1、_ SAE Technical Standards Board Rules provide that: “This report is published by SAE to advance the state of technical and engineering sciences. The use of this report is entirely voluntary, and its applicability and suitability for any particular use, including any patent infringement arising there
2、from, is the sole responsibility of the user.” SAE reviews each technical report at least every five years at which time it may be reaffirmed, revised, or cancelled. SAE invites your written comments and suggestions. Copyright 2007 SAE International All rights reserved. No part of this publication m
3、ay be reproduced, stored in a retrieval system or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of SAE. TO PLACE A DOCUMENT ORDER: Tel: 877-606-7323 (inside USA and Canada) Tel: 724-776-4970 (outside USA)
4、 Fax: 724-776-0790 Email: CustomerServicesae.org SAE WEB ADDRESS: http:/www.sae.org ARD5708 AEROSPACE RESOURCE DOCUMENT Issued 2007-03 Frequently Asked Questions About IEEE-1394b and SAE AS5643 RATIONALE This document is a compilation of frequently asked questions about the recently-published AS5643
5、 and AS5643/1 standards, both for the benefit of those currently implementing these standards and for those considering their use. Q1. Whats the difference between 1394a and 1394b? A1. IEEE-1394a-2000 and IEEE-1394b-2002 are both amendments to the original IEEE-1394-1995 standard. As per section 3.1
6、0.1 of IEEE-1394b-2002, “IEEE-1394a-2000 provides arbitration acceleration, improves the speed of bus resets, and adds suspend and resume power management capabilities to the bus.” Continuing on, later in this section, “IEEE-1394b-2002 supports the DS form of signaling used in IEEE-1394a-2000 and IE
7、EE-1394-1995 while adding a new form of signaling capable of much higher data rates between Beta ports.” This new form of signaling is a modified version of the 8b/10b coding scheme used for Fibrechannel and Gigabit Ethernet. It is DC-balanced, so it allows DC-free data transmission that is needed w
8、hen transformers or capacitors (to separate ground domains) are used in the signal path. The addition of a scrambler to all data and control signals results in a significant reduction (approximately 20 dB) in the radiated emissions when compared to earlier repetitive control patterns. Continuing on
9、in section 3.10.1 of IEEE-1394b-2002, “IEEE-1394b-2002 also supports new media, such as Category 5 (CAT-5) unshielded twisted pairs (UTPs), glass optical fiber (GOF), and plastic optical fiber (POF).” These new media allow maximum cable lengths greater than earlier limitations of 4.5 meters: 50 mete
10、rs for POF, 100 meters for GOF, and S100 operation over lengths of CAT-5 up to 100 meters. IEEE-1394a-2000 supports speeds of S100, S200, and S400. IEEE-1394b-2002 is backwards-compatible with IEEE-1394a-2000; in addition, IEEE-1394b-2002 extends bus speeds to S800 and S1600 and has architectural su
11、pport for S3200. Q2. Whats the difference between 1394b and the SAE implementation? A2. The SAE AS5643 base specification standardizes an approach for using IEEE-1394b-2002 in safety-critical/mission-critical applications for military and aerospace vehicles. It contains generic requirements that spe
12、cify data bus characteristics, data formats, and node operation. Additions include the use of asynchronous stream packets, a fixed frame rate synchronized with a STart Of Frame (STOF) packet, addition of a Vertical Parity Check, static assignment of channel numbers, pre-assignment of bandwidth, and
13、use of Anonymous Subscriber Messaging. The SAE implementation uses IEEE-1394b-2002-compatible devices. The AS5643/1 slash sheet establishes guidelines for the data bus cable and its interface electronics for a system utilizing S400 over copper medium over extended lengths. Whereas IEEE-1394b-2002 li
14、mits the distance between hops to be no more than 4.5 meters, the AS5643/1 slash sheet accommodates much longer distances (see cable selection guide contained in A7 below), and the hardware requirements and examples address many of the environmental conditions (electromagnetic compatibility, tempera
15、ture, vibration, etc.) that military and aerospace vehicles may experience. SAE ARD5708 - 2 - Q3. Where can I find enough information to understand 1394a and 1394b? A3. Start by obtaining the IEEE standards: IEEE-1394-1995, IEEE-1394a-2000, and IEEE-1394b-2002. “Firewire System Architecture”, by Don
16、 Anderson of Mindshare (), is the most well-known reference book on IEEE-1394; it explains aspects of the standard that may be ambiguous and supplies real-world technical examples, but it is no substitute for having the actual IEEE standards. The Texas Instruments website () is another good resource
17、 for IEEE-1394 application notes (http:/ revision 1.1 can be obtained here: http:/ Of course, the best way to learn more about the SAE 1394b standards is to get involved with the SAE AS1-A3 task group! Q5. What do you mean by “asynchronous streaming packets”? Are they asynchronous packets or are the
18、y streams? A5. Refer to section 6.2.3A of IEEE-1394a-2000 for a detailed write-up on asynchronous streams. An asynchronous stream packet is identical in format to that of an isochronous stream packet, except that it is transmitted during the asynchronous period, subject to the same arbitration requi
19、rements, including fairness, as other asynchronous request subactions. Asynchronous stream packets, like isochronous packets, use channel numbers for addressing whereas asynchronous request packets are mapped into 1394s 64 bit address space. Asynchronous stream packets do not require the allocation
20、of bandwidth, however allocation of channel numbers is required (may not be for AS5643), and may be easily filtered by hardware. Q6. In paragraph 3.1 of AS5643, it mentions that 1394 transmissions occur in a dual simplex manner on TPA and TPB. What does that mean? A6. Each port of an IEEE-1394b-2002
21、 physical device (PHY) contains the transmit differential signals (TPB) and the receive differential signals (TPA). As explained in paragraph 3.1 of AS5643, “packets can be transmitted on the TPB pair at the same time arbitration and or signaling is received on the TPA pair and packets can be receiv
22、ed on the TPA pair at the same time arbitration is transmitted on the TPB pair.” Q7. What is the maximum number of nodes, bus lengths, etc. on the bus? A7. Per IEEE-1394-1995 and IEEE-1394a-2000, the maximum number of nodes on a bus is 63. However, up to 1024 busses may be used with bridges connecti
23、ng them together. That provides for over 64,000 nodes in a system. Regarding cable length, IEEE-1394-1995, IEEE-1394a-2000 and IEEE-1394b-2002 define transmitter signaling, cable/connector loss, and receiver sensitivity. None of these specifications explicitly define a cable length, however with a w
24、orst case cable/connector the longest distance is 4.5 meters. The AS5643/1 slash sheet accommodates much longer distances adjusting a number of factors that affect maximum bus length, including bus speed, wire gauge and construction, termination methods, and the number of disconnects between hops; r
25、efer to the following cable selection guide for guidance. SAE ARD5708 - 3 - Cable AWG SizeS100 (50 MHz Fundamental Frequency)S400 (250 MHz Fundamental Frequency)S800 (500 MHz Fundamental Frequency)S100 (125 Mbps)S400 (500 Mbps)S800 (1000Mbps)S100 (125 Mbps)S400 (500 Mbps)S800 (1000Mbps)22 3.5 7.9 11
26、.2 28 13 9 165 73 5224 4.4 9.8 14.0 22 10 7 130 59 4126 5.4 12.3 17.5 18 8 6 105 47 3328 6.8 15.4 21.9 14 6 4 85 38 2730 8.6 19.3 27.4 12 5 3 68 30 21Cable AWG SizeS100 S400 S800S100 (125 Mbps)S400 (500 Mbps)S800 (1000Mbps)S100 (125 Mbps)S400 (500 Mbps)S800 (1000Mbps)22 0.15 dB 0.15 dB 0.25 dB 4.0 2
27、 2 4.0 2 224 0.15 dB 0.15 dB 0.25 dB 3.5 1.5 1.5 3.5 1.5 1.526 0.15 dB 0.15 dB 0.25 dB 2.5 1.5 1.5 2.5 1.5 1.528 0.15 dB 0.15 dB 0.25 dB 2.5 1 1 2.5 1 130 0.15 dB 0.15 dB 0.25 dB 1.5 0.75 0.75 1.5 0.75 0.75Cable Selection GuideThe following tables may be used to select an appropriate cable AWG size
28、based on the attenuation (insertion loss per unit length), transmissiondistance, and data rate requirements of the link. Maximum link lengths are listed based on no disconnects in the line. The second table lists aderating factor in feet for each disconnect in the cable run. In all cases a cable of
29、Quad construction was used (see below for cross sectionalview).Using Active Transformers (Assumes Minimum Differential P-P Voltage at TP2 = 1100 mV, Receiver Sensitivity at TP3 = 504 mV)Using Passive Transformers (Assumes Minimum Differential P-P Voltage at TP2 = 635 mV, Receiver Sensitivity at TP3
30、= 504 mV)Maximum Link Length* in Feet(*) Derating example: Assume 26 AWG cable for S400 speeds with three (3) disconnects in the link. Base maximum distance = 47 ft. De-rating length = 3 * 1.5 ft = 4.5 ft. Total maximum distance = 47 ft - 4.5 ft = 42.5 ft(*) The maximum link length for a given cable
31、 AWG is primarily determined by the performance characteristics of the transmitter and receiver and/or the use of other inline components (such as active or passive transformers). The lengths documented in this selection guide have been measured experimentally and compared to the eye mask defined in
32、 the IEEE 1394b specification (with no disconnects). Performance characteristics of the physical cable should be confirmed as well as the performance characteristics of any disconnects in the cable run.Typical Attenuation (dB / Disconnect)Typical Attenuation (dB / 100 ft)Derating Factor* in Feet Per
33、 38999 Disconnect in a Cable RunUsing Passive Transformers Using Active TransformersSAE ARD5708 - 4 - Q8. I find the explanations of assigning static IDs to nodes confusing. Im still not clear on if a node receiving a message decides to receive a message because it has an ID that is of interest or i
34、f it is based upon offset times from the STOF. Please explain. A8. Node IDs are not static and are not assigned; channel numbers are static and are assigned. A node receives an asynchronous stream or isochronous packet because that message is addressed to a channel number that the node is listening
35、to. When using isochronous streams, nodes request a channel number from the Isochronous Resource Manager (IRM). AS5643 eliminates the requirement for an IRM by having channel numbers and bandwidths pre-assigned. At minimum, a node will listen to the STOF channel and the channel that is specific to t
36、hat node by system design. A node may retrieve data (that is sent to it from the CC) at any time during a frame, but for latency-critical applications, it will be assured of the “freshest” data immediately following the nodes receive offset. Node IDs are used as source and destination addresses in a
37、synchronous transactions. Node IDs are assigned by the PHYs during every bus reset event. Since AS5643 does not use asynchronous transactions, Node IDs are generally ignored by the software developers. Channel numbers, however; are pre-assigned to each node in the network requirements documents (e.g
38、. Interface Control Document) and are coded into the software. Then the LLC is programmed to receive packets sent to that channel number. Packets sent to channel numbers not programmed into the LLC will be discarded by the LLC. Q9. The statement in the spec “Asynchronous and isochronous packets are
39、not required but may be utilized” leaves me, the reader, with the uneasy feeling that I may utilize them at my peril! I have the more general concern that 1394 has various features in it that if I try to use them in Mil-1394 will cause serious problems. Correct? If so, shouldnt the spec be very expl
40、icit about this? As examples, bus resets and dynamic node assignments seem to be forbidden but Im not sure how you make sure that they are inhibited. A9. Section 3.3.3.2 of AS5643 describes how asynchronous transactions may be used for command and control, but the system implementer must take care t
41、o ensure that they do not conflict with STOF offset times or system timing constraints in general. Their use will not cause serious problems but could cause the STOF timing to be erratic. If the system is designed to accommodate this erratic STOF timing, no harm will be done. Section 3.3.3.3 of AS56
42、43 describes how isochronous packets may be used for streaming audio and/or video applications. A method is described for determining how much spare bandwidth may be utilized for such purposes. As noted, the interleaving of isochronous and asynchronous stream packet types will have the effect of cre
43、ating “soft” indeterminate boundaries for the STOF times of each remote node. The system implementer must calculate the degree to which STOF times could vary for their implementation, and then determine if this “drift” in STOF times is acceptable for all latency-critical applications that depend upo
44、n such timing. As stated in the previous paragraph, their use will not cause serious problems but could cause the STOF timing to be erratic, so the system must be designed with this in mind. A lesson learned from one project was that the use of a very slow STOF rate (e.g. 10 Hz) allowed the mixture
45、of streaming video and audio with the data packets. Because the STOF rate was so slow, a disruption of STOF timing of a hundred microseconds or so did not present any problems. Also, the slow STOF frame rate was used only for Command and Control and not for Flight Control. Not sure what you mean by
46、dynamic node assignments. Perhaps youre referring to channel numbers rather than node IDs. If channel numbers is the question, it will be handled in the nodes software. There is no way to inhibit bus resets, but we suggest that there not be any code, in a node, that generates a bus reset. That way,
47、run-away code should not be able to execute a bus reset. Refer also to the answers to questions 13 and 31. SAE ARD5708 - 5 - Q10. It is unclear to me that the combination of AS5643 and the use of 1394b hardware and software will necessarily produce Mil-1394 implementations that are interoperable. Is
48、nt it the goal of a standard to allow different vendors to produce equipment that can interface to others vendors equipment and will be interoperable? A10. The objective of interoperability is different for military and aerospace vehicle applications than it is for consumer products. For example, a consumer expects to be able to connect their “Brand X” camcorder to their “Brand Y” computer and be able to capture video of their childs birthday party with a couple simple clicks of the mouse. However, it is not realistic to a
