Bonding Cable Shields at Both Ends to Reduce
waitress
Noi
by Tony Waldron and Keith Armstrong
Table of contents
write是什么意思1 Synopsis
2 Introduction
3 The cables tested
4 The sources testedam
5 The audio test gear ud
6 The test t-ups ud
7 Test results analyd by CMRR
8 Analysis of cable equivalent circuit
9 Analysis of the effect of the building’s ground system impedance
10 Some related issues
11 Poor equipment design is the real cau of ‘ground loop’ noi
12 Conclusions
13 Referencesnamp
1 Synopsis
Many equipment designers, including some in professional audio companies, have found that radio-frequency (RF) bonding the shields of cables to chassis/frame/enclosure at both ends helps greatly with EMC Directive compliance. They have also found that it does not compromi the signal quality when the equipment is installed – despite permitting power-frequency ground loop currents (earth loops) to flow in the cable shields. Professional audio companies using this technique have found that it also saves a great deal of time during installation and commissioning.
But, in many industries, ground loops are anathematid by long-standing tradition as a significant cau of noi in the cables’ signals. This aversion to ground loops is especially strong in the professional audio industry, which us balanced cables for audio signals and often goes to extreme lengths during installation and commissioning to prevent ground loops. This article address the ground-loop concerns in the pro-audio industry, but its results are easily applied to balanced (differential signalling) and unbalanced (single-ended signalling) shielded cables ud in other electronic application areas, for both analogue and digital signals. Its conclusions will be of value wherever ground loop currents in cable shields are currently avoided in the attempt to improve signal/noi ratios or noi margins.
The authors found plenty of anecdotes but few hard facts when investigating the effect of ground loops on signal noi, so performed some tests themlves and reached some very interesting and valuable conclusions. This article describes tests we performed on a variety of balanced audio cables nearly 30m long with metallid foil or braid shields, to determine the effects of power-frequency shield currents (ground loop currents) on noi. Several types of balanced audio cables were tested, including an extremely poor quality balanced audio cable with untwisted signal conductors and a capacitive imbalance exceeding 20%.
Analysis of the test results revealed, to the authors’ surpri, that power frequency currents in metallid foil or braid shields do not inductively couple significant noi into their internal conductors even at current levels which cau the cables to warm up. However, the voltage between the cable’s shield and its internal conductors is a significant source of noi for balanced signals which have high impedances to ground. For good quality pro-audio balanced cables and equipment, the signal noi created by the equipment’s common-mode rejection is comparable with that created by capacitive imbalance in the cables.
It appears that traditional equipment design methods that connect cable shields to conductors inside equipment are most probably to blame for the problems which have been blamed on ground loops. Equipment constructions that bond cable shields directly to the chassis/frame/enclosure are better for EMC compliance and allow signals to achieve the highest levels of quality regardless of the ground loop currents flowing in their cable shields.
2 Introduction
Professional audio equipment and systems supplied in the European Union (EU) must meet the EMC Directive, and unless they u its Technical Construction File (TCF) route to compliance pro-au
dio suppliers must declare conformity to EN 55103-1:1997 and EN 55103-2:1997 – the product-family harmonid EMC standards for audio, video, audio-visual and entertainment lighting control apparatus for professional u [1], [2]. Other types of equipment have other EMC standards applied.
When the EMC emissions tests are properly applied to equipment, systems or installations it is generally found that unless the shields of their external cables are radio-frequency (RF) bonded at both ends of their cables –
Equipment which us digital control and/or processing (e.g. which contains a microprocessor) or switch-mode technology will generally fail the emissions tests
Equipment which us analogue signal processing will generally fail the continuous RF immunity tests
RF bonding techniques which work well at radio-microphone, cellphone and wireless LAN frequencies (e.g. up to 2.5GHz for IEEE 802.11b, also known as Wi-Fi) require cable shields to be terminated using 360o electrical bonds between the shield and the cable connector shell, between that shell and the shell of the equipment’s mating connector, and between that shell and the equipment’s chassis/frame/enclosure. 360o bonding is sometimes called peripheral bonding or circu
mferential bonding.
On the other hand, cable shields which are only bonded at one end cea to provide shielding when their length exceeds one-tenth of the wavelength of the frequencies to be shielded against, so for example a cable 10m long only provides any significant shielding for frequencies below 3MHz. When cable lengths exceed one-quarter of a wavelength, shields which are bonded at one end only can become very efficient RF antennas – radiating RF noi and picking up RF from the environment more efficiently than if there was no shield at all. Although the RF noi in pro-audio products is usually caud by digital and switch-mode circuits, it appears as common-mode (CM) noi on all the analogue inputs and outputs too. The problem is that in the professional audio industry there is a long-established tradition of bonding cable shields at one end only, to prevent ‘ground loop’ currents from flowing. Ground loop currents are traditionally blamed for increasing the levels of hum and other mains-borne nois in the audio signals. In some installations cable shield currents can be so high as to overheat a single cable, making it unreliable and possibly even causing safety risks. Unfortunately, in a complex modern pro-audio installation the best way to bond the cable shields when using single-ended bonding is not always obvious, and it can take a great deal of time for even very skilled installers to determine the unique optimum solution when commissioning a pro-audio sys
tem. This time-consuming exerci often needs to be gone through again every time the system is changed.
From the above it might be concluded that there is an impas – either we have pro-audio products, systems and installations that u one-ended shield bonding and have good audio quality but fail EMC compliance – or we u double-ended shield bonding and achieve EMC directive compliance but with poor audio quality.races
One way that is often suggested to overcome this dilemma is to electrically bond a cable shield to the protectively-grounded chassis/frame/enclosure at one end, and ‘RF bond’ it using a capacitor in ries with the bond at the other end. The idea is that good RF bonding is achieved at both ends, aiding EMC compliance, while ground loop currents (at power frequencies) are prevented from flowing.
(Protective grounding is necessary for the safety of all mains-powered equipment that is not ‘double insulated’. Protective grounding is sometimes called protective earthing, but becau ‘ground’ and ‘earth’ are such misud and abud terms it would be better to u the phra
‘protective bonding’ instead. A structure’s protective ground (earth) network is best called a protectiv
occupy是什么意思e bonding network, and the word ‘earth’ rerved solely for the soil-penetrating electrodes that connect to the mass of the planet Earth. Having made the comments, ‘protective ground’, ‘protective grounding’, ‘grounding’ and ‘bonding’ are the terms ud here.) The problem with the capacitor-at-one-end approach is that capacitors with good RF performance over a wide frequency range are costly and must be designed to fit within the bodies of the cable connectors. The inductance inevitably associated with fitting capacitors to connector pins with flying leads or by a printed circuit board (PCB) reduces their ries-resonant frequencies and diminishes the range of frequencies over which they are effective. The authors have received an estimated price of US$4 each from Metatech Corporation () in October 2001 for 1000-off quantities of an EESeal™ RF capacitor asmbly which fits inside the shell of a 3-pin XLR connector (the traditional pro-audio balanced cable connector) and ‘RF bonds’ pin 1 of the XLR – the pin ud for shield bonding – to its metal shell and hence to the chassis/frame/enclosure of the equipment. Even at the 10,000-off price estimate of US$2.25 each this is a costly modification considering that XLR connectors usually cost between US$1.50 and $3.00.
Despite the fact that EESeal™ devices provide much better RF performance than any type of capacitor that could be connected externally to an XLR connector, Figure 1 shows that they only perf
orm very well over a limited range of frequencies, depending on their capacitance value. So there can be no guarantee that any given capacitance value will enable a particular product to meet the 150kHz to 1GHz emissions and immunity tests in [1] and [2].
北京爱迪国际学校
Higher-performance RF capacitors, such as annular or feedthrough types, could possibly be designed into XLR connectors and provide much better performance over a wider frequency range, but they are likely to be very much more costly than the EESeal™ example given above.
Another problem with the capacitor-bonding-at-one-end technique is that it does not remove the problem of deciding which end of the shield of each cable should be directly bonded to the grounded equipment. This probably means that interconnections would need to be fitted with
RF-bonding capacitors between their shield and chassis/frame/enclosure at both ends, knowing that one end would have its capacitor shorted out. This method would add to material costs without reducing the traditionally long installation and commissioning times. Single-ended shield bonding permits very high surge (transient) over-voltages to exist at the unbonded cable ends [8], mostly caud by lighting. Modern steel-framed buildings provide a reasonable amount of shielding from such atmospheric disturbances, but older wooden or brick buildings will not provide as much and open-air venues provide none. Cloud-cloud lighting within 2 miles radius couples well with horizontally-run cable shields and can inject 100V per metre (1kV for 10m length, 10kV for a 100m length). Cloud-cloud lightning is at least 10 times more common than cloud-ground strokes. Lightning experts have en arcing occurring at the unbonded ends of shielded cables during thunde
rstorms. Unreliability of electronic equipment and incread safety hazards can therefore be a feature of installations that u single-ended shield bonding, unless surge protection devices are liberally applied. In some installations, connecting cable shields to ground via RF capacitors will place tho capacitors at risk of damage from surges. Read chapter 9 of [5] for more about lightning protection for electronic systems.
Capacitive shield-to-ground bonding can be a uful ‘fix’ for individual interference problems in existing installations, especially when using good products like EESeal™ and spending some time finding the best capacitor value, and especially when they are ud to filter the signals on the balanced conductors (additional capacitors hardly add to the EESeal™ unit cost). But taking all the above into account we cannot recommended the u of RF capacitors to connect cable shields to ground it as an technically effective or cost-effective method suitable for universal application to pro-audio interconnections.
IEC 61000-5-2:1997 [3] describes the best practices to u for grounding and cabling to achieve EMC in installations, and it is referenced by many of the latest standards on installing telecommunications and computer systems. [4] and [5] are additional references for people interested in the techniques described in [3].
Some pro-audio products manufacturers, system integrators and installers, such as Cadac Electronics Ltd (e Figure 2) have found that [3] works very well indeed when applied to pro-
xinleiaudio installations – allowing them to comply with [1] and [2] whilst also achieving very high quality a
udio. Also, as the electromagnetic noi in modern environments continues to ri and as digital processing is increasingly ud for pro-audio and co-located equipment, compliance with the EMC immunity standard [2] is increasingly found to be important for the achievement of good quality audio.
[3] recommends directly bonding cable shields at both ends using 360o RF bonding techniques. It also recommends using a “Parallel Earth Conductor” (PEC) with a lower impedance than the shields, where necessary, to divert a large proportion of the power-frequency ground loop currents away from the shields and preventing them from overheating.
A PEC can be a dedicated conductor, or it can be new or existing metalwork, as long as it is bonded to the frame/chassis/enclosure of the equipment at both ends of the cables concerned (effectively in parallel with their shields). In many pro-audio installations there are usually large numbers of shielded cables following any given route. Bonding all their shields at both ends would considerably reduce the currents flowing in each shield below the levels which could cau overheating – so an additional PEC might not be needed.
A very valuable benefit of employing [3] is that the days (sometimes weeks) that skilled installers usu
ally spend trying to find the best way to bond the ends of each of hundreds of cable shields to minimi hum and noi is no longer needed. Products designed to u the techniques described by [3], and installed accordingly, generally achieve excellent audio quality immediately – by design. If a system/installation is changed – providing the techniques described in [3] are still followed – excellent audio quality is again achieved automatically without the need for highly-skilled modifications to the cable shield grounding scheme.
So the u of [3] permits the design of products and systems that are EMC compliant for legal supply in the EU, have the desired audio quality, and save a great deal of time (and money) in their installation. The additional material costs and asmbly time required to implement [3] in products and systems is small, and in any ca is vastly outweighed by the time and cost (and, increasingly, audio quality) benefits it achieves.grd什么意思
The effective u of the techniques described by [3] require that the electronic equipment connected at each end of the cable bond all cable shields directly to their low-impedance chassis, frame, cabinet, or enclosure to prevent the 50/60Hz currents in the shields from interfering with their audio circuitry. This shield-bonding technique is already commonplace in telecommunication and computer equipment, where it is usually necessary for the maintenance of adequate signal integrity as well as f
or EMC compliance.
Unfortunately, many of the cable connectors traditionally ud in the pro-audio industry (e.g. XLRs) do not yet permit 360o shield-bonding direct to the chassis/frame/enclosure. Instead, they connect to the shield using one of their connector pins, necessitating a ‘pigtail’ type of wired shield-chassis connection. Pigtail shield connections are singled out for attention by [3] as very bad practice, but careful attention to detail such as pigtail length and routing, and the u of RF filtering, have allowed short pigtails to be ud in some EMC-compliant pro-audio equipment. Products that employ powerful digital signal processing or high-speed data links may have EMC compliance problems with any practical length of pigtail or connector pin. The u of pigtails and connector pins for shield bonding is likely to become more problematic if/when the EU’s EMC emissions and immunity tests are extended beyond 1GHz, as they are likely to be in a few years time.
Despite the very good results achieved for both EMC compliance and audio quality by following [3], the pro-audio industry is in general still very wary of bonding cable shields at both ends. The long-established tradition of avoiding ground loops at any cost appears to have led to the commonly-held view that cable shield currents directly cau hum and noi problems due to noi coupling within the cables themlves.
The authors could not find any published papers, articles or data on the details of the noi coupling caud by shield currents in balanced audio cables, so decided to test a few cables to find out why [3] gives such good results in pro-audio installations.
惊喜 英文