1、AEROSPACE INFORMATION REPORTAIR1394REV.AIssued 1978-02Revised 1998-12Reaffirmed 2009-11Cabling Guidelines for Electromagnetic CompatibilityFOREWORDChanges in this revision are format/editorial only.1. SCOPE:These cable practice recommendations tend toward design guidance rather than standardization.
2、 EMC achievement tests can be standardized, but the means for achievement should not be constrained. The material can best be described as an essay on cabling, and the theme is that a cable is just a part of a complete circuit, the interconnect circuit. Cable EMC performance is thus determined large
3、ly by circuit design; it is unrealistic to expect cabling techniques to compensate for improper impedance, symmetry or waveform in the circuit.2. REFERENCES:There are no referenced publications specified herein.3. BACKGROUND:Cables are system elements containing interconnect circuits, and these caus
4、e more interference than do circuits contained inside boxes. Circuits in general exhibit a class of EMI problems related to conduction which includes crosswalk, ground loops, common impedance coupling and sneak circuits. All result from a unique characteristic of electric conduction; the current del
5、ivering energy from source to load must flow in a closed path, a circuit. As illustration, consider the functional flow diagram for any system and compare with the wiring schematic; function lines usually will not be found to correspond to wires in any consistent, simple manner. Perhaps 20% of the f
6、unctional flow lines do correspond one for one with nominally “complete” transmission lines; e.g., pairs or triples, coaxes or triaxes. The rest of the function lines are implemented with single wires sharing a common return or, frequently, finding a return elsewhere. Even the nominally “complete” t
7、ransmission lines are seldom truly complete; typical lines lose several percent of the fundamental return and most of the harmonic return in other cables and in structure. The functional flow lines which represent one-way energy flow are, in other words, implemented with circuit flow, that is, loop
8、flow. This results in large apertures and shared impedances, both of which cause interference. Conduction and induction interference is moreSAE 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 r
9、eport is entirely voluntary, and its applicability and suitability for any particular use, including any patent infringement arising therefrom, 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 cancell
10、ed. SAE invites your written comments and suggestions. Copyright 2009 SAE International All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without
11、 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) Fax: 724-776-0790 Email: CustomerServicesae.org SAE WEB ADDRESS: http:/www.sae.org SAE values your input. To provide feedbackon this Technical Report, please vis
12、it http:/www.sae.org/technical/standards/AIR1394ASAE AIR1394 Revision A- 2 -3. (Continued):of a chronic problem in cable circuits than in single chassis circuits, or even in long distance transmission circuits. Chassis circuits and long distance lines are unified designs, the latter having matched d
13、rivers and receivers. Cable circuits in contrast are not really designed. Typically, the driving and receiving end circuits are given to the cable designer as constraints, and he then must make the best of them. For these two reasons; i.e., circuitous flow and fragmented design, cables require caref
14、ul attention early in any program. This AIR, therefore, stresses integrated design of the entire interconnect circuit.4. DEFINITION OF INTERCONNECT CIRCUIT:The term “interconnect” refers to wiring which connects circuitry inside one box to circuitry inside another box some distance away. Noting that
15、 there are “boxes” within, and upon, other “boxes” in complicated units, the essential feature of an interconnect is that it traverses system structure. A “box circuit” in contrast stays within, or on the (metal) surface of one unified chassis.An interconnect circuit consists of cable wires plus the
16、ir terminating circuitry in the boxes at each end and at any intermediate junction box. These latter driving and receiving elements will be called end-circuits; they are, nevertheless, incomplete circuits, being part of the complete interconnect circuit.Circuitry which is complete within one chassis
17、 may be called box circuitry. Multi-rack circuits like computer circuits are “box circuits” if the racks are bonded into a continuous bay, but should be thought of as “interconnect circuits” if the racks are electrically distinct from each other. Figure 1 illustrates these definitions.5. COUPLING AN
18、D CONTAINMENT:The first and most obvious prerequisite for interference is transfer of energy, that is, coupling. At great distance, coupling consists of emission, path and reception. At close range; e.g., between two circuits in a cable, coupling consists of induction and conduction. Coupling is con
19、trolled by the “geometric” measures, twisting, balancing, and shielding in contrast to the “modulation” measures of Section 6. Coupling results from poor containment of energy within physical envelope and so the term “containment” can be used to describe both the goals of bottling up and excluding e
20、nergy. The second section of the guide, “Interconnect Circuit Design” deals entirely with prevention of coupling between a circuit of interest and some environment.The degree of containment that should be designed into a circuit for compatible operation in the system (e.g., whether or not to twist,
21、shield or balance) depends upon many factors. The least understood, yet most important factor, is the one generally called “influence factor”, and this is taken up next.SAE AIR1394 Revision A- 3 -FIGURE 1 - Interconnect Circuit6. INFLUENCE FACTOR:In order for one interconnect circuit to interfere wi
22、th another, the coupled signal of Section 5 must be of certain frequency, have certain modulation, and occur at certain times, and these specifics depend on the frequency, modulation and timing of the receptor. The need for twisting, shielding, balancing and so forth, therefore depends on this mutua
23、l “time-frequency” relationship, as well as on the “space-amplitude” relationship; i.e., energy transfer. The modifying effect of time-frequency coincidence upon the situation has been called “Influence Factor”:1Effective interference signal = (Coupled signal) x (Influence Factor)In the example of t
24、he reference, the influence factor for powerlines versus telephone lines describes the harmonic power falling within the speech bandwidth. Influence Factor is decreased by time and frequency control measures, such as gating, coding, frequency translation and rise time control. The modifying effect o
25、f Influence Factor upon containment design is difficult to foresee. Nevertheless, the protection afforded by time-frequency measures must be considered in order to avoid shielding overdesign.1. From telephone practice; e.g., Telephone Influence Factor for Power lines near telephone lines.SAE AIR1394
26、 Revision A- 4 -6. (Continued):An extremely large interfering signal can cause a category of response different from that just discussed. Demodulation, overloading or damage can occur when the signal exceeds normal limits. The Influence Factor concept does not apply to such conditions.7. GROUND:In e
27、lectrical usage “ground” means a current sink of infinite capacity. An ideal ground must, therefore, be large enough to return a sink current to its source wherever that might be. An ideal ground must also be highly conductive and accessible. Not all regions have a ground; if a continuous conductor
28、extends and is accessible throughout a region then that conductor forms “ground” for that region. This intuitive concept of ground is analogous to the frame of a mechanical system.If a system cable pattern is made up of many small loops (dense population of boxes highly interconnected) and if a grou
29、nd is desired, then it should be a plane. If two such regions situated some distance apart are joined into one system by long cables, then a criterion for the ground link is needed. To require that both ends of the link be at the same potential could be unrealistic. The general criterion is that the
30、 ground link be wide enough and close enough (to the link cables) to enable formation of a good image of the link cables. This requires that the ground link carry the full net distributed charge on the link cables. This is the criterion for an elongated ground.Usage has blurred the meaning of the te
31、rm “ground”. In common usage “ground” can mean either a true ground as described here, or just a wire that eventually connects to power common. It is recommended that the term “ground” be specifically modified if other than true ground is meant. To this end, the classic upside down “tree” symbol sho
32、uld be discarded; it no longer has meaning. The chassis connection symbolFIGURE 2is definitive and useful: Whether or not chassis is a good ground, there is no doubt about the meaning of this symbol. In most aircraft and spacecraft the structure is the only true ground available, so the chassis conn
33、ection symbol becomes synonymous with a ground connection. All other connections to ground-like conductors should be identified with modified symbols, for exampleFIGURE 3for which a complete description is provided.SAE AIR1394 Revision A- 5 -8. MODELS FOR ENVIRONMENTAL COUPLING:When two circuits are
34、 being designed to work together with no third party, then the problem variables are mutual ones: mutual inductance, mutual capacitance, relative power and relative timing. In the more usual case, an individual interconnect circuit is designed, tested, and qualified for an anticipated environment an
35、d specified level of emission. Constraints for this problem are based on some kind of assessment of Influence Factors (see Section 6) between the circuit and the eventual system. The problem variables in this artificial half-situation are environmental ones. Recommended models for use in the importa
36、nt low frequency range up to 1 MHz are described below.Emission from the circuit to the environment is best measured in terms of net monopole source current and charge. The net current in the interconnect wire bundle can be defined to be the reading of a clamp-on ammeter. This variable corresponds t
37、o the dominant mutual inductance of the complete two-circuit situation. The net charge per unit length of the bundle is not readily measured because a suitable probe is not readily available. The rod antenna of MIL-STD-461 provides an indirect measurement. Net charge is, nevertheless, the variable w
38、hich determines the radial (monopole) electric field of the bundle. Net charge corresponds to the dominant mutual capacitance in the complete two-circuit situation. Dipole and higher order source variables exist and can be significant. However, it is seldom necessary to consider multipole sources du
39、ring circuit design; twisting and light shielding will always control the higher order fields, should control be needed.The effect of environmental magnetic field on the circuit under design can be represented by voltage generators of equal value put in series with each wire of the circuit bundle. T
40、he common value of voltage equals the derivative of the expected magnetic flux threading the loop between bundle and ground, plus any IR drop in ground. This value for the equivalent generator can be called the “reference shift” to which the circuit is subjected. This reference shift corresponds to
41、the effect of mutual inductance and resistance in the complete two-circuit situation. The effect of environmental electric field on the circuit can be represented by current generators in shunt with each wire. Diagrams of the four models just described appear in Figures 4 and 5.SAE AIR1394 Revision
42、A- 6 -FIGURE 4 - Environmental Coupling Models - MagneticSAE AIR1394 Revision A- 7 -FIGURE 5 - Environmental Coupling Models - Electric9. INTERCONNECT CIRCUIT DESIGN:9.1 Preliminary Design:Electrical energy can be transferred by means of conduction, induction or radiation. Of these, conduction is th
43、e most versatile; together with induction and relay coupling, electrical conduction accounts for almost all cable transfer in use today. Conduction is not necessarily the choice of the future. Hybrid techniques for interchange of information and small amounts of energy over moderate distances are be
44、ing developed. Light coupling links from gap-size to tens of meters in length are in use. Experimental work continues in electro-chemical techniques similar to neuron transmission and synaptic coupling. Hybrid techniques will become increasingly useful in spacecraft and non-metal aircraft. Conductio
45、n is yet the most feasible mechanism and because of this the preliminary designer must make an early assessment of structural conductivity and circuit isolation.SAE AIR1394 Revision A- 8 -9.1 (Continued):The cost of achieving EMC in cabling is roughly the sum of circuit cost and ground plane cost. I
46、n some aircraft under development wiring is put into conduit because the structure does not provide a good enough ground plane; this cost can be traded with the cost of optical or other high performance isolation.The quantity of isolators needed in a system given no ground plane depends in large mea
47、sure on the number of loops in the functional flow map. A radial pattern has no loops and can in principle be put together without isolators by floating the circuits in the radial terminator boxes. In contrast, a function loop automatically requires at least one isolator.9.2 Return Current Rule:Rega
48、rdless of how moderate may be the goal for containment according to system analysis (see Sections 5 and 6), there is a minimum quality EMI design criterion which should always be met. This is the one-to-one rule implied earlier. Interconnect circuit current paths should correspond one-for-one with f
49、unctional flow paths. The rule in effect requires that return flow must either be bundled with forward flow or be imaged in adjacent ground; return flow can never be allowed to seep back through a route-of-opportunity in another cable. Consider the interconnect circuit wiring in Figure 6. This set of conductors includes all wires and wire shields labeled with a common function. The net current in this set due to nominal generators (exclude environmental current) should be the one-to-one rule,a. by zero, or elseb. return completely in adjacent ground.The ne
