GB2174572A - Radio telephone systems - Google Patents

Abstract

In order to increase the density of use of channels in communications apparatus, especially cordless telephones, means (B,C) are provided for automatically adjusting the transmission power of each transceiver of the cordless telephone link in response to the strength and/or reliability of a signal received by at least one of the transceivers in the cordless link. Before establishing communication via a selected channel certain checks are made including - measuring the radiation strength on that channel, transmitting a signal of a predetermined duration if that strength is below a predetermined level, again measuring the radiation strength on that channel and occupying the channel if the radiation strength has not risen above a predetermined level (Box A). If the radiation strength does not rise above that level, then this is because existing users of the channel are "raising their voices" so as to be heard over the intruder. The intruder then knows not to interfere, but to seek a different channel. <IMAGE>

Description

SPECIFICATION Communications apparatus and method This invention relates to a communications apparatus and method, particularly to "cordless telephone sets " and more particularly to a method and apparatus for enabling many such telephone sets to operate on a particular channel within a given area or volume.

Generally, a cordless telephone set is a telephone terminal connected to an exchange line or an extension line and is integrated with or accompanied by a normal telephone set. It consists of two parts which are connected by a radio link, namely a fixed device (the base unit) and a portable device (the "handset") permitting the same basic functions as a normal telephone within a limited area around the fixed part.

A cordless telephone set may be either: a) a fully functional apparatus on which calls can be set up and answered by the portable device, or b) a limited apparatus where the portable device has no facility for initiating outgoing calls.

Prior art types of cordless telephone sets normally have their base units connected to an exchange line, connecting to the Public Switched Telephone Network (PSTN). However, the present invention is not limited to such types, and includes portable and/or mobile base units, which can communicate by radio with a radio network, such as a cellular system, and if desired have a facility for connection to the PSTN.

With the predicted growth of the use of cordless telephone sets and cellular radio, particularly in commercial application, there is a need to maximise the number of communication links that can be made from a given area or volume using a given number of channels, i.e. to increase the user density. Several good systems exist at present giving reasonable density and good flexibility in the range within which a handset can operate from a base unit. However, as more such telephone sets are used and especially as it becomes desirable to communicate data other than speech by cordless links, so will these systems find their limitations.

It is therefore desirable to increase the use of each channel separately without necessarily reducing the range of operation of handsets.

It is envisaged that by increasing user density, it will be possible for internal or private switchboards, e.g.

in offices and factories, which require vast lengths of wiring to be installed to each individual telephone instrument from the switchboard, to be replaced by a communications system according to the present invention, wherein each user has a portable handset, the switchboard is replaced by a base unit connected to the PSTN, and further base units (which may be portable) - or sub-base units - are optionally provided in various other locations in the building, depending on its size. Such a communications system is the subject of our co-pending U.K. Patent Applications Nos: 8428159; 8424308 and 8500452.

According to one aspect of the present invention, we provide a communications apparatus which comprises first and second communications devices each comprising a radio transmitter and a radio receiver, wherein each device is provided with means for automatically adjusting the transmission power of its respective transmitter in response to the strength and/or reliability of a signal received by at least one of the receivers.

Preferably the apparatus is a cordless telephone with the first device being a fixed base unit, optionally connectable to the PSTN, and the second device is a portable handset. The fixed base unit and the portable handset for use in the communications apparatus also form individual aspects of the invention.

Preferably the transmission power of the transmitter of any one device is adjusted in response to the strength of signals received by the receiver of the other device, for example in response to an encoded command transmitted by the other device, which command is dependent on the strength of signals received by the said other device. The transmitter of each device may transmit a first encoded command when the signal strength received by its receiver is less than a predetermined minimum and transmit a second encoded command when that signal strength is greater than a predetermined maximum.On receipt of the first command, a device which is transmitting at less than maximum power will increase its transmission power incrementally, and on receipt of the second command, a device which is transmitting at more than a predetermined minimum power, decrease its transmission power incrementally.

Preferably the apparatus further includes means for measuring the reliability of a signal received by each device, the measure of reliability being the bit error rate.

Preferably the transmitter of each device transmits a third encoded command when the bit error rate of the signal received by it rises above a first predetermined level and wherein the transmission power of a device is increased on receipt of the third encoded command by its receiver. More preferably the transmission power is increased to maximum on receipt of the third encoded command. Means may be provided within each device for increasing its transmission power to maximum when the bit error rate of the signal received rises above a second higher predetermined level.

Also according to the invention a method is provided of occupying a channel in an interactive communication device comprising the steps of measuring the radiation strength on that channel, transmitting a signal of a predetermined duration if that strength is below a predetermined level, again measuring the radiation strength on that channel and occupying the channel if the radiation strength has not risen above a predetermined level.

The present invention will now be described by way of example with reference to the accompanying drawings, in which: Figure lisa diagrammatic representation of the transmission ranges of two adjacent base units and two associated handsets of a cordless telephone system operating in a known prior art manner; Figure 2 is a flow diasgram for use in explaining the present invention;; Figure 3 is a diagrammatic representation of various base and portable units for explanation of the invention, Figure 4 is a schematic field strength diagram of the units of Figure 3 in operation, Figure 5 is a schematic field strength diagram of units arranged in a straight line, Figure 6 is a diagrammatic representation of a network of cordless telephones according to the invention, and Figure 7 is a block circuit diagram of an implementation of a unit for use in the invention.

Referring to Figure 1, two fixed base units A and B are shown having normal transmission range R.

Associated with each unit is a handset HA and H8 each having a normal transmission range R. In theory, for there to be no serious interference between the links established between HA and A and HB and B, when communicating on the same channel, the base units must be distanced 3R apart, HA must not stray outside the circle centred on A, radius Rand H5 must not stray outside the circle centred on B, radius R. In fact, in practice it is customary to divide the transmission areas into regular hexagons, as shown, with the base units at the centres thereof. It is then generally permissible for, say, handset HA to move anywhere within the hexagon centred on A and still transmit thereto and receive therefrom even though, at times, the distance between HA and A may be slightly greater than R.At such times, transmission and reception occurs at weaker strength. However, an arrangement such as described does severely limit the number of handsets operating on the same channel in any one or in adjacent hexagonal cells. It is normal practice for only one handset to be operating on any given channel in a cell at one period of time.

In the preferred embodiment of the cordless telephone set of the invention now to be described with reference to Figures 2 to 7, both the base units and the handsets comprise a transmitter and a receiver and communicate on a digital modulation format on a single radio channel in time division multiplex mode.

Typically there are 40 channels available for use by the base units and handsets. All information including speech is in digital form. Each part of the cordless telephone has a maximum effective radiated transmitter power of 10 mW.

The present invention enables a greater stacking of base units and handsets to be accomplished in any given area at volume by means of a new procedure for allocating channels which involves monitoring of received power levels, adjustment of transmission power levels, and monitoring of bit error rates. The procedure enables a radio frequency channel to be taken up and occupied on the initiative of a base unit or a handset. The procedure is the same for both. In the case of an incoming call, it will be the base unit and in the case of an outgoing call, the handset, that initiates the channel seizure. Reference hereinafter will just be made to the "initiating device" forthe unit that initiates channel seizure and the other unit as the "counterpart device". The procedure will be described with reference to the flow diagram of Figure 2.

Procedure for the channel seizure The initiating device searches for a potentially idle channel by scanning through the channel frequencies.

A channel is considered to be potentially idle if the initiating device senses that the radio frequency field strength, Re, on the specific channel is below a specified limit, T. The specified limit is preferably 44dB relative to 1 micro volt per metre.

The initiating device then transmits a test signal at maximum power for a short duration and, after a slight delay, then repeats the field strength test. The signal transmitted may, for example, be 101010..., lasting for 2;3 ms. If a change in field strength is detected which raises Re above T, the channel must not be considered idle and another potentially idle channel must be found by recommencing the sequence.

If no change in the field strength above the threshold is detected, the channel is considered idle insofar as the initiating device is concerned. This routine is shown in box A of Figure 2.

The reason for a possible field strength change is as follows.

Once an idle channel is found, a link to the counterpart device is established by methods well known in the art. The initiating device then measures the signal strength on that channel and if it is unnecessarily high to maintain satisfactory communication, it instructs the counterpart device to reduce its transmission power.

The choice of maximum necessary signal strength and the factors governing that choice are discussed later in this description. Thus, if the signal is stronger than a maximum limit, Relax the counterpart device is instructed to reduce its power. Similarly if the signal received by the initiating device falls below a preset minimum, Remin, it instructs the counterpart device to increase power. Likewise, the counterpart device measures signal strength from the initiating device and instructs it to reduce or increase its power accordingly. Preferably, the transmission power of each device can be varied in four steps between maximum power and the operating point and the device awaits instructions from its counterpart before reducing or increasing to a second, third, etc., level. By this means the two devices do not radiate more power and therefore occupy more "space" than is necessary. Should the handset move closer or further away, orthe background noise level change, then each device will on its own account adjust its power in steps as instructed by its counterpart.

This monitoring and adjustment of power levels is illustrated in box B of Figure 2.

As well as this signal strength test, each device performs another test, which is the monitoring of bit error rate. The bit stream received has a parity check e.g. every 32 bits. This is checked continuously for correct parity and should the error rate rise above 1 bit in 1000, then a command is transmitted to the counterpart device to increase power to maximum. In this description, a bit error rate greater than 1 in 1000 will be referred to as "corruption". An increase in power to maximum is used not just to restore reliability but also to serve as a strong warning to other devices to "keep out". When a 101010.... test signal arrives at a device already in operation, and results in corruption, that device instructs its counterpart to increase to maximum power.At this time the initiating device is again measuring the field strength so, if the field strength rises above the threshold (e.g. 44dB relative to 1 V/m) as a result of this power increase, then the channel is determined not to be idle.

Thus it can be seen that where a channel appears at first to be potentially idle, this test routine can determine if it is in fact idle. The bit error rate check routine is shown in Box C in Figure 2.

Although the channel has now been determined to be idle insofar as the initiating device is concerned, this is not yet the case for the counterpart device. The latter therefore then repeats the whole idle channel seizure test routine of Figure 2 again to establish that the channel is validly idle for itself. Once this has been accomplished, normal communications can proceed on the channel.

The procedure is now described in further detail with reference to Figure 3 in which a base unit Y and two other devices U and V, one of which is a base unit, are shown. U and V are, for simplicity of explanation, shown equidistant from base unit Y. Further devices X, Wand Z are also shown. Their distance from Y is for the present arbitrary.

Suppose that U and V are communicating satisfactorily and are employing mutual power control whereby each unit measures the signal strength received from the opposite unit and commands the opposite unit to reduce transmit power until the signal received is below the preset maximum, Relax, or to increase power if the signal received falls below the preset minimum, Remin. This is analogous to telling someone "there's no need to shout" or "please speak up".

If, while this is going on, another signal arrives at, say, V, of sufficient strength to corrupt the signal from U, the V commands U to increase power up to full power to restore the signal and serve as a warning to whichever unit is causing corruption. After a period of time at full power, unit U once again reduces its power until the signal at V is once again within the preset limits or, if the corrupting signal is still present, then until the signal is insufficient to prevent corruption in which case U is commanded to return to high power. In such a situation of prolonged interference from an outside signal, the power control will continuously decrease in steps and increase cyclically. This is tolerable.

Consider now the case where base unit Y wishes to establish a link with unit W. Y first performs a field strength test and finds that the field strength on the channel used by Wand V is below a predetermined threshold T. This is the case because the nearest operating units U and V are operating sufficiently below maximum power. Base unit Y thus determines that the channel is, at first sight, potentially idle, so it transmits, at full power, its test signal. Should the test signal corrupt the signal received by U, then U will command V to increase to full power, whereupon the field strength at Y will increase above threshold T. Y, upon repeating the field strength test, will find that the channel is not in fact idle and will act accordingly by trying a different channel.The threshold T, maximum power P and separation of base units are such that when one unit is operating at full power, then the field strength at a point within a distance equal to the separation of the base units will, other factors aside, be above the threshold. Should, on the other hand, the test signal not corrupt the signal received by U or V, then there will be no responding power increase and the channel will be free for use by Y.

Once Y, as the initiating device, has located an idle channel for use with W, the latter then repeats the idle channel routine to ensure that it can communicate with Y without interfering in communications between X and Z.

It can be seen therefore that there are occasions when a channel which would, but for the power control routine, be determined as occupied, will in fact be available for use. It is clear therefore that more units can operate in this manner within a given area and that the separation of base units can be less than would otherwise be necessary.

The arrangement will be beneficial in a given situation provided that the following equation is satisfied: T-3 + ESt + MC < Re where T is the free channel threshold 3 is a power difference at which the received signal lies below the threshold T.

E St is the amount by which the transmitted power has been reduced. If the power is reduced in steps St, then s, St is the total for the number of steps by which it has been reduced.

Mc is a margin for co-channel corruption and Re is the received signal strength at the operating point i.e. the signal strength received. It lies between the present limits Remax and Rye minx These terms are further explained by Figure 3 which is a power diagram of the W, U, V, Y unit situation of Figure 3 with distance along the base. For simplicity, power diminution has been represented by straight lines rather than more accurate inverse square law relations but this should not detract from the explanation.

Units U and V, though equidistant from Y, have their horizontal separation in a different dimension represented by a discontinuous extension to the base line. Lines 5 and 6 represent the strength of signals from Y and U respectively. The reference power may correspond to a field strength of 1 AV/m.

Re is a design parameter that is required to be kept as low as possible. The lower the operating point, the more the power can be reduced. Re also determines how close U (or V) can operate to Y without being corrupted by the full power test at Y. Thus Figure 3 shows in fact the minimum separation of Y and U for the chosen Re with U and Vtransmitting with signal strengths represented by dotted lines 7. Y and W can then (assuming no interference from X and Z) communicate on similar power levels as shown on the same channel. Nevertheless, it can be seen that this is a closer separation of Y and U than would be allowed if U were transmitting at full power, in which case T would be exceeded.

Figure 5 represents another case to be considered, where Y, U and V are on a straight line. In this case, consider V to be a fixed base unit and the separation of Y and V to be just less than the minimum separation of a base unit and a handset for separate calls to be set up at each one according to the prior art, i.e. if V were to transmit at maximum power, then Twould be exceeded at position Y (dotted line 8). U and V are communicating as before and when Y transmits at full power, the difference at U between the signals from V and Y is just about Mc, the margin for co-channel corruption. Thus if U was closer to Y, Y could not enterthe channel because the signal at U would be corrupted and U would command V to increase to maximum power whereupon T would be exceeded at Y.However, U and V can operate anywhere within the range 9 and Y will not corrupt. In this case, unit W can operate with Y on the same channel, provided that W is sufficiently far from U that it does not corrupt the signal at U when transmitting a test signal at full power.

The response test signal is further described below.

Clearly there are now two handsets U and Wand two base units Y and V operating on the same channel within a range where there was previously only room for one base unit and one handset while at the same time, the maximum range of a handset has not been reduced.

There are a number of parameters here which can be adjusted accordingly to the required specification.

No mention has so far been made of the maximum transmission power. This determines the absolute maximum range of a handset from a base unit but should be kept low to increase the density of users. There is a trade-off between user density and range and this power is preferably of an order already established in the trade. Threshold T is similarly a compromise between density and reliability while Mc preferably corresponds to a bit error rate of 1 in 1000 which is found to be an acceptable error rate. Thus the remaining parameters to be selected are Re and the separation of the base unit. As explained above lower values of Re limit the minimum separation of an initiating device (Y) and a unit in operation (U) for a new linkto be established and therefore limit the maximum density while higher values also limit the maximum density by the likelihood of T being exceeded.The chosen range of Re lies somewhere between these effects. Thus again there is a trade-off which depends on the application and factors such as the median range of operation of a handset from its base unit, rate of set up and duration of calls, minimum acceptable density at peak use and other such factors must be considered.

This description has referred to a base unit Y as the initiating device but it could equally be a handset that initiates a call.

Preferably the power reduction in response to a command is made in a series of discrete steps, making one power reduction step and awaiting a command to make another. If the communicating units are at their minimum step, then the signal strength may rise considerably above Relax when the units move very close together. In this situation clearly there is more signal margin than usual and an interfering unit would have to come even closer before it corrupts.

A numerical example is now given with reference to Figure 3 and by way of illustration only. The figures given in no way limit the scope of the invention. All field strength values are relative to 1 microvoit per metre.

Let P represent the maximum transmission power and let Remin = 65dB, Relax = 70dB, Mc = 1 OdB, T= 44dB, and St = 15dB.

UnitY Unit U UnitV Initial P-1 5dB(transmit) 66dB(receive) condition 40dB(receive) 66dB(receive) P-15dB(transmit) Since Y receives at 40dB from U when the latter transmits at P-1 5dB, then by reciprocity, when Y transmits at P, then U receives at 40 + i 75 dB = 55dub. Thus: Unit Y full PowerTest P(transmit) 55dB(receive) Margin at U 11 dB From this example, the margin at U (11dB) just exceeds Mc (10dB) so there is no corruption. The link between U and V is "robust" and Unit Y can use the same channel.

Consider another case where the three units are closer together.

Unit Y Unit U Unit V Initial P-30dB 66dB (transmit) (receive) condition 40dB(receive) 66dB P-30dB (receive) (transmit) Thus: UnitYfull power test P(transmit) 70dB(receive) Margin at U -4dB In this example, the signal received by U from Y is 4dB above the signal received from V. Since there is now no margin, it causes U to instruct V to increase to full power, and vice-versa (as both U and V are equidistant from Y). U and V increase transmission powers by 30dB, thus increasing the received power at Y to 40+30dB = 70dB. This exceeds the 44dB threshold at Y, causing the latter to determine that the channel is not in fact idle. Simultaneously there is now a satisfactory positive margin of 26dB at U and V.

The preferred embodiment is now described in further detail with reference to Figure 6, in which thirteen base units numbered 10 to 22 are shown. Each of the units 10 to 16 stands at the centre of and serves a generally hexagonal cell 23' to 29', as described above with reference to Figure 1 (the prior art situation ).

Suppose first that a handset is being operated on one channel at point 35 (channel M) in conjunction with a base unit 10. The base unit 10 measures the level of signal from the handset 35 and reduces its transmission power to the minimum aceptable level such that its "range" is shown by the circle 36. Simultaneously, the base unit 10 instructs the handset 35 to reduce its power until defined by the circle 37. These units are now taking up the minimum area necessary for communication.

Suppose now that a user wishes to open a channel with a handset at point 38. Handset 38 searches for an idle channel and determines that channel M is free because the field strength is below the threshold (44dB relative to 1 micro-volt/metre). It then transmits the test signal at maximum power for 2-3ms and checks the field strength again. If handset 38 is too close to the nearest devoce operating on that channel, in this case base unit 10, then base unit 10 senses a corruption of its signal by the test signal of handset 38 and instructs handset 35 to increase its power to maximum. When the power output of the handset 35 is thus increased, the signal strength at handset 38 rises and the handset 38 detects this rise.Should the field strength rise above the predetermined threshold level of 44dB relative to 1 microvolt per metre, then the handset 38 determines that the channel is being used and seeks a different idle channel.

If, on the other hand, handset 38 is somewhat further away from base unit 10, then the test signal is not strong enough to corrupt the signal being received by base unit 10 so handset 35 does not increase its transmit power and handset 38 determines that the channel is free to be used.

From this, it is clear that two or more handsets can satisfactorily operate within a single cell 23 and that further base units can therefore be added to serve these handsets. Six further base units 17 to 22 are shown at the corners of cell 23 and the most efficient spatial arrangement is the hexagonal one as shown. Further base units (not shown) could also be provided at locations corresponding to the mid-point of any three adjacent base units. This means that for a given area (two dimensions) or volume (three dimensions), a considerable increase in base unit density can be achieved over the prior art arrangement of Figure 1, thus allowing handset density to be increased accordingly. A very substantial user density increase can be achieved as a result of the facility of adjusting transmission power provided with each handset and base unit.

It can be seen from this explanation that where an initiating part transmits a test word, it is not necessarily the nearest unit that responds by increasing power. With the mutual power control described, a nearer unit may command a more remote unit to increase power.

Alternative arrangements are possible. Self power control may be employed whereby each unit controls its own power in response to the received signal strength and in response to corruption of that signal. This is less efficient in that it will result in excess transmission power but it requires less circuitry to implement as it dispenses with the need for power commands and gives a stronger warning to initiating units to "keep out".

Similarly a base unit could command the handset to alter power while employing self control on itself. This would impose the circuitry burden on the base unit. These are less preferred arrangements.

Because with mutual control, it may be the more remote unit that responds, it may be that this response does not raise the field strength at the initiating part above the threshold. This will depend on the selection of T and the size of the cells but does give an opportunity for greater densities than with self-control. On the other hand, a leapfrog effect can be envisaged where the response of a more remote unit corrupts a further unit. It is expected that this would only be a problem in the rare situation where the whole network is substantially saturated. A preferred further feature provides for self-control in the event of serious corruption of the signal. Thus if the bit error rate rises above 1 in 100, a unit can of its own accord increase to full power.

In such situations, this may be necessary where an incoming instruction to increase power would be garbled. This is in fact shown in box C in Figure 2.

Referring now to Figure 7, an implementation of power control in a handset or base unit is shown. An input stage 40 passes a signal to an intermediate frequency amplifier 41 capable of measuring the received signal strength to provide a Receive Signal Strength Indicator output (RSSI) 42, as well as an amplified output on line 43. A controller 44 which comprises logic circuitry programmed to accomplish the test routine of Figure 2 receives signals on lines 42 and 43. The controller passes two signals, "output amplifier control" 45 and "command to opposite unit" 46, to an output amplifier47 which is fed by a voltage controlled oscillator 48 to give an output to an antenna 49.

The controller performs two functions. It measures the incoming signal strength at 42 and outputs a command 46 for transmission if necessary and it decodes incoming commands 43 and alters the output transmission power at 45 when necessary.

Controller 44 typically has four power levels at which it can control the amplifier 45 to give signals at the antenna of, say, -40dB, -30dB, -20dB and -10dB relative to maximum transmission output level. Likewise controller 44 has preset signal levels T, Relax and Remjn with which it compares RSSI 42. Relax and Remin may be, for example 65dB and 60dB respectively above I jlV/m. Should RSSI 42 fall below Remin then controller 44 will output a command 46 to be transmitted to the opposite part commanding itto increase power. Likewise if RSSI 42 is above Relax then a reduce power command is transmitted.

Claims (16)

1. A communications apparatus which comprises first and second communications devices each comprising a radio transmitter and a radio receiver, wherein each device is provided with means for automatically adjusting the transmission power of its respective transmitter in response to the strength and/or reliability of a signal received by at least one of the receivers.

2. Apparatus according to claim 1 which comprises means for adjusting the transmission power of each transmitter in response to an encoded command transmitted by the other device, which command is dependent on the strength of signals received by the said other device.

3. Apparatus according to claim 2 wherein the transmitter of each device comprises means for transmitting a first encoded command when the signal strength received by its receiver is less than a predetermined minimum and for transmitting a second encoded command when that signal strength is greater than a predetermined maximum.

4. Apparatus according to claim 3 wherein, on receipt of the first command, a device which is transmitting at less than maximum power comprises means which increases its transmission power in incremental steps and, on receipt of the second command, a device which is transmitting at more than a predetermined minimum power, comprises means which decreases its transmission power in incremental steps.

5. Apparatus according to claim 1 which comprises means for adjusting the transmission power of each respective transmitter in response to the strength of signals received by each respective device.

6. Apparatus according to any one of the previous claims further including means for measuring the reliability of a signal received by each device.

7. Apparatus according to claim 6 comprising means for measuring the bit error rate of the received signal.

8. Apparatus according to claim 7 wherein the transmitter of each device comprises means for transmitting a third encoded command when the bit error rate of the signal received by that device rises above a first predetermined level and wherein means are provided for increasing the transmission power of a device on receipt of the third encoded command by the recieverofthatdevice.

9. Apparatus according to claim 8 which comprises means for increasing the transmission power to maximum on receipt ofthe third encoded command.

10. Apparatus according to any one of claims 6 to 9 wherein means are provided within each device for increasing its own transmission power to maximum when the bit error rate of the signal received thereby rises above a second higher predetermined level.

11. Apparatus according to any one of the previous claims in the form of a cordless telephone wherein the first device is a fixed base unit and the second device is a portable handset.

12. A fixed base unit for use as a communications device in a communication apparatus as claimed in any one of claims 1 to 11.

13. A fixed base unit according to claim 12 which is adapted to be connected to the Public Switched Telephone Network.

14. A portable handset for use as a communications device in a communication apparatus as claimed in any one of claims 1 toll.

15. A method of occupying a channel in an interactive communication device comprising the steps of measuring the radiation strength on that channel, transmitting a signal of a predetermined duration ifthat strength is below a predetermined level, again measuring the radiation strength on that channel and occupying the channel if the radiation strength has not risen above a predetermined level.

16. Apparatus substantially as hereinbefore described with reference to Figures 2 to 7 of the accompanying drawings.