WO2003001757A1

WO2003001757A1 - Pulse-shaping method for reducing the radio frequency emissions

Info

Publication number: WO2003001757A1
Application number: PCT/GB2001/002740
Authority: WO
Inventors: Rolf Peter Startin
Original assignee: Aim Aviation (Jecco) Limited
Priority date: 2001-06-20
Filing date: 2001-06-20
Publication date: 2003-01-03

Abstract

A data transmitter for data communication of a data signal along a wired data communication path comprises a sample value generator operable to provide sample values representing a transition waveform; a digital to analogue converter operable to receive sample values from the sample value generator and to produce a corresponding analogue transmission waveform; the data transmitter being arranged to generate an output waveform corresponding to the data signal in which transitions in the waveform are formed by the analogue transition waveform.

Description

PULSE-SHAPING METHOD FOR REDUCING THE RADIO FREQUENCY EMISSIONS

This invention relates to the field of data communications. More specifically it relates to data communications in wired systems.

Digital data is commonly transmitted in the form of so-called "square waves". Figure 1 of the accompanying drawings schematically shows a square waveform 10 which oscillates between a lower signal level 12 and an upper signal level 14 as a function of time. Positive transitions 20 take the waveform from the lower level 12 to the upper level 14 and negative transitions 22 take the waveform from the upper level 14 to the lower level 12. In the transmission of binary data, normally the upper level 14 represents logical one, T, and the lower level 12 represents logical zero, '0', but of course the polarity convention could be the other way round. The square waveform shown in Figure 1 therefore represents the binary digits (bits) '10110010'. This form of representation is called the 'non-return to zero' (NRZ) format. An alternative representation is the 'return to zero' (RZ) format. In the RZ format a logical one is represented by a positive level, a logical zero is represented by a negative level and there is a null ('zero') region 26 between bits. Figure 2 of the accompanying drawings is a schematic illustration of a square waveform 8 in which the bits '10110010' are represented in RZ format.

A square wave may be resolved into a large number of sinusoidal components. The Fourier transform of a perfect square wave, i.e. a square wave which has instantaneous transitions 20 and 22 between horizontal levels, is an infinite series of sinusoidal components.

When a square wave is transmitted in wired electrical communication systems the upper and lower levels 14 and 12 are represented by voltage or current levels. The harmonics that make up a square electrical waveform have a great deal of radio frequency energy and they can induce corresponding electrical signals in conductors forming part of other nearby equipment. This type of interference is commonly termed "radio frequency interference" (RFI). RFI can cause the malfunction of other equipment that is sensitive to this form of interference, so to reduce RFI the transmission wire carrying the data communication signal, or the other equipment itself, is conventionally surrounded by a metallic screen.

In real situations the transitions 20 and 22 between the voltage levels of an electrical square wave are not absolutely instantaneous. The gradient of the transitions 20 and 22, i.e. the rate of change of the amplitude, is termed the slew-rate. Because the slew- rate is not infinite the number of harmonics that comprise such a square wave is also not infinite. The slew-rate can be deliberately decreased to limit the number of harmonics and therefore the amount of RFI. However, the slew-rate can only be decreased to a value that does not cause the positive and negative transitions 20 and 22 to overlap. Therefore, there is a trade off between the data bandwidth of the signal and the degree to which the slew- rate can be reduced. Figure 3 of the accompanying drawings schematically shows a modified square wave 30 in which the positive transitions 32 and negative transitions 34 have a finite and substantially constant slew-rate, i.e. the slew-rate has been limited. When data is transmitted using a slew-rate limited square waveform the amount of

RFI generated at particular frequencies may be less than that generated when non slew-rate limited square waveforms are transmitted. However, wires carrying these modified waveforms will still require screening in electromagnetically sensitive areas.

In aeroplane electronic systems there is a severe restriction on RFI emissions above a certain magnitude at frequencies greater than 150 kHz. The extent of allowable RFI is governed by an international standard such as one of the so-called "RTCA DO 160" limit lines, an example of which is shown schematically in Figure 4 of the accompanying drawings. This is one of many civil and military electromagnetic emission specifications, but is a very good and widely adopted civil standard. Any RFI emissions above this limit line may interfere with the electronic communications systems on the aeroplane. The limit line includes three frequency bands 2, 4, 6 at which the aeroplane electronics are considered particularly susceptible to RFI, and so the permissible RFI level is lower at these bands. In conventional systems shielding or screening of the wiring that transports the electronic signals contains these emissions. In the case of a large jet aeroplane this shielding may add several hundred kilograms to the weight of the aeroplane. This heavy shielding reduces the useful payload of the aeroplane, as well as increasing manufacturing costs and fuel consumption.

The present invention provides a data transmitter for data communication of a data signal along a wired data communication path, the transmitter comprising: a sample value generator operable to provide sample values representing a transition waveform; and a digital to analogue converter operable to receive sample values from the sample value generator and to produce a corresponding analogue transition waveform; the data transmitter being arranged to generate an output waveform corresponding to the data signal in which signal level transitions in the output waveform are formed by the analogue transition waveform.

The invention recognises that the transmission of a waveform that has been modified by replacing transitions in the waveform with numerically synthesised transitions has the advantage that a reduced amount of RFI may be generated when the waveform is transmitted along a wire.

The synthesised transitions may substantially follow any pre-determined function: examples include, but are not limited to, exponential, linear and sinusoidal functions. Preferably the synthesised transitions substantially follow a sine wave, or more generally one or more sine waves. This has the advantage that the RFI generated when the modified waveform is transmitted occurs at well defined and predictable frequencies. Namely, RFI occurs at frequencies that correspond to the harmonics of the sine wave that was used to synthesise the synthesised transitions. The modified waveform can be tailored so that the harmonics are in a frequency window that is remote from sensitive frequency windows such as those used for data communication. For example, it would be possible to avoid the frequency windows 2, 4, and 6 schematically illustrated in Figure 4 which relate to the frequencies used for data communications on board an aeroplane.

Although the synthesised transitions may substantially follow a sine wave that spans any range of values it is preferable that the synthesised transitions substantially follow a sine wave which ranges over substantially 180°. By using appropriate sine wave ranges, the gradient of the synthesised transitions at both the start and the end point of each synthesised transition can be substantially equivalent to the gradient of any subsequent DC waveform section or any subsequent synthesised transition. This is advantageous because there are then no sudden changes in the amplitude of the waveform which may otherwise cause RFI when transmitted down a wire.

Since the RFI produced by transmission of such modified waveforms can be both reduced and more predictable in frequency content, the heavy shielding that is conventionally used to enclose the transmission wire can be reduced or dispensed with. In an aeroplane the reduction or removal of the shielding allows for a larger payload to be carried by the aeroplane and for a reduction of fuel consumption.

The sample values representing the transition waveform may be stored in a memory or other storage device (e.g. a read only memory) or may be calculated as required. 03/001757

4

Further respective aspects and features of the invention are defined in the appended claims.

Embodiments of the present inventions will now be described, by way of example only, with reference to the accompanying drawing in which: Figure 1 schematically illustrates a 'non-return to zero' square waveform;

Figure 2 schematically illustrates a 'return to zero' square waveform; Figure 3 schematically illustrates a slew-rate limited waveform; Figure 4 is a schematic graph of intensity in decibels per microvolt against frequency in Hertz; Figure 5 schematically illustrates a data transmitter;

Figure 6 schematically illustrates a data communication system; Figure 7 schematically illustrates a data communication system configured in spur geometry;

Figure 8 schematically illustrates a data communication system configured in a star geometry;

Figure 9 schematically illustrates a data communication system configured in ring geometry;

Figure 10 schematically illustrates a data communication system configured to allow two-way communication; Figure 11 schematically illustrates a modified waveform, in non-return to zero format, in which dwell periods are maintained;

Figure 12 schematically illustrates a modified waveform, in non-return to zero format, which does not have dwell periods;

Figure 13 schematically illustrates a modified waveform, in non-return to zero format, in which transitions do not take place between successive identical data bits;

Figure 14 schematically illustrates a modified waveform, in return to zero format, in which dwell periods are maintained;

Figure 15 schematically illustrates a modified waveform in return to zero format which does not have dwell periods; Figure 16 schematically illustrates a data receiver;

Figure 17 is a plot of 26 sample values; and Figure 18 is a chart illustrating measured RFI against frequency. Figure 5 is a schematic illustration of a data transmitter 40. The transmitter 40 may be the originator of the a data signal to be transmitted, or may simply act as a modulator or driver to transmit a data signal generated in other apparatus (not shown) and supplied to the transmitter by an input port 42. It is well within the ability of the skilled man to implement appropriate control electronics to generate a raw data signal to be transmitted.

In the case when the original waveform is generated externally, a microprocessor or microcontroller 44 detects transitions in the input data signal and activates a counter 46 when a transition is detected. The function of detecting transitions and initiating a counter may alternatively be performed by discrete logic circuitry. In the case when the original waveform is generated internally, the microprocessor or microcontroller 44 may be used to both generate the original waveform and to activate the counter 46.

The counter 46 addresses a look-up table 50 contained within a memory 48. The look-up table 50 contains data that is accessed by the counter 46 to generate synthesised sections of a waveform. The memory 48 may be any type of memory unit suitable for storing numeric data. For example, the memory unit may be a random access memory (RAM), for example a non-volatile RAM such as a flash RAM, a read only memory (ROM) or an erasable programmable read only memory (EPROM). Preferably the memory 48 is a non-volatile memory unit to provide a more secure store for the look-up table 50.

The data stored in the look-up table 50 comprises two components, an index (which may be implied - see below) and a numeric value. After it has been activated the counter 46 counts from a beginning index to an end index. It is not necessary for the values stored in the look-up table 50 to be in the correct order required to form the synthesised waveform section, although a particularly convenient method of operating such a system is for the index to be represented by an address within the memory, with each address storing the corresponding numeric value. When the values are addressed in the order that they are required then the correct waveform is created.

The counter 46 counts at a rate which is determined by a clock 52. The clock 52 may also be used to control other devices within the transmitter 40. For example data values may be recalled from the memory 48 at a rate controlled by the clock 52 or by a different clock (not illustrated). The counter 48 and clock 52 are operated for at least a period of time that allows the correct amount of data to be accessed from the memory unit to replace the transitions of the original waveform. In a particular example embodiment, the transition waveform lasts for substantially 10 microseconds and represents a substantially 180° section of a sine wave, so that the sine wave has a fundamental frequency of about 50 kHz (in fact for reasons of design convenience a fundamental frequency of 50388.8 Hz is used). During the period of 10 microseconds, 26 sample values are output from the memory unit, the sample values being substantially evenly spaced in time.

A digital to analogue converter (DAC) 54 is connected to the memory 48 to receive the numeric values as they are read out. The DAC 54 produces an appropriate voltage or current level according to the digital value of each code read out of the memory 48. The resolution of the DAC 54 is sufficient for converting numeric values as they are stored in the look-up table 50. The DAC 54 may be chosen from a large range of commercially available DACs which have sufficient resolution to cope with the incoming signal. An exemplary DAC 54 would be an eight bit DAC supplied by Texas Instruments under the part number TLC7524. However, the skilled man will appreciate that a simpler solution such as a resistor ladder could of course be used.

The output from the DAC 54 is an analogue electrical waveform which represents the same digital information as the original data signal. A filter 56 is connected to the DAC 54 to receive the analogue waveform. The filter 56 filters out spurious signals that may be contained within the analogue waveform. A driver 58 is connected to the filter 56. The driver 58 provides gain (if needed) to adjust the voltage or current levels of the filtered analogue waveform before the waveform is transmitted, and also provides an appropriate output impedance for driving a communication line. The driver 58 may be, for example, a dual power operational amplifier in a push-pull configuration. The modified waveform is supplied from the transmitter (via an output port 60) to a wired data bus 62.

In Figure 5, the microprocessor 44 could in fact feed the DAC 54 directly, avoiding the need for intervening components, for example in situations where the sample values representing the transition waveform are generated directly (e.g. as needed) rather than accessed from a memory store. A number of situations are envisaged within the scope of the present invention:

(a) as shown in Figure 5, with the digital data to be transmitted being supplied to the microprocessor 44 in order that a "modified" waveform can be generated (b) as shown in Figure 5, but with the microprocessor being the source of the digital data, that is to say, the microprocessor providing the information to be transmitted as well as controlling the generation of an appropriate output waveform. (c) As in either (a) or (b), where a read-only memory 48 or another storage device stores the sample values needed to construct a transition waveform, (d) As in either (a) or (b), where the sample values are generated "as needed", for example by the microprocessor 44. In all of these and other configurations, the skilled man will be aware of many different ways of generating the transition waveform. For example, a series of samples which are substantially evenly spaced in time could be used, or in another example a series of samples which are substantially evenly spaced in amplitude could be used, but where the timing of the conversion of each sample is controlled so as to give an appropriate (e.g. sine-like) output waveform. Similarly, in all of these and other configurations, there may or may not be an

"original" waveform in existence in the system. That is to say, the term "modified waveform" should not be taken to imply that an original waveform has necessarily been altered in some way - there may have been nothing more than a data signal, or in some embodiments such as those in which the microprocessor generates the information to be transmitted, the output waveform may in fact be the first expression of a data signal.

Various example arrangements for data communication will be discussed briefly with reference to Figures 6 to 10.

Figure 6 schematically illustrates a communications system comprising a transmitter 40 connected to a receiver 68 via a data bus 62. Preferably the data bus 62 comprises a twisted-pair communication line. The receiver 68 may be a conventional receiver capable of receiving both the original waveform and the modified waveform, alternatively the receiver 68 may be a receiver that has been specifically designed to receive the modified waveform.

A suitable receiver 68 is schematically illustrated in Figure 16 and will be described below.

A number of receivers 68 may be arranged to receive a modified waveform. The receivers 68 may be connected to one or more transmitter 40 according to a number of different geometries: Figure 7 schematically illustrates a spur geometry; Figure 8 schematically illustrates a star geometry; and Figure 9 schematically illustrates a ring geometry.

Figure 10 schematically illustrates a system in which a transceiver unit 70

(comprising at least a transmitted 40 and a received 68) is able both to transmit and to receive modified waveforms and thus allow two-way data communication. A number of transceivers 70 can also be connected together according to spur, star and ring geometries.

A large number of different transmitter/receiver, transceiver/transceiver, transmitter/transceiver or transceiver/receiver geometries may be easily envisaged by a person skilled in the art of data communications. Various ways in which transitions in a data communications signal may be represented by numerically synthesised waveform sections will be described below with reference to Figures 11 to 15.

The transmitter 40 can be used to produce a modified waveform from an original square wave waveform. An example of such a square wave is the waveform 10 schematically illustrated in Figure 1. The transitions 20 and 22 in the square wave 10 can be replaced by synthesised sections of a waveform that follows a function that is substantially a sine wave, or a waveform which is built from a predetermined combination of two or more sine waves. A waveform sections corresponding to a sine wave spanning from -90° to +90° replaces the positive-going transitions 20 and waveform sections corresponding to a sine wave spanning from +90° to +270° replaces the negative-going transition 22. When such a modified waveform is transmitted then a reduced amount of

RFI is produced compared that which would have been produced if the original waveform was transmitted.

The skilled man will of course be aware that a sine wave is equivalent to a cosine wave that has been phase shifted by 90°. Similarly, the function sin(θ ± n.360°), where θ is an angle in degrees and n is an integer, is taken to be entirely equivalent to the function sin(θ).

Sample values corresponding to a sine wave are stored in the look up table 50 in the memory 48. By noting that a sine wave has a number of symmetries the number of sample values which need to be stored can be reduced, to reduce the size of the look-up table 50.

Table 1 is a simplified example illustrating a look-up table, stored in a 32 byte memory, that has 32 sample values relating to a sine wave. Since the values relating to the sine wave have 3-bits a DAC 54 of at least 3-bit resolution (in this simplified example) would be required to process data from this table.

Table 1

A section of a sine wave spanning 0° to 90° can be simulated by accessing the first (leftmost) column of values in Table 1 in order, i.e. the index is addressed from 00000 to 00111 in sequence. A section of sine wave spanning 90° to 180° can be simulated by accessing the second column of values in sequence, but by noting the symmetry of the sine wave the same section can also be simulated by accessing the first data column in reverse order, i.e. the index is addressed from 00111 to 00000. Accessing the third and fourth data columns in order simulates a section of a sine wave spanning 180° to 360°. That is, the index is addressed from 10000 to 11000 in sequence, although the same section could also be simulated by taking the complement of the data values held in the first two data columns.

As a worked example, the section of a sine wave spanning 180° to 360° can be simulated by accessing only the first data column. First the complement is taken of this data (i.e. the first column of Table 1) accessed in ascending order, then the complement is taken of the same data accessed in reverse order. It can be seen that it is possible to simulate an entire sine wave from a table that only holds values relating to a section of sine function spanning 0° to 90°. Furthermore, all the values necessary to represent a sine wave can be obtained from the values representing any one quadrant of a sine wave no matter what the range of angles the quadrant spans. If this technique is used then the counter 46 needs to be operable in either sense, i.e. counting upwards or downwards, and a controllable inverter (e.g. a set of exclusive-OR gates is needed at the output of the memory 48.)

The spectral quality of the waveform outputted from the DAC 56 is related to a number of factors including amplitude quantisation and DAC linearity. Spurious signals relating to these factors can be filtered out using the electronic filter 56. The filter 56 in this embodiment is a fourth order Bessel filter with a Salen-Key topography and a roll-off of 75 kHz. Its main role is to filter out digital aliasing, clock breakthrough and coupled digital noise in the analogue circuit. As mentioned above, however, the sample values used in the present embodiment form a group of 26 sample points occupying a 10 microsecond transition period. These points are expressed to eight bit resolution, being rounded versions of ((sin(x) + 1) * 128) with x varying from -90° to +90°. The sample points can either be inverted to give a falling transition edge or can be read in the opposite order. An example set of sample points according to this arrangement is given in Table 2.

Table 2

jrees sin(x) (sin(x) +1) * 128 sample value (decimal)

-90 -1 0 0

^■82.8 -0.99211 1.005376 1

•75.6 -0.96858 4.005647 4

^■68.4 -0.92978 8.953498 9

^■61.2 -0.87631 15.7709 16

-54 -0.80902 24.35033 24

^■46.8 -0.72897 34.5565 35

^■39.6 -0.63742 46.22844 46

^■32.4 -0.53583 59.18208 59

^■25.2 -0.42578 73.21314 73

-18 -0.30902 88.10033 88

^■10.8 -0.18738 103.6089 104

-3.6 -0.06279 119.4942 119

3.6 0.062791 135.5058 136

10.8 0.187381 151.3911 151

18 0.309017 166.8997 167

25.2 0.425779 181.7869 182

32.4 0.535827 195.8179 196

39.6 0.637424 208.7716 209

46.8 0.728969 220.4435 220

54 0.809017 230.6497 231

61.2 0.876307 239.2291 239

68.4 0.929776 246.0465 246

75.6 0.968583 250.9944 251

82.8 0.992115 253.9946 254

90 1 255 255

A plot of these 26 points is shown as Figure 17.

Figure 11 schematically illustrates a first example of a modified waveform 76 that may be produced by the transmitter 40 from an original waveform which is the NRZ square waveform 10 schematically illustrated in Figure 1. Synthesised sections corresponding substantially to a sine wave spanning from -90° to +90° replace the positive transitions 20 of the original waveform 10. Synthesised sections corresponding substantially to a sine wave spanning from +90° to +270° replace the negative transitions 22 of the original waveform 10. In this example the waveform 76 has high and low DC regions 72 and 74 since only the transitions of the original waveform are replaced. By varying the fundamental frequency of the sine wave section replacing the transitions and/or the bit rate, it is possible to eliminate some or all of the DC regions 72 and 74. Figure 12 schematically illustrates a second example of a modified waveform 78 that may be produced by the transmitter 40 from the original waveform 10. In this example there is no dwell time between a synthesised section corresponding substantially to a sine wave spanning from -90° to +90° and a synthesised section corresponding substantially to a sine wave spanning from +90° to +270°. In other words the modified waveform does not have any high DC regions. In this way, a logical one, '1 ' is denoted by a synthesised section corresponding substantially to a sine wave spanning 360°, whilst a logical zero, '0', is denoted by a low DC region 74. In this example the modified waveform 78 has a mark- space ratio that significantly less than the mark-space ratio of 1 : 1 exhibited by the original waveform 10. Because of this, some methods of testing a data bit may yield an error. To overcome this problem either the waveform 78 or the receiver 68 can have its threshold(s) appropriately modified. Alternatively, all instances of a ' 1 ' could be advanced in time with respect to the start bit (or synchronous bit) such that the centre of the bit becomes aligned with the receiver's timing requirements. Figure 13 schematically illustrates a third example of a modified waveform 80 that may be produced by the transmitter 40 from the original waveform 10. The waveform 80 is similar to the waveform 78 in that a '1' is represented by a synthesised section that corresponds substantially to a sine wave spanning 360°. In this example, however, a high DC region 86 is maintained between successive Ts. The transmitter 40 can also modify data that has been encoded in the RZ format.

Figure 14 schematically illustrates a first example of a modified waveform 88 that may be produced by the transmitter 40 from an original waveform which is the RZ square waveform 8 schematically illustrated in Figure 2. The waveform 88 is analogous to the waveform 76 in that the DC regions 72 and 74 between the leading and trailing transitions corresponding to a single bit are preserved. Synthesised sections corresponding substantially to a sine wave spanning from -90° to +90° replace the positive transitions 20 of the original waveform 8. Synthesised sections corresponding substantially to a sine wave spanning from +90° to +270° replace the negative transitions 22 of the original waveform 8. Figure 15 schematically illustrates a second example of a modified waveform 90 that may be produced by the transmitter 40 from the original waveform 8. The waveform 90 is analogous to the waveform 78 in that a '1' is represented by a section that corresponds substantially to a 360° sine wave section spanning from -90° to +270°. In the waveform 88 a '0' is also represented by a section that corresponds substantially to a 360° sine wave, but in this case the sine section spans from +90° to +450°.

In both waveform 88 and waveform 90 the null periods 26 are preserved. The original waveforms 10 and 8, and the modified waveforms 76, 78, 80, 88, and

90 are given by way of example only. A person skilled in the art of digital communications could readily envisage other original waveforms that could be processed according to the apparatus described herein to give various modified waveforms.

Figure 16 schematically illustrates a data receiver for use with the transmitter 40 described above.

A data-carrying signal received at an input port 64 is supplied to a clock extractor 70 and a thresholder 69. The clock extractor 70 extracts a clocking signal from the received signal by conventional techniques. This provides clocking time-points (e.g. the centre of each data bit as represented by the received signal), at which time-points the thresholder compares the amplitude of the received signal with one or more thresholds to determine whether a current data bit is a 1 or a 0. The output of the threshold unit, a data signal comprising ones and zeroes, is supplied to a line driver 66 or other appropriate circuitry to forward, decode or be controlled by the received data.

As an example, an aeroplane communication system using a square waveform transmitted at 9600 baud could be easily modified by incorporating a transmitter 40. In this case the transmitter 40 could be set up to replace the transitions of the square waveform with synthesised sections which correspond substantially to a sine wave with a fundamental frequency of substantially 50 kHz. Other baud rates and sine wave frequencies could be used depending on the communication protocol that was in use. It is noted that the fundamental frequency of 50 kHz is very much less than the lower frequency limit of 150 kHz at which the RTCA EMI specifications even begin to define acceptable levels.

Figure 18 is a chart illustrating empirical RFI measurements against frequency for an example transmitter according to an embodiment of the invention, driving a simulated avionic data network with 255 receiving nodes. The lower curve is the measured RFI output, and the upper (straighter) curve is a part of the RTCA/DO-160 category L line. The measured levels are close to the noise floor of the measurement equipment and are far below the RTCA line. A slight peak is seen at a frequency of 50 kHz, the fundamental frequency of the transition waveform.

The transmitter 40 or receiver 68 can be easily integrated with the existing communications systems of an aeroplane, so that retrofitting of the transmitter 40 is straightforward and cost effective.

Within the scope of the present invention the transmitter, receiver and other components of the systems described above may be implemented at least in part by (for example) a general purpose data processing device under the control of computer software, by firmware controlled devices such as field programmable gate arrays, by bespoke data processing devices such as application specific integrated circuits, by hard- wired circuitry, or by combinations of these. In the case of a software or firmware content, the software or firmware may be provided by a transmission medium (e.g. a network or the internet) or a storage medium (e.g. a memory or a disk storage medium).

Claims

1. A data transmitter for data communication of a data signal along a wired data communication path, the transmitter comprising: a sample value generator operable to provide sample values representing a transition waveform; and a digital to analogue converter operable to receive sample values from the sample value generator and to produce a corresponding analogue transition waveform; the data transmitter being arranged to generate an output waveform corresponding to the data signal in which signal level transitions in the output waveform are formed by the analogue transition waveform.

2. A transmitter according to claim 1, in which the sample value generator comprises: a memory storing a set of sample values; and control logic for addressing the memory to output sample values corresponding to a required transition waveform.

3. A transmitter according to claim 1, in which the sample value generator comprises logic for deriving sample values corresponding to a required transition waveform.

4. A transmitter according to any one of claims 1 to 3, comprising an electrical filter operable to filter the analogue waveform generated by the digital to analogue converter.

5. A transmitter according to any of the preceding claims, in which the transition waveform substantially comprises at least a section of a sinusoidal waveform.

6. A transmitter according to claim 5, in which the transition waveform substantially comprises a 180° section of a sinusoidal waveform.

7. A transmitter according to any one of the preceding claims, comprising: a detector for detecting signal level transitions in the data signal;

8. A transmitter according to claim 7, in which the sample value generator is operable to provide the sample values in response to the detection of a signal level transition in the data signal.

9. A data communications system comprising: a data transmitter according to any one of claims 1 to 8; a data receiver; and a wired data communication path linking the data transmitter and the data receiver.

10. A data transmitter substantially as hereinbefore described with reference to the accompanying drawings.

11. A data communications system substantially as hereinbefore described with reference to the accompanying drawings.

12. An aeroplane electronic control system comprising a data communications system according to claim 9 or claim 11.

13. A method of communicating a data signal along a wired data communication path, the method comprising the steps of: providing sample values representing a transition waveform; digital to analogue converting sample values from the sample value generator to produce a corresponding analogue transition waveform; and generating an output waveform corresponding to the data signal in which signal level transitions in the waveform are formed by the analogue transition waveform.

14. A data communications method substantially as hereinbefore described with reference to the accompanying drawings.

15. Computer software for carrying out a method according to claim 13 or claim 14.

16. A software providing medium by which software according to claim 15 is provided.

17. A medium according to claim 16, the medium being a transmission medium.

18. A medium according to claim 16, the medium being a storage medium.