SFF-8410：高性能SCOPP测试与性能规范

需积分: 9 18 浏览量更新于2024-07-16 收藏 1.16MB PDF 举报

SFF-8410.PDF 是由 SFF（Small Form Factor）委员会制定的一个标准，该委员会后来成为了SNIA（Storage Networking Industry Association）的SFF Technology Affiliate（TA）技术工作小组（TWG）。这个标准是在SNIA的指导下形成的，旨在规范和优化小型形态系统的硬件设计和性能。SFF-8410 的第 16.1 版本包含了高性能硬盘驱动器（HDD）的测试和性能要求，对数据传输速率、接口兼容性、电源管理和散热等方面进行了详尽的规定。 SFFspecifications 可在 SNIA 的官方网站上获取：http://www.snia.org/sff/specifications 或通过 Seagate 的 FTP 服务器：ftp://ftp.seagate.com/sff。对于有兴趣参与 SFF TWG 活动的人士，可以访问 http://www.snia.org/sff/join 进行会员申请。完成或正在开发的完整 SFF 标准列表可在相同网站的 SFF-8000.TXT 文件中找到。 SNIA 的 TWG 政策与程序是指导 SFF TWG 工作的基础，这些信息可以在 http://www.snia.org/sff/specifications/SFF-8032.PDF 查阅。该文档鼓励对 SFF-8410 规范的改进意见，提交至 SNIA 的反馈平台：http://www.snia.org/feedback。 SFF-8410 Rev 16.1 版本的核心内容聚焦于硬盘的高级功能，包括但不限于硬盘的数据传输速度（如 SAS、SATA 和 PCIe 接口），硬盘的热插拔和电源管理特性，以及确保在紧凑型系统中的可靠性和散热效率。此外，还包括了关于硬盘安全性能的要求，比如硬盘错误检测和修正机制，以及数据保护措施。这份标准对于制造商、数据中心管理员和存储系统设计师来说，是一项重要的参考指南，帮助他们在设计和选型过程中满足严格的性能和兼容性标准。

Published SFF 8410 Rev 16.1

HSS copper testing and performance requirements Page 15

Testing techniques for HSS duplex copper interconnect in this speed range lack

good standardization and methods. This deficiency leads to unintended

incompatibilities between suppliers and users. Further, the specific

conditions existing when a duplex connection is required are explicitly

addressed. This document is not a general specification for testing all high

speed copper interconnect but rather is aimed at the architectures used in

Fibre Channel, Gigabit Ethernet, 1394, and the like. Although the

specifications are written around a 150 ohm balanced transmission the methods

are readily adaptable to other transmission line impedance levels such as the

110 ohm used by 1394.

It is assumed that a completed cable assembly (media with connectors on both

ends) constitutes the primary interconnect component of interest. This does

not mean that the raw cable media is not important in producing good

interconnect but rather means that the focus of this document includes all the

effects of adding connectors, of placing connectors on boards, and of mating

the finished assembly to a shielded bulkhead. Except for the electromagnetic

compatibility tests, all tests apply to both shielded and unshielded

constructions. The test fixturing for testing raw media is not considered

directly in this document.

All tests except the impedance profile are intended to apply to all

constructions of cable assembly including those with equalizers and those with

known non uniformities such as intermediate connection points.

The general requirements are:

• for a launched signal with the most degraded allowed parameters to traverse

the interconnect without further degrading beyond the allowed input

specifications for the receiver

• for the interconnect to not export more noise to other parts of the system

than specified (anti pollution requirements)

• to behave like a matched transmission line for devices and other cable

assemblies attached to the ends

The received signal and exported noise are influenced both by the properties of

the launched signals and by the degradations occurring during the transmission

process. The goal is to deliver adequate signal quality to the receiver and to

not export excessive noise under the worst allowed cases. This means that both

the properties of the launched signals and the properties of the interconnect

must be considered in the testing processes.

Real launched signals always contain some level of imperfection so it is not

practical to require perfect launch signals. Similarly it is essentially

impossible to create a launched signal that is degraded to all the allowed

limits at the same time. There is a requirement to accommodate these two facts

into the testing strategy. Essentially there are two risks associated with

this issue.

One risk comes from inadequately characterizing the launched signals used in

the tests. Since the properties of the signals at the receiver are the only

properties that matter to an operating system, the standards only consider the

final result at the receiver. If launched signals used during the test are

degraded more than that allowed then the interconnect will be called on to

Published SFF 8410 Rev 16.1

HSS copper testing and performance requirements Page 16

cause less degradation so that the result at the receiver will still be within

specification. The use of excessively degraded launched signals places unfair

burden on the interconnect. Conversely, if the launched signals are better

than allowed, the interconnect may cause more degradation than allowed for the

interconnect but still deliver compliant signals to the receiver. This

condition permits defective interconnect to be measured as good interconnect.

The way to avoid these risks is to execute an adequate characterization of the

launched signals and to compensate in the test requirements for the amount of

excess goodness or badness in the launched signals. Figure 1 illustrates this

general scheme.

Another risk derives from the fact that some parameters in launched signals can

be corrected by interconnect that introduces degradation of equal and opposite

sign. If this happens it gives a false sense of goodness since launched

signals from other transmitters may have the parameters degraded in the same

sense as the interconnect with a resulting doubling in the negative effects at

the receiver.

In general, if the polarity of connection of the interconnect to the launched

signals is reversed then the polarity of the degradation in the interconnect is

also reversed (for those parameters that are sensitive to polarity, like

balance degradation and near end cross talk). Therefore the second risk can be

managed by performing a second test with reversed connections but changing

nothing else.

In order to avoid this risk one must take two measurements: (1) with the +

signal of the transmitter connected to the + line of the PUT and the - signal

of the transmitter connected to the - line of the PUT and (2) with the + signal

of the transmitter connected to the - line of the PUT and the - signal of the

transmitter connected to the + line of the PUT. Figure 2 illustrates this

scheme.

Summarizing, the combination of real launched signal properties and

interconnect properties causes additional burden on the testing process. The

compensation for real launched signals is likely to be a one time cost for the

same transmitter. The polarity reversal, however, requires that two

independent tests be executed because one cannot be sure which sense of

degradation is present in the interconnect under test.

During the calibration processes for the tests the properties of launched

signals are measured. Procedures are specified that do not require the

adjustment of the launched signal to the maximum allowed degradation. By

noting how much degradation could be added to the actual launched signal before

exceeding the maximum degradation and adding this difference to the

requirements for the received signals one achieves the equivalent effect as

actually degrading the launched signals as far as measuring the properties of

the interconnect is concerned. Said differently, if the launched signals are

better than allowed (as is usually the case) then the requirements on the

received signals are tightened by the same amount that the launched signals

were better. Similarly, if the launched signals are more degraded than allowed

then the received signal range is broadened.

This process eliminates a major problem with creating calibrated degraded high

frequency signals, uses the linear property of copper interconnect to good

advantage, and allows the properties of the interconnect to be fairly and

accurately measured.

Published SFF 8410 Rev 16.1

HSS copper testing and performance requirements Page 18

This section specifies the framework used for the high frequency performance

requirements and test methods to be used for measurement and verification of

properties of the interconnect. These requirements apply to the physical

connector and to its electrical neighborhood. See section 5.3. Since the

connector does not exist alone in a serial link the electrical properties of

(1) the wire (or other conductor) termination to the connector contact and (2)

any other non-uniformities within the electrical neighborhood of the connector

are intrinsic parts of the performance of the connector and its associated

interconnect. The electrical performance requirements are on the finished

cable assembly, board, the mated partner to the connector and whatever exists

within the electrical neighborhood of the physical connector. Tests that are

used to measure the high frequency performance properties are termed HSS tests.

5.1 HSS test levels

Two broad levels of HSS tests are described:

5.1.1 Level 1 test definition

Level 1 tests are those that are needed to ensure that the interconnect is

capable of (a) transporting minimum integrity (maximum degradation) launched

signals to the far end of the interconnect without excessive additional

degradation to the launched signals and without exporting excessive

interference to other parts of the system in the process and (b) to operate as

an effective transmission line between HSS ports.

5.1.2 Level 2 test definition

Level 2 tests are those that characterize specific contributors to the level 1

results and may thereby aid in diagnosing and revealing the source causes of

degradations measured in level 1 tests. Level 2 tests are expected to be

useful for designing and manufacturing interconnect components that satisfy

level 1 tests but are not individually required as direct performance measures

of the interconnect. It may be required to use specific level 2 tests to

establish the test conditions for level 1 tests.

5.1.3 Relationship between level 1 and level 2 tests

This separation of tests into two levels will be specified in detail in later

sections. By separating the testing requirements into the two very different

levels, testing resources may be more efficiently utilized compared with the

former schemes that required all tests to be individually satisfied. In

effect, only the level 1 tests need be used to verify an interconnect for sale

or use by both the supplier and the user. The level 2 tests are available to

the interconnect designer and manufacturer to more efficiently create designs

and manufacturing processes that produce good interconnect. Figure 3 shows a

graphical relationship between the two levels of test.

剩余102页未读，继续阅读

二零零零

粉丝: 2
资源: 41

SFF-8410：高性能SCOPP测试与性能规范

SFF-8485 SGPIO协议详解：电气与逻辑规范

SFF-8419规范：SFP28电源与低速接口详解

SFF-8663 Rev1.7规范：QSFP+ 28Gb/s笼形结构机械兼容性定义

SFF-8614.PDF

SFF-8419.PDF

SFF-8613.PDF

SFF-8644.PDF

SFF-8485.PDF

SFF-8472.pdf

SFF-8431.PDF

最新资源