GB2444753A

GB2444753A - Audio amplifiers with compensation for Earth loop currents

Info

Publication number: GB2444753A
Application number: GB0617822A
Authority: GB
Inventors: Edward Campbell Bentley
Original assignee: Individual
Current assignee: Individual
Priority date: 2006-09-11
Filing date: 2006-09-11
Publication date: 2008-06-18
Anticipated expiration: 2026-09-11
Also published as: GB2444753B; GB0617822D0

Abstract

A simple and inexpensive technique for compensating the effect of Earth loop currents in an amplifier arrangement comprises the insertion of a resistance R8 in the Earth line associated with the amplifier U1A so that both ends of the load R14 are subjected to identical signals caused by Earth loop current flowing in the loop V3, R12, R1, R8. The stray Earth-side resistance of the signal input circuit is denoted R1. Although the unwanted signal developed in R1 is amplified by U1A and fed to the top of the load R14 an identical signal is generated in the resistance R8 and coupled to the Earthed end of the load R14 so that no net signal due to Earth currents flows in the load. The Earth currents may be due to electromagnetic fields or may be due to currents injected into the Earth network by other amplifiers. The technique may be useful in instrumentation and control circuits in avionics, computers and vehicle applications.

Description

1 2444 Ainpi if iers This invention relates to audio power amplifiers,

using both transistors and valves.

With valve circuits in power amplifiers it seems that the lower the harmonic distortion factor the better the resulting amplifier would sound. This remains true even when two such amplifiers are fed from a common power supply and used as a stereo pair.

Transistor amplifiers can easily be made with harmonic distortion factors much lower than that displayed by the best valve amplifiers, however, when a stereo pair of such amplifiers was run from a common power supply the resulting sound was found to be harsh and discordant; this was true even when use of one such amplifier in mono gave results which were indistinguishable from a good mono valve amplifier in side by side comparison. A further puzzle was that the way one wired up the interconnections between two transistor mono amplifiers *. to run them from a common power supply would affect the sound quality. Despite much empirical experiment it has not previously been possible to achieve a level of sound * 25 quality from a stereo transistor amplifier using a common power supply comparable to that obtained from even a quite modest valve stereo unit.

* ***** * *

Background

It was easy to find convincing proof in the literature eg Wireless World 11/1977 "Audible amplifier distortion is not a mystery" to show that even when transistor amplifiers were first introduced in the 1960's mono circuits could be readily made which introduced no audible distortion; and designs have improved since the 1960's technically.

It appears that when transistor stereo amplifiers were first introduced there was widespread concern over the sound quality when compared to the previous generation of valve units. Tests (which showed that good mono transistor units introduced quite inaudible distortion) were carried out to try and convince the world at large that transistor stereo units would also sound perfect.

The debate (valve vs transistor) ran for many years and still continues, see for example Electronics and Wireless World, Vol 94 p. 694/695 Vol 94 "Science and subjectivism in audio engineering", DRG SELF pp694/695, which once again used the inaudibility of distortion from a mono transistor amplifier to prove' that such amplifiers were perfect even when run in stereo from a common power supply.

There is a fundamental difference between valve and transistor units-the valve circuits work at high voltage low current whereas the transistor circuits run at lower .. : voltage and higher current to produce the same level of output power. Also valve circuits use an output transformer keeping the high loudspeaker currents isolated from the rest of the circuit. It has been * considered whether there might be some problem connected with current flow which tended to make for problems in running two mono transistor amplifiers in a single box from a common power supply. Wireless World 01/1978, p44, Letter of Mr Baxendall suggested that earth loop effects would contaminate the operation of an otherwise good sounding transistor amplifier, so this was an area worth exploring further.

Prior art was reviewed, such as described in Wireless World dating from the early 1960's; when transistors with significant power handling capacity became available; these enabled engineers to design transistor power amplifiers for the first time. Wireless World 01/1965, page 7 was found in which a small discussion was given relating to the present problem and its cure.

In fig 13 of Wireless World 01/1965, page 7 an earth loop will exist when two transistor power amplifiers are fed from a common power supply, if they are fed with stereo signals coming from a source with a common earth. The loudspeaker earthy' return current from one channel can either flow directly back to the power supply zero point through the impedance of the direct connecting wires, or it can flow round the loop through the preamplifier earth connections and back to the power supply zero through the earth conductor of the second channel. The 1965 author *"e found that' the current pulses in the earth line are asymmetrical, causing severe even harmonic distortion *..

(which)... causes each channel to become distorted by the even harmonic products of both channels, causing :. unpleasant distortion particularly when one channel has a ::.::; transient such as a cymbal clash' Prior attempts at solution This phenomenon would account for the present problem, which has been clearly well known for many years. The 1965 author, Dinsdale, found a partial cure by inserting a small resistance into the earth loop in such a way as not to accentuate the even-harmonic distortion', see riB in figl4 of Wireless World 01/1965, page 7,8 a solution followed (see Figure 1, the circuit diagram) in the well known QUAD 303 power amplifier(R115,2.2ohm resistor). The QUAD 303 was one of the amplifiers tested in the 1970's (see Wireless World 11/1977 "Audible amplifier distortion is not a mystery") and found to introduce no audible distortion in mono. However the proposed cure in Wireless World 01/1965, page 7,8 only reduced the extent of the problem, and whilst not eliminating it did increase the amplifier distortion on mono signals.

To assess the extent of the problems remaining after adopting the cure from Wireless World 01/1965, page 7,8, tests were run on a QUA]J303 unit. In tests, the output from the silent channel' when the other channel is delivering 5Vpk/pk at 50 Hz into a 10 ohm resistor shows spurious signals-a basic structure looking like even harmonic distortion being +ve going humps above a baseline, and spikes which will represent charging pulses from the power supply. * **. * S S...

What, if any, will be the audible effect of this stereo S...

: operation' distortion? :. John Linsley Hood wrote of experiments which he carried * out (see Electronics World 11/1998, p964) showing that *....S * low levels of harmonic distortion of about 0.1% would normally be completely inaudible because of masking' by the actual programme material being reproduced.

However it was found that in the special case where the added distortion came into frequency synchronisation with a multiple of the low frequency input signal, then even very low levels of distortion-below 0.1%-would give rise to a change in timbre of the low frequency tone being amplified. In the case of the present problem, the distortion introduced by stereo operation was in the form of harmonics of the original input signal to each channel, and would necessarily be in frequency synchronisation with a multiple of the input signal-thus even when, as in the QUAD3O3, the problem is reduced to a low level, one would expect some kind of change in timbre to be observed in stereo operation.

When a class B amplifier such as the QUAD 303 is operated with very low quiescent current, so that crossover distortion is produced, this distortion will be created in synchronisation with a multiple of the input frequency being amplified. It was possible to verify Electronics World 11/1998, p964 by using a QUAD 303 unit in mono and reducing the quiescent current whilst continuing to listen to programme material. As the bias current fell below the correct setting a lightening' of the musical pitch was heard-bass notes appeared to be of higher frequency with less impact' *..

Extent of remaining problems :. Regarding the present problem, and considering the tests * mentioned above on the QUAD 303 using the Wireless World * S....

* 01/1965, page 7,8 solution the stereo operation distortion is about l3mV peak with an RNS value of approximately 4.1/root2.l3mV =2.3mV; the base signal was 5v pk/pk applied to the other channel. Assume this signal to have been applied to the silent channel at the same time, it has an RNS value of ".5.1/root2 =1.77V RNS. We have a harmonic distortion figure of 0.0023/1.77% =0.13% approximately. This is because (from Wireless World 01/1965, page 7,8) each channel becomes distorted by the even harmonic products of both channels. According to Electronics World 11/1998, p964 this level of distortion will be audible.

Previous alternative attempts at solution An alternative solution to the earth loop problem was suggested (Wireless World 02/1965, p84 and Wireless World 02/1965, p82) wherein a single point star' earth point was used to prevent the unwanted earth currents circulating in the vulnerable input stage ground conductors. This solution looks good on paper and has been adopted by many manufacturers of good quality transistor stereo power amplifiers. This scheme was tested, but was disappointed-results were still not comparable to a good quality valve stereo amplifier. The Wireless World 01/1965, page 7,8 author's comment on the Wireless World 02/1965, p84 suggestion was very favourable-see addendum to Wireless World 02/1965, p84 * *** but it noted that the even harmonic distortion signal was * *** still present but would have less effect being developed **** across low impedance connecting wires. 25

Consideration of single point earthing suggests that the scheme would indeed work well if the star point had * S....

* negligible impedance. When measured, however, a typical soldered joint has an impedance of about 0.01 ohms from one incoming wire to an adjacent one. In a setup using single point earthing loudspeaker currents in the ampere range passing from input to output give millivolt order voltages between eg vi and vO. The effect of such voltages will be to induce distortion signals into each channel's input proportional to the current flowing between PSU 0 VOLTS and the 0/P stage ground connection for each channel. This is because the input stage ground will not be at the same potential as the loudspeaker earth return, giving distortion which negative feedback cannot remove being outside the NFB loop.

A further problem with single point earthing can be seen from the diagram in Wireless World 02/1965, p84. An earth loop exists from the star point around the early stages of the power amplifier and back via the other channel to the star point. Any difference in potential between the two wires leading from the star point to the early stages will cause a current to flow in the ground conductors of the preamp as before; in addition any stray electromagnetic fields within the amplifier enclosure will induce unwanted currents in the loop which each channel will amplify as if they were wanted signals. The severity of this problem depends upon the physical layout of the stereo amplifier; it will be serious if any high current carrying conductors run close to the loop. * * S..

The writer found an article in Electronics Today S...

.. : International (1984, page 34) in which John Linsley Hood *25 alleviated some of the problems with star earthing by :. taking the input stage earth connection from the star * point via a single conductor to prevent currents S.. *S.

* circulating around an input loop, and by taking one conductor from the star point to the zero point in each of the two power amplifiers used, the loudspeaker return being to each amplifier power zero point. Capacitative decoupling was applied to each of the power amplifiers thus reducing the AC currents flowing in the star point.

This scheme was tried and was found to be a useful improvement, thus confirming the ideas being formulated so far. However the results were still found to be inferior to a good valve system, so further literary research was carried out.

Consideration of Wireless World 02/1965, p82 shows that a further earth loop will exist if an earthed ancillary unit such as a CD player is connected-any fluctuations in the potential of the GRAND EARTH POINT' compared to mains earth potential will cause an earth current to flow along the Earth mains lead-and also along the coaxial cable outers to the earthed ancillary unit. This gives rise to a superimposed input voltage which the amplifier treats as a signal. If, as would be preferable, the amplifier was grounded to mains earth at its chassis input connection, then fluctuations in potential between mains earth and the GRAND EARTH POINT' will once again give rise to unwanted voltages being superimposed on the amplifier input., as currents flow through the internal loop. * S * I*.

Thus, the problem of an earth loop was still present. S...

Further research revealed Electronics World 05/1996, page S...

402 "Scissors overcome earth loop problems"; a little known solution to earth loop problems described as earth scissors'. In essence the earth scissors when compensating a following amplifier measure the current * flow along the earth line (see Figure 2) and produce an in phase output voltage at point B. This corresponds to the voltage set up along the ground conductors of the following equipment as measured at point A. The result is to offset the effect of the ground current in the input circuit of the following equipment. The common mode rejection ratio of the following equipment needs to be high for this scheme to work well. A diagram of the operation is shown in Figure 3.

In this example the basic amplifier' has one of its inputs between R4 and R5 both of equal value; then R9 was adjusted until common mode signal rejection in the basic amplifier removed the unwanted earth signal represented here by source v3. Best cancellation was found to occur at almost the precise values shown in Electronics World 05/1996, page 402 "Scissors overcome earth loop problems", but the success of this scheme will depend upon the internal impedances in the following amplifier( represented by R4 and R5) being in a 1:1 ratio-if they are not the scissors' is less successful with the given component values, which would need adjusting.

It is an object of the present invention to address the above-mentioned disadvantages.

According to a first aspect of the present invention, there is provided an amplifier having ground current : ... compensation means. * S S...

Preferably, the ground current compensation means are * 25 external to the amplifier. S..

Preferably, the ground current compensation means are passive.

Preferably, the ground current compensation means are in the form of a suitably chosen resistor. Preferably, the resistor has a value suited to the amplifier.

The amplifier may be a multi-channel amplifier. The amplifier may be an audio amplifier The amplifier may be inverting. The amplifier may be non-inverting. The amplifier may be a power amplifier.

According to another aspect of the invention, there is provided a method of compensating for ground currents in an amplifier, comprising: a) determining the forward gain of an amplifier; b) determining the impedance of input circuit ground conductors of the amplifier; C) calculating a value for a compensating resistor for the amplifier based on the determined values; and d) applying a suitable resistance to the amplifier.

A method of calculating a compensation resistance for an amplifier comprises steps a) to C) of the former aspect.

All of the features described herein may be combined with any of the above aspects in any combination. * .

For a better understanding of the invention, and to show S...

.. : how embodiments of the same may be carried into effect, *25 reference will now be made, by way of example, to the accompanying diagrammatic drawings in which: ***..

* Figure 1 is a circuit diagram of a prior art attempt to prevent an earth loop; Figure 2 is a circuit diagram showing an "earth scissors"

prior art attempt to solve the earth loop problem;

Figure 3 is a circuit diagram of another implementation of an earth scissors; Figure 4 is a circuit diagram of an earth loop compensation circuit according to an aspect of the present invention; Figure 5 is a circuit diagram of another implementation; Figure 6 is a circuit diagram of an implementation for two channel amplifier; Figure 7 is a circuit diagram for an implementation for a basic compensated non-inverting amplifier; Figure 8 is a circuit diagram for a simplified version of the Figure 7 circuit; Figure 9 is a circuit diagram of an implementation for two channel use with a non-inverting amplifier; a. * *

::: Figure 10 is a circuit diagram showing an implementation to a real non-inverting power amplifier design; * * a *. S *25 Figure 11 is a circuit diagram showing an implementation for two channel use with a non-inverting amplifier; and S.....

* Figure 12 is a circuit diagram of an implementation for 5 channel use.

Given that the purpose of the earth scissors is merely to produce a cancelling signal to offset the unwanted ground signal' the surprising realisation has been made that it might be possible to provide an internally generated cancellation signal produced by the amplifier itself. This would avoid the need for an external scissors' circuit.

The proposed circuit is shown in Figure 4. In this circuit the basic gain is about 20 (Rl1/(R10 R9)).R9 represents the source impedance of the stage eg a preamplifier driving the circuit-the actual value of this will affect the gain; it was assumed that the circuit was driven by a source impedance of 100 ohms. Accordingly x amps current through the earth loop simulated by V3,R12, gives x.Rl volts across the input and of course 20.x.Ri volts after amplification across the load. Then a suitable value for R8, so that 20.x.Rl volts was dropped across it when x amps current flows through it, will raise the potential of the bottom of the load resistor R14 by just the amount that the top of R14 is raised by the spurious signal, ie no additional current will flow due to the spurious earth signal in the load R14. Here if Ri = 0.01 ohms and gain=40/2.1=19.048 the calculated value for R8 would be R1.19.048 =0.1905 ohms. Using : Pspice, a proprietory simulation package, gave a i**.

practical value for R8 of 0.191 ohms, to achieve cancellation of the unwanted earth current signal. This * 25 correlates well.

Tests on a real system * *04*S* * . For this simple arrangement an inverting amplifier is needed. The QUAD 303 is an inverting amplifier, and in Figure 1, Rii5 the 2.2 ohm resistor, was replaced by a variable resistor in what was to be the silent' channel.

This resistor was conveniently placed between the bottom of the voltage amplifier output and the bottom of the input of the unity gain output stage comprising TR1O3,104,107,105,106,1L and 2L. The source (the output resistance of a QUAD 33 prean'tp in this case) resistance was set to 1.7k which was placed across the input of the silent channel of the QUAD 303.As before the active channel of the QUAD 303 was arranged to give 5V pklpk across a 10 ohm resistor, and the value of the variable resistor was adjusted until the best nulling of the unwanted earth loop signal was achieved. A 50 Hz signal was used, and the unwanted signal is down in the noise' which provides a very satisfactory result. As was expected the nulling effect did not depend on the level of output in the active channel. The value of the variable resistor was then measured and a fixed component of that value was installed. The procedure was then repeated for the other channel so that both were compensated. The requisite values for each channel in the first unit dealt with were 0.28 ohms for the compensating resistor. Other units studied required 0.27 ohms and 0.3 ohms. The precise value will depend of course on the actual forward gain of the unit in question, and the impedance of the input circuit ground conductors.

Given the order of magnitude difference between the value * 25 of the compensating resistor required for a QUAD 303 and *** the value of Rl15 used in the original circuit, it is hard to say that the manufacturers of the QUAD 303 thought of this scheme themselves. Putting in appropriate values to simulate the QUAD 303 we get the values shown in Figure 5.

With the observed 0.28 ohm value of compensating resistor, (R8 in Figure 5) and the known value of the gain(30) we find that the input cable earth impedance (Ri) needs to be 0.0093 ohms, for nulling of the spurious ground signal, a realistic value.

Simulating this using Pspice gives a spurious output signal of 3mV pk/pk.

Putting in the QUAD 303 value of 2.2 ohms for R8 in the above diagram gave a spurious output level of 3.2V pk/pk which is about 1000 times worse and so clearly indicates that the problem had not been properly addressed.

What has been found, therefore is a means of removing the effects of ground loop currents at negligible additional production cost. The currents whose effects are removed include those induced in the ground loop by external

electromagnetic fields. Figure 6 shows a simple

:. connection scheme for a 2 channel amplifier following the QUAD 303 basic layout using a single power supply. Star earthing is not used so none of the abovementioned ... :20 drawbacks of this system apply. The amplifier negative feedback works just as if zero currents flow in the overall earth loop so precise stereo operation is achieved for 2 mono amplifiers in a single box. * *

The resulting sound is very good. The solution found, which can work equally well for any number of amplifiers in a single box fed from a single power supply, since each channel is internally compensated, is surprisingly good at reducing noise.

Implications The solution found may well have applications in consumer electronics, where it will apply whenever more than one amplifier is to be run in a given equipment in stereo or multi channel use (home cinema; car stereo; domestic audio visual etc) .There could well be industrial applications in instrumentation and control wherever unwanted ground currents appear; computing, avionics and vehicles spring to mind.

Further developments for Non Inverting amplifiers As will be evident from the above discussion work has been done on the simple inverting amplifier configuration. This has the minor drawback that amplifier gain depends upon source impedance of the signal feed to the amplifier, so in cases where this cannot be well defined, perfect cancellation of ground current signals cannot be relied upon. The obvious cure in such cases is to use a buffer in front of the compensated amplifier to give defined source impedance; alternatively the scheme will work in the non inverting configuration, wherein the system gain does not depend on the source impedance, :20 making the compensation scheme more versatile. One non inverting configuration is shown in Figure 7, there may well be others. S. * S * ..*

Here, the cancelling signal is taken across R2 since the non inverting configuration requires a phase inverted cancelling signal compared to the inverting configuration.

The above gave an unwanted signal of ±300mV when R2 was bypassed and ±0.75 my when as shown. In Figure 7 Ri and R8 can be combined with the same result to give the circuit of Figure 8.

A two channel non inverting amplifier combination is shown in Figure 9.

V3 represents the unwanted earth current signal, again cancellation is effective in both channels at once.

Finally, an example is shown in Figure 10 of application of the non inverting cancellation scheme to a real amplifier design.

Cancellation of the ground current signal is again effective.

A FURTHER NON INVERTING AMPLIFIER CONFIGURATION

In this version, shown in Figure 11, for the top' amplifier, any unwanted loop currents will set up offsetting voltages across (R5 +R14) and Rl5,which can cancel the effects of the loop currents if the ratio of (R5 + R14) and R15 is correctly chosen. Similarly, for the bottom' amplifier, unwanted loop currents will set up offsetting voltages across (R15 + R9) and R14 which again can cancel the effects of I. * the loop currents if the ratio of (Rl5 + R9) to R14 is properly chosen.

S

*****S * 25 The advantage here is that all of the amplifier signal input earth connections can be at the same potential and earthed to the equipment chassis; and the cancellation effect does not depend upon the source impedance of the source driving the amplifier, this being a non inverting arrangement.

The only drawbacks are:-that this arrangement only works for 2 channels, ie ordinary stereo; and current pulses from psu common zero (shown as a big black dot) to ground are not cancelled. This latter is not a problem since these current pulses will in practice be small and they do not appear as harmonics of the signal being axnplif led.

DEVELOPMENT OF THE SCHEME FOR MULTI CHANNEL USE

The scheme may be readily adapted for multi channel use to run a multiplicity of amplifiers in a single box from a single power supply. An example for use with 5 channels is shown in Figure 12.

The embodiments described above show that for very little extra cost an amplifier having more than one channel can be significantly improved by the addition of a suitably calibrated resistor.

Attention is directed to all papers and documents which are filed concurrently with or previous to this :. specification in connection with this application and which are open to public inspection with this specification, and the contents of all such papers and *:::2o documents are incorporated herein by reference.

All of the features disclosed in this specification (including any accompanying claims, abstract and * drawings), and/or all of the steps of any method or process so disclosed, may be combined in any combination, except combinations where at least some of such features and/or steps are mutually exclusive.

Each feature disclosed in this specification (including any accompanying claims, abstract and drawings) may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise. Thus, unless expressly stated otherwise, each feature disclosed is one example only of a generic series of equivalent or similar features.

The invention is not restricted to the details of the foregoing embodiment(s). The invention extends to any novel one, or any novel combination, of the features

disclosed in this specification (including any

accompanying claims, abstract and drawings), or to any novel one, or any novel combination, of the steps of any method or process so disclosed. * * * *** **** * S .5. I... * S I S. S I. S. * . * S.. *

S.....

S S

Claims

Claims 1. An amplifier incorporating passive ground current

compensation means.
2. An amplifier as claimed in claim 1, in which the ground current compensation means are external to the amplifier.
3. An amplifier as claimed in claim 1 or claim 2, in which the ground current compensation means are in the form of a suitably chosen resistor.
4. An amplifier as claimed in claim 3, in which the resistor has a value suited to the amplifier.
5. An amplifier as claimed in any preceding claim which :..::: is a multi-channel amplifier.

*:::20
6. An amplifier as claimed in any preceding claim which is an audio amplifier.
7. An amplifier as claimed in any preceding claim which S..... . * * is inverting.
8. An amplifier as claimed in any one of claims 1 to 6 which is non-inverting.
9. An amplifier as claimed in any preceding claim which is a power amplifier.
10. A method of compensating for ground currents in an amplifier, comprising: a) determining the forward gain of an amplifier; b) determining the impedance of input circuit ground conductors of the amplifier; C) calculating a value for a compensating resistor for the amplifier based on the determined values; and d) applying a suitable resistance to the amplifier.
11. A method of calculating a compensation resistance for an amplifier comprises: a) determining the forward gain of an amplifier; b) determining the impedance of input circuit ground conductors of the amplifier; C) calculating a value for a compensating resistor for the amplifier based on the determined values.
12. An amplifier substantially as described herein with reference to Figures 4 to 11. * S
13. A method of compensating for ground currents in an amplifier substantially as described herein with *::::2o reference to Figures 4 to 11.

S *SS

S
14. A method of calculating a compensation resistance for : .. an amplifier substantially as described herein with S.....

reference to Figures 4 to 11.