Screened cables are used to reduce coupling of the inner cable conductors with the environment they pass through, but the screen functions differently at high frequencies and at low frequencies, and also has different effects as between electric field and magnetic field coupling.

Low frequency cable shielding

For low frequencies an overall screen, grounded only at one end, provides good shielding from capacitively coupled interference but none at all from the magnetic fields, which induce a noise voltage in the loop formed when both source and load are grounded. To shield against a magnetic field, both ends of the screen must be grounded. This allows an induced current (I_S) to flow in the screen which will oppose the current induced in the centre conductor, effectively minimizing the loop area seen by the complete signal circuit. This effect begins to become apparent only above the cable cut-off frequency F_c = 2π·ω_c, which is a function of the screen inductance and resistance and is around 1 - 2kHz for braided copper screens or 7 - 10kHz for aluminium foil screens. Above about five times the cut-off frequency, the interference induced in the centre conductor is constant with frequency, which means that the magnetic shielding effect is increasing.

The same principle applies when shielding a conductor to prevent magnetic field emission. The return current must flow through the screen, and this will only occur (for a circuit which is grounded at both ends) at frequencies substantially above the shield cut-off frequency.

But in general, circuits are not separately connected to ground at each end of a cable. To minimize low frequency magnetic field pick-up, one end of the circuit should be isolated from ground, the circuit loop area should be small, and the screen should not form part of the circuit. You can best achieve this by using screened twisted pair cable with the screen grounded at only one end. The twisting minimizes magnetic coupling, and the screen will reduce external capacitive coupling, but there is now the question of which end is best grounded. Any voltage difference between the screen and the inner conductors will compromise the shielding since there is close capacitive coupling between them. Therefore the screen should be grounded at the same end as the circuit, to minimize this voltage difference along the cable.

There is no absolute reason not to ground the screen at each end of the cable, but there may be a reason related to installation practice. If there is a significant voltage difference between the earths at each end of the screen, current will flow in the screen if both ends are connected. The impedances involved could be very low, and the current that flows could be enough to damage the cable. The preferred solution is to install the cable in a mesh-bonded ground network using parallel earth conductors (PECs), which will minimise the voltage differences, but not all installations follow this practice. As a product designer, you have to be aware that this limitation seriously affects the performance of screened cable.

High frequency cable shielding

Once the cable length approaches a quarter wavelength at the frequency of interest, screen currents due to external fields become unavoidable. A high impedance at one end of the cable becomes transformed into a low impedance a quarter wavelength away - see the description of transmission line properties - and screen currents flow in a standing wave pattern. The magnitude of the current is related to the characteristic impedance of the transmission line formed by the cable and any external ground reference. Even below resonant frequencies, stray capacitance can allow screen currents to flow.

The distribution of the cable screen current and voltage will be determined by the boundary conditions of the system: that is, current will fall to zero at locations of open circuit and will be a maximum at low impedance points. Thus, current will be greatest when the physical conditions correspond to multiples of half- and quarter-wavelengths at particular frequencies. The voltage on the screen will be greatest at the opposite, high impedance points. Since these points vary along the length of the cable as the frequency changes, it is impossible to devise a screen configuration open at one end or the other which provides broadband protection. To get proper RF performance, you have to connect the screen at each end to the screening metalwork of the connected unit.

But the good news is that at high frequencies the inner and outer of the screen are isolated by the skin effect, which prevents currents on the surface from passing into the bulk of the conductor (skin effect is also discussed in the section on enclosure shielding).

skin depth δ = 6.61/ μ_r · σ_r · F) cm

where μ_r is relative permeability and σ_r is relative conductivity of the screen, F is in Hz; e.g for copper at 10MHz, δ  is 0.0208mm

The current density within the screen reduces by 8.6dB for every skin depth penetration from the surface. Therefore signal currents on the inside of the screen do not couple with interference currents on the outside, and vice versa. Both-end grounding or multiple grounding of the screen does not introduce interference voltages on the inside to the same extent as at low frequencies (see above). This desirable effect is compromised by a braided screen due to its incomplete optical coverage and because the strands are continuously woven from inside to out and back again. It is also more seriously compromised by the quality of the screen ground connection at either end.

Transfer impedance

The screening performance of shielded cables is best expressed in terms of surface transfer impedance Z_T (STI). This is a measure of the voltage induced per unit length on the inner conductor(s) of the cable with respect to the screen, by an interference current flowing down the cable outer screen.

A perfect screen would not allow any voltage to be induced on the inner conductors and would have an STI of zero. Practical screens will allow some coupling. STI will vary with frequency and is normally expressed in milliohms per metre length. At low frequencies it is equal to the DC resistance of the screen, but at higher frequencies the STI is dominated by the effect of mutual coupling between the screen and the inner conductor. The parameters which affect the STI are

mutual inductance - screen to inner

mutual capacitance - screen to inner

leakage capacitance - inner to external (“optical coverage”)

skin effect and the compromising effect of braid weave

The graph here compares STI versus frequency for various types of cable screen construction. The decrease in STI with frequency for the better performance (double braided) screens is due to the skin effect separating signal currents on the inside of the screen from noise currents on the outside. The subsequent increase is due to field distortion by the holes and weave of the braid. A solid copper screen does not suffer from this, and its STI continues to reduce with increasing frequency.

Note that most inexpensive types have a worsening STI with increasing frequency (a typical single wire braid is the popular RG58/U coax). A laminated foil screen with drain wire is approximately 20dB worse than a single braid, due to its higher resistance and to the field distortion introduced by the drain wire, which carries the major part of the longitudinal screen current. Because it is not coaxial, the drain wire has much worse magnetic coupling with the inner conductors.

Once the frequency approaches cable resonance then STI figures become harder to measure as both the cable itself and the outer (screen) circuit must be perfectly matched.

For a braided construction Z_T in the higher frequency region is dominated by the leakage inductance between the overall braid and the inner conductor. For a good single layer braid (the quality is determined by the weave of the braid) this inductance is of the order of 1nH/m.

Try your own calculation for skin depth:

Click here for Questions