[Skip to content]

TWI
Search our Site
.

Am I interfering with you?

TWI Bulletin, March 1985

- guidelines on grounding and shielding

by Andrew Perryman

Andrew Perryman, BSc, CEng, MIEE, is a Principal Research Engineer in the Control Engineering Department.

Unwanted electrical interference can affect the function and reliability of sensitive apparatus in an electrically noisy environment, such as in the vicinity of welding operations. The main sources of interference are described here, with methods by which their effect may be reduced.


A welding workshop presents many opportunities for the occurrence of unwanted electrical pickup. Welding power supply contactors; cables lying one upon the other; high frequency high voltage arc starting; and machine contactors - all are potentially hazardous regarding pickup. The level of pickup may be sufficient to destroy some apparatus even though there is no direct electrical connection. However, an understanding of how pickup can occur may help to reduce the effects of this problem.

There are two ways in which two circuits can be coupled without having direct physical contact - inductive and capacitive coupling. If there is contact between the circuits, resistive coupling may also occur. Each of these sources of interference is dealt with below.

Inductive coupling

When a current flows through a conductor it generates a magnetic field which is proportional to the level of the current ( Fig.1). If some of this field is cut by a loop then a voltage will be generated which is in proportion to the rate of change of field enclosed in the loop (which in turn is related to the rate of change of current level).

Fig.1. Magnetic field generated by passage of an electric current through a conductor. (H - magnetic field, I - current, A)

Fig.1. Magnetic field generated by passage of an electric current through a conductor. (H - magnetic field, I - current, A)

The voltage may be expressed as

b2633e1.gif

where

v is the induced voltage;

N is the number of turns, in this case, one;

and

b2633e2.gif

is the rate of change of magnetic flux enclosed by the loop. 

Ø (Webers) is more easily expressed in terms of the flux density B (Tesla) and the area A, where Ø = BA.

B is related to the field strength H (Ampères per meter) by the permeability of free space µ o (a constant):

B = µ o H

If any other medium was linking the loop to the current source a relative permeability (µ r ) would be added giving:

B = µ o µ r H

It is the level of H which is related to the current flowing, I. This is given by Ampère's law as

b2633e3.gif

for a long straight wire carrying current.

I is the level of current and r is the radius from the conductor to the sensing loop.

These then give an expression for voltage of

b2633e4.gif

Say the current flowing was sinusoidal having a maximum level Ι and frequency ω radians per second then:

I = Ι sin ωt

b2633e5.gif

Examining this simple equation highlights a number of significant points.

The voltage pickup ( Fig.2) is:

  1. Proportional to the loop area (A);
  2. Inversely proportional to the distance from the source of the field (r);
  3. Proportional to the frequency ( ω).
Fig.2. A changing magnetic field, H, when cut by a conducting loop will cause a voltage v to be generated within the loop, proportional to the area of the loop, A, and its distance from the current carrying conductor (r) and the rate of change of H

Fig.2. A changing magnetic field, H, when cut by a conducting loop will cause a voltage v to be generated within the loop, proportional to the area of the loop, A, and its distance from the current carrying conductor (r) and the rate of change of H

This last point confirms intuitive thinking. Any waveform may be reduced to a sum of contributing waveforms. The faster the edges of the waveform, the greater the component of the higher frequency elements required to synthesise it, and hence the greater voltage pickup of these high frequency components.

The presence of this type of loop is a common fault when monitoring, and can easily be overcome. Say, for instance, a voltage is being monitored, and the apparatus (such as an oscilloscope or ultraviolet recorder or a voltmeter) is earthed. In this situation, ( Fig.3) the area (A) encloses a field H which may not necessarily be related to the welding set being monitored. The loop of sensing wires and earthing will pick up a voltage which, in turn, will cause a current to flow around the loop disturbing the monitor's reading. This earth loop must be broken to prevent interference. However, safety factors must be observed when accomplishing this breaking of the earth loop. It may not be permissible to detach the oscilloscope's earth (E) (at the power input) as a means of breaking the loop. Consult your safety officer for further advice before taking this step. (This step also requires that the apparatus 'floats' at a potential controlled by the 'O'V being measured!)

Fig.3. The loop of sensing wire and earthing which may be present when using an oscilloscope to monitor voltage can generate interference as shown

Fig.3. The loop of sensing wire and earthing which may be present when using an oscilloscope to monitor voltage can generate interference as shown

A second means of overcoming the loop is to employ a differential input amplifier with good common mode rejection ratio. This then looks at the signal sensed by the leads and rejects the error voltage which has been picked up by both leads (the common mode voltage).

The leads making up a conducting loop should be run together as closely as possible to ensure that they pick up stray voltages identically. If they are separated then the area between them offers a path for the magnetic field to pass through, so generating a voltage ( Fig.4a). Running two wires alongside each other is not as good as twisting them together. This is easily explained when considering how the area between each crossover of the twist is in the opposite sense to the previous one so that the voltage generated in the first loop is cancelled by the next ( Fig.4b).

Fig.4. Running two wires alongside each other is not as effective as twisting them together for eliminating interference:
Fig.4. Running two wires alongside each other is not as effective as twisting them together for eliminating interference:

Fig.4. Running two wires alongside each other is not as effective as twisting them together for eliminating interference:

a) Two parallel wires give a large enclosed area, A, with no cancelling of induced voltages
b) Two twisted wires give small enclosed areas, A 21 and A 2 cancelling each loop voltage

These two methods of minimising interference can be combined as shown in Fig.5.

Fig.5. Interference arising from earthing loop can be minimised by using twisted wires with a differential input stage (as the latter has high input impedance, there is no induced current flow)

Fig.5. Interference arising from earthing loop can be minimised by using twisted wires with a differential input stage (as the latter has high input impedance, there is no induced current flow)

Summary of methods for avoiding inductive pickup

  1. Minimise areas for pickup;
  2. Minimise number of loops;
  3. Maximise distance of loop from source of alternating field;
  4. Minimise transmission of field, e.g.:
    a) Use short lengths of current carrying cable;
    b) Run lengths of cable closely for forward and return paths (preferably twisted). This reduces emission by adding opposite sense of field being generated by both forward and return paths;
    c) Use lowest possible frequency. The rate of change of current is proportional to the frequency and the current level.

Capacitive coupling

Between any two surfaces there is a capacitance. If a potential is applied between two surfaces, a charge Q appears on them. If the area of the surfaces is A and the distance between them d, the capacitance (C) between the plates is

b2633e6.gif

where D is the displacement charge density (Q/A);

E is the applied field (V/d);

and

ε is the permittivity between the surfaces.

ε is the product of the medium's permittivity ( εr) and the permittivity of free space ( ε o ).

ε o is linked to µ o (the permeability of free space) and the speed of light (c) by the relationship:

b2633e7.gif

By definition of the MKS system µ o = 4 πx 10 -7 henry/meter which then gives ε o as 8.85 x 10 -12 farads/meter where c is 3 x 10 8 m/sec in vacuo.

The stray capacitance relationships may be examined as the dual of the electromagnetic pickup source.

b2633e8.gif

then Q = CV

b2633e9.gif

where i is the current induced to flow by a rate of change of source voltage dV/dt.

Now substituting for C, we get the relationship

b2633e10.gif

If the applied voltage was sinusoidal having frequency ω radians per second and peak level

b2633e11.gif

then 

b2633e12.gif
b2633e13.gif

As before this simple equation highlights a number of significant factors i.e. that the current injection is:

  1. proportional to the common area of source and sensing surface (A);
  2. inversely proportional to the distance between the surfaces (d);
  3. proportional to the frequency ( ω).

This last point is important when considering any waveform as the sum of its harmonics. The sharper the edge of the waveform the larger the high frequency components and the increase of the level of current likely to be injected when considering it as a source.

Let us consider a voltage measuring system which is located near a TIG set with a high frequency, high voltage, arc starting mechanism ( Fig.6).

Fig.6. Stray capacitance and current paths induced when a voltage measuring system is located near a TIG set with a high frequency, high voltage arc starting mechanism

Fig.6. Stray capacitance and current paths induced when a voltage measuring system is located near a TIG set with a high frequency, high voltage arc starting mechanism

The stray capacitance will inject a current into the measuring system. If, however, a shield is constructed around the measuring system then the injected current may be diverted from the sensitive leads and carried to earth ( Fig.7). The injected current then flows in the shield and is only reflected into the measured signal when the stray current generates a magnetic field which cuts the sensing wires or measuring system.

Fig.7. Method of diverting injected current from measuring system by use of screening shield conducting to earth

Fig.7. Method of diverting injected current from measuring system by use of screening shield conducting to earth

The shield should ideally be a complete envelope with good current carrying qualities. Metal enclosures enhance immunity from this type of pickup and a great deal can be achieved if they are used carefully. A metal case alone is insufficient; the parts of the case, such as panels, must be electrically connected otherwise the panel, say, merely sits at some potential between earth and the source without contributing to the shielding ( Fig.8).

Fig.8. When the metal case used for shielding a monitor is not electrically connected, it will not contribute to shielding. The voltage of the panel will be related to the voltage of the source in the ratio of the stray capacitances

Fig.8. When the metal case used for shielding a monitor is not electrically connected, it will not contribute to shielding. The voltage of the panel will be related to the voltage of the source in the ratio of the stray capacitances

Vpanel = 

b2633e15.gif

x V source

The way in which the panel is connected is significant when considering shielding. The connection must pass current, and when current flows a magnetic field is induced which opposes the flow of the current, i.e. a stray inductance. This inductance is most troublesome at high frequencies which is also when most current is injected by the stray capacitance. Thus multiple current carrying paths are better than one ( Fig.9a and 9b). (The best solution is to bond the panel solidly to the case - say by soldering - to ensure even current flow around all the edges). Multiple contact may be achieved by a large number of points each with an electrically conductive path left clear around the fixing. An alternative is to have a large number of 'fingers' which wipe the edge of the panel as it is set home. This allows each 'finger' to make good electrical contact.

Fig.9. Methods of diverting induced current from monitoring systems:  Fig.9a) Induced current is diverted from the monitoring system by the single path (which has an inductance):

Fig.9. Methods of diverting induced current from monitoring systems:

Fig.9a) Induced current is diverted from the monitoring system by the single path (which has an inductance):

Fig.9b) Induced current is diverted from the monitoring system by the multiple paths. (Each path has inductance but the effective impedance to current is reduced by the increased number of available paths)

Fig.9b) Induced current is diverted from the monitoring system by the multiple paths. (Each path has inductance but the effective impedance to current is reduced by the increased number of available paths)

Summary of methods for avoiding capacitive coupling

  1. Minimise impedance of shielding path;
  2. Maximise distance of apparatus from source;
  3. Make shield equally spaced around sensing wires to reduce coupling of stray current's magnetic field;
  4. Use lowest possible frequency and voltage. The injected current is proportional to the level and rate of change of level of source voltage.

Resistive coupling

A current flowing through a resistance (R) induces a voltage drop (V) ( Fig.10). The voltage is proportional to the current (i) giving: V = iR.

Fig.10. Voltage drop resulting from current flowing through a resistance

Fig.10. Voltage drop resulting from current flowing through a resistance

The resistance is given by the material's coefficient of resistivity (ρ), the length under consideration (1) and cross sectional area A:

b2633e14.gif

The current flowing through the stray resistance may not be related to the welding source being monitored. It could be associated with the measuring system or the way in which it is connected to another system. It does, however, modify the sensed voltage, adding a component of error ( Fig.11).

Fig.11. Voltage dropped by current passing through resistance of sensing lead introduces an error voltage

Fig.11. Voltage dropped by current passing through resistance of sensing lead introduces an error voltage

Providing an alternative sensing path may alleviate this, but only provides an alternative path - the 'O'V sensing lead will still carry some of the return current.

The solution to these difficulties is similar to that recommended for avoiding inductive couplings, i.e. differential amplifiers or a re-referred earth position (check safety with local safety adviser before removing the earth).

Summary

If the mechanisms for coupling stray signals into measured quantities are considered, the effect of the phenomena may be counteracted. The errors will not be removed but will be reduced, the reduction required depending on the application and its requirements. The techniques described above may be used individually or in combination to achieve the required attenuation. The attenuation achieved using such techniques may be sufficient to enhance the reliability and function of sensitive apparatus in a noisy environment and must be considered when utilising sophisticated apparatus in the welding environment.