Filtering other lines--Power and Signal

Telecommunications, data and signal ports may all carry interference signals into and out of equipment. Reducing conducted or radiated emissions with differential mode filters is not practical if they are caused by the essential spectrum of the signal: screened cables and connectors will be required instead of or as well as filtering. Common mode filtering may however be applied to frequencies within the desired signal spectrum, since the wanted signals are differential rather than common mode. Common mode signals are always unwanted.

Power and 0V "bounce" noise in digital processing circuits is often overlooked when predicting the specification of signal filters. The wideband digital circuits used in many modern products (computers, PLCs, and increasingly in instrumentation) create a lot of internal noise. This tends to exit via I/O and other cables as common mode noise with a much higher frequency content and amplitude than would be expected from knowledge of the intended signals, and this should be considered when specifying a filter. Unfortunately, it is impossible to predict and can usually only be quantified by measurement, ideally by the manufacturer of the digital device, who should then take steps within his product to reduce these noise emissions to reasonable levels. System integrators and installers need to be aware, when considering filtering and shielding, that such spurious noises exist and are completely unrelated to the signals which are intended to be carried by the cables.

Filtering at interfaces is usually the responsibility of the equipment designer, but there are two areas where the systems engineer can make improvements if necessary: using filtered connector adaptors, and using ferrites.

--- Using filter connector adaptors and ferrites

Filtered connector adaptors

Various makes of filter adaptors are available for common connector styles: the D sub- miniature series are the most common and popular. These adaptors are very simple, being nothing more than parallel capacitors between each line of the connector and the connector shell. As such, they will only work properly if used against a connector port on the equipment whose shell is terminated to a zone boundary, or interface ground/earth reference. If there is no such connection- e.g., if the equipment is housed in a plastic case and the shell is unconnected- then the capacitors form a high-pass interconnection between all the pins on the connector. This is likely to affect the signal circuit, and won't provide any filtering.

The value of the capacitors is important: too high a value will also affect the desired signal, but too low a value will be ineffective for suppressing lower-frequency noise.

Values between 47pF and 2200pF are typical. These capacitors appear in parallel with the capacitance of each conductor to the screen in the cable, which is a function of cable length and specification. To properly specify the use of such an adaptor, you need to know the maximum capacitance to ground/earth that each circuit is able to drive. Multiply the cable length by the conductor capacitance to screen, and subtract this from the overall drive capacitance specification of the equipment to find the maximum allowable filter capacitance. For example:

Belden 9305 cable, capacitance to screen 164pF/m, length 10m; maximum circuit drive capacitance 2000pF:

Max. C_filter = 2000 - (164 x 10) = 360pF

As with filtered connectors themselves, most adaptors can be supplied with a limited range of capacitor values. It may be necessary to check the voltage rating of the capacitors, especially if the signal line is intended to be isolated. Variations on the design of adaptor may include series chokes in one or both legs of each line, n or L- configurations, and varistor clamps in parallel with each capacitor.

Using ferrites

The most usual application of a ferrite component is as a sleeve over a cable to attenuate common mode currents on the cable. A particularly useful type of ferrite component is the split tube in a plastic "clip-on" housing. They are very easy to apply as a retro-fit to a cable (and to remove when found not to do much), and are available in a wide range of shapes and sizes, including fiat types for ribbon cables.

The effect of magnetic material around a conductor

Current flowing through a conductor creates a magnetic field around it. Transfer of energy between the current and the magnetic field is effected through the "inductance" of the conductor-- for a straight wire the self-inductance is typically 20nil per inch. In a non-permeable material, such as air or a plastic sheath, the magnetic field strength and magnetic flux density are equal. Placing a magnetically permeable material around the conductor increases the flux density for a given field strength and therefore increases the inductance.

Ferrite is such a material; its permeability is controlled by the exact composition of the different oxides that make it up (ferric, with typically nickel and zinc) and is heavily dependent on frequency. Also the permeability is complex and has both real and imaginary parts, which translate into both inductive and resistive components of the impedance "inserted" into the line passed through the ferrite. The ratio of these components varies with frequency -- at the higher frequencies the resistive part dominates (the ferrite can be viewed as a frequency dependent resistor) and the assembly becomes lossy, so that RF energy is dissipated in the bulk of the material and resonances with stray capacitances are avoided or damped.

--- Typical ferrite impedance with frequency

The effect of cable currents

Elsewhere the distinction between common and differential mode in cables is discussed. This distinction is particularly relevant to the application of ferrites.

Differential mode:

The magnetic field produced by the intended "go" current in each circuit pair is substantially cancelled by the field produced by its equal and opposite "return" current, provided that the two conductors are adjacent. Therefore any magnetic material, such as a ferrite sleeve, placed around the whole cable will be invisible to these differential mode currents. This will be true however many pairs there are, as long as the total sum of differential mode currents in the cable harness is zero.

Placing a ferrite around a cable, then, has no effect on the differential mode signals carried within it.

Common mode:

A cable will also carry currents in common mode, that is, all conductors have current flowing in the same direction. Normally, this is an unintended by-product of the cable connection, and the current amplitudes are often no more than a few microamps, but are usually the main cause of interference problems. The source of such currents for emissions is usually either…

++ ground/earth-referred noise at the point of connection, which may have nothing to do with the signal(s) carried by the cable, or

++ imbalance of the impedance to ground/earth of the various signal and return circuits, so that part of the signal current returns through paths other than the cable harness.

A screened cable may also carry common mode currents if the screen is not properly terminated to a noise-free reference. Even though the currents may be small, they have a much greater interfering potential because their return path is essentially uncontrolled. Also, incoming RF or transient interference currents are invariably generated in common mode and convert to differential mode (and so affect circuit operation) due to differing impedances at the cable interfaces, or within the circuit.

Common mode currents on a cable do generate a net magnetic field around the cable. Therefore, a ferrite inserted around the cable will increase the cable's local impedance to these currents.

Circuit impedances

In the same way as was discussed earlier with respect to mains filters, when a ferrite is placed on a cable it functions between source (equipment) and load (cable) impedances. A quick glance at the equivalent circuit section that maximum attenuation due to the simple impedance divider will occur when these impedances are low. For cable interfaces, low source impedance means that the ferrite should be applied near to a capacitive filter to ground/earth or to a good screen termination. For open or long cables, the RF common mode load impedance varies with frequency and cable length and termination: a quarter wavelength from an open circuit, the impedance is low, a few ohms or tens of ohms; a quarter wavelength from a short circuit, the impedance is high, a few hundred ohms. (This property of cables is well known to antenna designers but looks like magic to the rest of us.) Since you do not normally know the length and layout of any cable that will be attached to a particular interface, and since the impedance is frequency dependent anyway, it is usual to take an average value for the cable impedance, and 150 -ohm has become the norm. This is the common mode impedance: it has nothing to do with the cable's characteristic impedance or any differential circuit terminations.

--- The equivalent circuit for ferrites

Ferrite impedances rarely exceed a few hundred ohms, and consequently the attenuation that can be expected from placing a ferrite over a cable is typically 6-10dB, with 20dB being achievable at certain frequencies where the cable section a low impedance. --- the actual attenuation for two types of core at two different circuit impedances. Note the different vertical scales for the two plots.

--- Ferrite attenuation versus impedance

Choice and application

Size and shape

There are two rules of thumb in selecting a ferrite for highest impedance:

++ where you have a choice of shape, longer is better than fatter;

++ maximum impedance demands the maximum amount of material in a given volume.

The impedance for a given core material is proportional to the log of the ratio of outside to inside diameter but directly proportional to length. This means that for a certain volume (and weight) of ferrite, best performance will be obtained if the inside diameter fits the cable sheath snugly, and if the sleeve is made as long as possible. A string of sleeves is perfectly acceptable and will increase the impedance pro rata, though the law of diminishing returns sets in with respect to the attenuation.

Number of turns

Inductance can be increased by winding the cable more than one turn around a core; theoretically the inductance is increased proportional to the square of the number of turns, and at the low frequencies this does indeed increase the attenuation. But it is usual to want broadband performance from a ferrite suppressor and at higher frequencies other factors come into play. These are:

++ the core geometry already referred to; the optimum shape is long and snugly-fitting on the cable, and this does not lend itself to multiple turns;

++ more importantly, inter-turn capacitance, which appears as a parasitic component across the ferrite impedance and which reduces the self resonant frequency of the assembly.

The normal effect of multiple turns is to shift the frequency of maximum attenuation downwards. It will also increase the value of maximum attenuation achieved but not by as much as hoped. The source and load impedances are critical in determining the effect: the lower the impedances, the less the effect of parasitic capacitance. To gain improved attenuation at very high frequencies (say, above 300MHz) it is usually better to string more ferrite tubes or toroids along the set of conductors, with only one pass of the cables through each.

Capacitance to ground/earth

Because a ferrite material is in fact a ceramic, it has a high permittivity as well as permeability, and hence will increase the capacitance to nearby objects of the cable on which it is placed. This property can be used to advantage especially within equipment.

If the ferrite is placed next to an ground/earthed metal surface, such as the chassis, an L-C filter is formed which uses the ferrite both as an inductor and as a distributed capacitor. This will improve the filtering properties compared to using the ferrite in free space. For best effect the cable should be against the ferrite inner surface and the ferrite itself should be flat against the chassis so that no air gaps exist; this can work well with ribbon cable assemblies.

Saturation

As with other types of ferrite, suppression cores can saturate if a high level of low- frequency current is passed through them. At saturation, the magnetic material no longer supports an increase in flux density and the effective permeability drops towards unity, so the attenuation effect of the core disappears. The great virtue of the common mode configuration is that low frequency currents cancel and the core is not subjected to the magnetic field they induce, but this only happens if the core is placed around a cable carrying both 'go' and 'return' currents. For three-phase star connected power all three phases and the neutral must pass through together. For heavy power cables, best performance from ferrites is achieved when all the relevant conductors are bunched tightly together and held in the centre of the toroid or tube, away from the walls. If the ferrite gets hot, it is not substantial enough or else the cables through it need better control.

If you must place a core around a single conductor (such as a power supply lead) or a cable carrying a net low frequency current, be sure that the current flowing does not exceed the core's capability; it is usually necessary to derive this from the generic material curves for a particular core geometry.

Next:

Prev: Filtering Mains Power

top of page  Article Index   Home