗与普通电抗元件没有区别。它反映了阻断电流的效果。

  • 华强电子元器件网转型中电子元器件柜台将走向何方?

    华强电子元器件网转型的具体规划是,公司将逐步将实体市场的运营调整为以电子和数字产品为主的综合商业运营商,强化体验购物功能。

  • 在电子元器件采购商城购买电子元件时,您需要了解哪些知识

    在电子元器件采购商城购买元件也需要一些知识。我们需要注意很多方面,例如型号、交货日期、包装和批次。在购买电子元件时,我们不仅要知道它是什么型号,还要知道详细的规格,然后才能购买正确的元件。买部件就像买衣服一样。我们还需要辨别真伪。我们必须确保产品是正品和原创的。如果你不知道,往下看。以下是您需要了解的有关购买电子元件的知识。

  • 基于buck电路的太阳能光伏电池模拟器性能改进试验

    尽管技术日益发达,但世界也面临着能源危机,各国对太阳能发电的研究也在不断发展。然而,由于太阳能电池研究成本高,太阳能电池模拟器应运而生。目前,我们正在设计一种基于buck电路的太阳能电池模拟器,采用arm控制,并加入电流PI控制方法,以提高系统的动态性能和稳态精度。

  • 传感器制造商研究半导体材料,使得气体传感器满足各行业需求

    传感器有许多实际应用,几乎涵盖所有行业。目前,几乎所有行业都在朝着更小的设备形状因素发展,传感器的尺寸也在减小。在本文中,我们将介绍三种微型传感器,生物传感器,气体传感器和加速度计及其实际应用。在电子产品尺寸不断减小的世界中,它们显示了向小型和低功耗传感器的持续发展趋势。

  • 中小企业、科研机构去电子元器件采购商城需要注意哪些事项?

    中小型企业、科研机构和电子元器件采购商城应该如何选择购买电子元件?如今,华强北商城向广大用户介绍并注重电子元器件的购买。

  • 贴片电容的价格与什么有关?贴片电容未来发展好吗?

    如果贴片电容产品质量和价格摆在我们面前,我相信大多数制造商仍然会选择低价格贴片电容。在当今的市场环境下,各行各业的竞争非常激烈,这也给我们带来了很多考验。为了生存,企业经常打价格牌。贴片电容价格已经降下来了。贴片电容质量和服务有改进吗?

  • 电子元器件采购平台:电子元器件采购需要做哪些工作?

    一般来说,电子元器件采购应做以下工作:1.电子元器件供应商开发:根据公司需要寻找对方的材料供应商,需要熟悉相关材料行业的生产厂家、各种材料的电子元器件制造商品质量状况、材料的实时市场情况、未来几个月的市场情况,以及物料的整体交付状态...

  • 温度传感器在火灾报警系统中的应用

    一线总线技术具有节省I/O资源、结构简单、成本低、总线扩展和维护方便等特点。因此,它被广泛应用于分布式温度传感器系统中。生产的单总线智能温度传感器DS18B20是一种典型的采用单线总线技术的产品。DS18B20是同一系列或不同系列的单线设备,因此主机必须能够决定如何正确访问单线总线上的每个设备。

  • 计流量传感器测量管的结构是什么?有什么分类?

    高精度质量流量计的流量传感器增加了测量管的半径,改善了应力分布,减少了疲劳损伤,提高了抗振动干扰能力。因此,流量传感器的测量管的形状和结构多种多样。本节仅从不同角度讨论分类:根据测量管流向和工艺管道流向的布置,有两种方式:平行和垂直。

  • 碳膜、金属膜和金属氧化物膜的电阻差

    金属膜电阻(金属射电阻)也采用真空喷涂技术喷涂在瓷棒上,但用金属膜(如镍铬)代替碳膜,并在金属膜车上的螺旋图案上形成不同的电阻值,在瓷棒的两端喷上贵金属。虽然它比碳膜电阻昂贵,但它具有噪声低、稳定性好、受温度影响小、精度高等优点。因此,金属膜电阻被广泛应用于先进的音响设备、计算机、仪器仪表、国防和航天设备中。

调查问卷

请问您是:

您希望看到什么内容: