|

The connecting cables between devices and their connectors
are the preferred route by which electromagnetic interference propagates in the
external environment and by which, from the outside, it couples to the
electrical conductors that exchange signals between the devices of the system,
thus creating situations of susceptibility.
It should be remembered that cabling is an area of
uncertain responsability; it is never clear who should choose the cables, their
route, or their location inside the system. Therefore it may
happen that an apparatus is designed perfectly from the EMC point of view, but
that the choice of the connecting cables between the apparatus itself and the
outside world is not made with equal care. Thus, the apparatus-cable whole may
not pass the tests of EMC.
There are essentially three mechanisms by which phenomena of emission and
susceptibility manifest themselves in electrical connections:
· radiated emission from cables, which act as transmitting
antennae;
· radiation susceptibility of the cables, which act as
receiving antennae;
· electric and/or magnetic coupling between cables.
In the control of interfering emission the principle of reciprocity rules, in
the sense that what is done to a cable to reduce radiated emission is also valid
for increasing its resistance to radiation susceptibility.
As far as cable screens are concerned, it must be remembered that their
effectiveness depends on the value of the SE of the cable braid and on the way
in which the screens are terminated. From the termination of cable screens point
of view, connectors come into play; these are an indispensable means of
connection between cables and devices.
Only recently has the great difflculty of seeing cables and connectors as two
separate entities been overcome and have they begun to be considered as one.
In the literature there exist many studies aimed at verifying and quantifying
the effect of the connection of a cable screen to the chassis ground of an
apparatus (the so-called pigtail); everyone agrees on the criticality of this
element with regard to screening efficiency and radiation.
In any case the influence of the type of connector starts to make itself
felt at a few MHz, or the frequency at which the length of the cable is equal to
a tenth of a wavelength.
Overlooking the cable contribution to the system EMC performance can ruin the
effect of the best system shielding. In fact, cable shielding very often is the
weakest link in the overall system shielding.
For a given frequency, the key contributors to the overall performance of a
shield placed upon a cable are :
· shield material
· termination method (clamp, solder, pigtail, type of
connector)
· geometry of the installation (cable lenght, cable height,
orientation vs. the field pointing vector, type of conductive ground)
In the final analysis, each of these contributors behaves differently, depending
on the nature of the field : near E-field, near H-field or far field.
SHIELD MATERIAL AND THICKNESS VS. THE TYPE OF EMI
A cable shield can be made of a thin aluminum
film flashed over a paper or mylar substrate, a copper braid, a thinly wrapped
metal foil, a corrugated pipe. All these material have basic shielding
properties which depend on two mechanisms reflection and absorption.
Reflection is a radiated phenomenon and is the result of the mismatch between
the impinging wave impedance and the shield barrier impedance.
Absorption depends on the skin effect. At frequencies high enough for the skin
effect to occur, absorption becomes significant. If the shield is a solid tube,
this absorption will increase exponentially. If the shield is a braid, all the
minuscule rhombic aperture due to the weave of copper wires will spoil the
absorption effect by making the shield more and more trasparent as frequency
increases. Depending on the optical coverage, this effect can show up as low as
1 to 10 MHz, ruining the Shielding Effectiveness (SE) of the braid.
SHIELD TERMINATION
Where and how the shield is terminated can
radically affect the performance of a shield.
Once a cable shield is mounted, its termination may be the weakest link in the
chain, especially at higher frequencies.
As a first approximation, the quality of a shield can be associated to its DC
sheath resistence, altough this statement becomes invalid above the HF region.
Since it is very difficult to have the shield terminated by a connector clamp or
pigtail whose impedance is much less than the shield material impedance, the
termination hardware is always the limiting factor of the in situ performance. |
|
For example, a
braid shield resistance of a 0.75 m piece of shielded cable, with a
braid resistance of 3 mohm/m is equal to 2.25 mohm. A typical bonding
resistance of the braid to the serrated clamp is 0.5 mohm. An ordinary
hand-tightened connector shall exhibits a contact resistance of about 3
mohm to the receptacle. Finally, the contact resistance of the connector
socket flange to the rack wall, assuming it is normally tool tightened
using four screws, is again 0.5 mohm.
Since there are two ends to this cable, the total termination resistance
is :
Rterm = 2 (0.5 + 3 + 0.5) = 8
mohm > 2.25 mohm of the braid shield resistance
Il contributo delle
terminazioni sulla resistenza complessiva dello schermo è maggiore di
più di tre volte quella della calza considerata da sola.
|
 |
|
The termination contribution to the whole shield resistance will
be more than three times that of the braid alone. Not only can an improper
shield termination spoil the cable shield efficiency, it can be the source of
secondary mechanism which can cause more EMI than if there were no shield at
all.
The key for preventing or solving such problems is always
the same : identify any existing EMI current paths.
OVERALL GEOMETRY OF THE
INSTALLED CABLE
Lo schermo di un cavo può essere
visto come la congiunzione di due linee di trasmissione, con lo schermo che è
il mezzo di trasferimento tra loro. I
due modelli circuitali sono :
-
Il circuito interno 2
ha parametri ben controllati come le impedenze di sorgente Zg e
di carico Zl, e impedenza caratteristica Z02.
Questi parametri definiscono i valori di corrente e tensione in ogni punto
nel circuito, includendo gli effetti di riflessione e onde stazionarie.
-
Il circuito
esterno 1 ha parametri molto poco controllabili, in quanto ognuna delle
impedenze di terminazione Zg1 e Zg2 possono variare da
0 a infinito, a seconda delle condizioni di grounding, e la sua impedenza
caratteristica Z01 dipende dal rapporto h/D, ovvero l'altezza del
cavo sopra il piano di massa rispetto al diametro del cavo.
|
|
L'efficienza di uno schermo è infatti la
misura della percentuale di energia trasferita dal circuito 2 al circuito 1 nel
caso di emissione EMI, oppure trasferita dal circuito 1 al 2 nel caso di
suscettibilità.
L'impedenza caratteristica del
circuito1 ha un notevole effetto su questo trasferimento di energia,
specialmente a valori multipli di landa/4.
Perciò, per un dato schermo
installato, con certe terminazioni ci saranno comportamenti diversi, a seconda
della sua lunghezza e della sua altezza sopra il relativo piano di massa. |
 |
|
UNBALANCED VS. BALANCED SHIELDED CABLES
La definizione di base dell'efficienza di
schermatura (SE), è :
SE(dB) = 20 log [Vindotta senza schermo / Vindotta
con lo schermo ]
Il metodo di terminazione dello schermo influenza
fortemente tale risultato e può nascondere completamente i parametri da
misurare.
A causa di questi problemi, una misura preferibile per valutare la qualità
dello schermo, è quella che sfrutta la grandezza denominata
impedenza superficiale di trasferimento Zt.
Questa grandezza fornisce in pratica una figura di merito come parametro
assoluto, piuttosto che relativo come la SE definita sopra. Inoltre, la
misura di Zt è relativamente facile da effettuare.
DEFINITION OF SE FOR CABLE SHIELD
The basic definition of SE is :
SE(dB) = 20 log [Vinduced without shield / Vinduced with shield ]
The shield termination method strongly influences the result and can even
totally obscure the parameter being measured, i.e., the quality of the shield
alone.
Because of these problems, another measure of the shield quality, the shield
surface transfer impedance ZT is preferred. It provides a figure of merit which
is an absolute parameter, instead of being a relative term like SE. Also, the
measurement of ZT is relatively easy to perform.
TRANSFER IMPEDANCE OF COAXIAL CABLE
SHIELDS
To define the quality of a cable in the EMC sense,
the concept of surface transfer impedance has been introduced.
Consider figure, where we show the screen of a cable in which a surface current
Is flows and let DV be the voltage drop that is generated inside the cable over
a length ,DX.
Then the surface transfer impedance Zt
is defined as
Zt = DV
/ (Is· Dx)
(ohm / m)
The lower the ZT the
better the shield quality for EMI reduction.
Below about 100 KHz, Zt is practically equal to the shield DC resistance.
The transfer impedance will consist of two components :
1. penetration component representing the energy diffusion through the
metal of the screen
2. coupling component representing the H field diffusion through the
rhombus-shaped holes
The complete Zt for the
braided wire shield can be written :
Zt = RDC + jwM12
At very low frequency, the STI is equal to the direct current resistance of the
braid.
With increasing frequency, two effects will occur :

|