GB1585121A

GB1585121A - Assessment of distortion in binary waveforms

Info

Publication number: GB1585121A
Application number: GB2334376A
Authority: GB
Original assignee: Lind L F; Trend Communications Ltd
Current assignee: Lind L F; Trend Communications Ltd
Priority date: 1976-06-05
Filing date: 1976-06-05
Publication date: 1981-02-25
Also published as: FR2354007A1; FR2354006A1

Description

(54) IMPROVEMENTS IN OR RELATING TO THE ASSESSMENT OF DISTORTION IN BINARY WAVEFORMS (71) We, TREND COMMUNICA TIONS LIMITED, a British Company, of St. Alphage House, 2 Fore Street, London, and LARRY FREDERICK LIND, a citizen of the United States of America, of 1 Almond Close, Wivenhoe, Colchester, Essex, do hereby declare the invention, for which we pray that a patent may be granted to us, and the method by which it is to be performed, to be particularly described in and by the following statement: This invention relates to methods of and apparatus for assessing bias distortion present in binary waveforms and especially though not exclusively - in binary waveforms used for telegraphy and data transmission, when received after transmission.

The invention is concerned with the assessment of bias distortion in both continuous (i.e. isochronous) binary waveforms and in "stop/start" waveforms - such as are used for telegraphy, telex and so on.

When a binary waveform is transmitted over telegraph lines, it is especially liable to two types of distortions; namely, bias distortion and fortuitous distortion. Although other distortions can occur in binary waveforms, they are usually of little importance in telegraphy and data transmission binary waveforms transmitted at relatively low speeds over telegraph lines. Bias distortion is regular in nature, and manifests itself as a constant change at reception in the markspace ratio from the transmission markspace ratio of 1 1. Bias distortion can be present in positive- or negative-going edges arriving later than expected, though sometimes with this distortion an edge arrives earlier than expected.Fortuitous distortion is random in nature and can occur with positive- or negative-going edges which arrive either earlier or later than expected, notwithstanding the presence of bias distortion. Bias distortion is often of greater interest to a user, for if it is present on a received binary waveform, it is possible electronically to correct for it, by suitable adjustment of the reception equipment.

Because fortuitous distortion is random in nature, it usually has to be tolerated provided it does not exceed certain levels, and thus often there is no point in assessing accurately the precise level of fortuitous distortion in a received waveform.

When a binary waveform is received, a third distortion can arise in the receiving equipment itself. This is speed distortion, which results from an incorrect setting of a clock pulse source in the receiving apparatus. This form of distortion can be corrected by an appropriate adjustment of the clock pulse source so that the clock pulse rate in the receiving equipment is synchronised with the clock pulse rate at the transmitter of the waveform.

Equipment has been designed for measuring both bias distortion and fortuitous distortion, which equipment is connected to the line over which the binary waveform is transmitted in order to assess the amount of distortion present in the waveform following its transmission. In general, such equipment is complex and difficult to use, and furthermore it is often the case that an operator has to be well-experienced to gain an accurate overall picture of the amount of bias distortion present, and not to be misled by an exceptional distortion peak.

According to one aspect of this invention, there is provided a method of assessing any bias distortion present in a received binary waveform, comprising: adjusting a clock pulse source to deliver to an element clock a preselected number of clock pulses for each binary element period, the element clock requiring said preselected number of pulses to be driven through each element period; re-setting the element clock on the arrival of one of a positive- or negative-going edge of the waveform; starting an averaging counter to count clock pulses on the arrival of the next following edge of the opposite sense; stopping the counting of clock pulses in the averaging counter when a pre-set number of clock pulses have been counted by the element clock following the point at which said next following edge should have arrived if no distortion is present; repeating the counting sequence by starting and stopping the averaging counter in the aforesaid manner for a chosen number of times; averaging the counts obtained thereby; and then displaying the averaged count to indicate the amount of any bias distortion present in the waveform.

It will be appreciated that in this aspect of the invention, the bias distortion present in the binary waveform is assessed several times and then averaged before a display of the amount thereof is given. This serves greatly to assist an operator in assessing accurately the amount of bias distortion present and thus the information obtained from the method of this invention is likely to have greater credibility and usefulness in adjusting apparatus designed to receive the binary waveform to eliminate the effect of the bias distortion.

When performing the method of this invention it is preferred to repeat the said counting sequence 64 times and then to average the 64 counts obtained thereby, in order to achieve an assessment of the amount of any bias distortion present. The figure of 64 counting sequences is chosen for several reasons: firstly it is sufficiently high to average out the effects of random fortuitous distortion which may be present on the waveform, and secondly it is especially easy to average powers of 2 when operating on digital signals. Whilst a yet better result might be obtained for 128 counts, this high number is not chosen because the usual telegraphy and data transmission rates would lead to a considerable delav before the display of the distortion following the initiation of the assessing method.

It can be shown mathematicallv that the averaging technique of this invention provides a better estimation of the true value of bias distortion than would be obtainable simply bv considering individual binary elements. leaving aside questions of fortuitous distortion. The expected arrival time of a signal edge can be expressed bv a probabil itv density function having a standard deviation o. The standard deviation is a measure of the width of the function, and if 12 identical but independent probabilitv densi tv functions are added together. o,,, = oVo.

fi this is averaged. both the overall sum and 0next are divided by ii: that is. onet/n = o/Vn and this function decreases as ti increases.

Thus when nH. onet/noO. which implies that the mean of n cases is the precise value.

In the method of this invention, it is preferred for the clock pulse source to be adjusted to deliver clock pulses at a rate of 100 for each binary element period, and for the preset number of clock pulses which causes the stopping of counting of clock pulses by the averaging counter to be, precisely 50 pulses following the point at which said next following edge should have arrived as determined by the element clock i.e. the element clock has advanced to a count of 50 after the completion of counting of the immediately preceding period. Thus, the counting of clock pulses is stopped at precisely the mid-point of the binary element period (as determined by the element clock) which commenced when said next following edge should have arrived - but because of bias distortion that edge might arrive early or late.If the edge arrives early, the count by the averaging counter will be greater than 50, and whereas if the edge arrives late, a count of less than 50 will be recorded. Of course, if there is no bias distortion, the edge will arrive precisely when it should, as determined from the clock pulses, and so a count of 50 will be recorded.

In order to display the averaged count with the optimum effect it is preferred for the averaged count to be suitably encoded for driving a display comprising a linear array of lamps, and preferably of lightemitting diodes. An averaged count of 50 (with the preferred clock pulse rate) should illuminate the central lamp of the array, indicating no bias distortion, with counts of more than 50 illuminating lamps to one side of the central lamp and designated mark bias distortion, and with counts of less than 50 illuminating lamps to the other side of the central lamp and designated space bias distortion.

Because the clock pulse source is preferably arranged to deliver 100 pulses per element period. each extra unit of averaged count corresponds to an increase of 1% in the assessed bias distortion: in the preferred display arrangement, it is convenient to provide a separate lamp for each 3C/e change in bias distortion. Conveniently the lamp array consists of 43 separate lamps. such that distortions of up to 40roc each side of zero bias distortion can be shown: this requires 41 lamps and the two end lamps of the linear array can then be used to show when the assessed distortion exceeds 40% either as mark bias distortion or space bias distortion.

It will be appreciated that the element clock may be arranged to require numbers of pulses to drive the clock through one element period other than the 100 mentioned above. For such cases. the clock pulse source must suitably be adjusted to provide the required number per element period in each element period. Also, the precise point in the binary element period at which counting of clock pulses in the averaging counter is stopped may be other than the half-way point of the element clock period, though the half-way point is the optimum for it allows for the measurement of distortions of up to 100%. However, since bias distortion is typically of the order of +6%, the count may be terminated earlier than the half-way point and the decoding arrangement for the display modified accordingly.

It is preferred for the element clock to be reset in the method of this invention at the arrival of a positive-going edge, with the counter being started on arrival of the next following negative-going edge. This is because the "rest" state of a binary waveform is usually following a negative-going edge, and thus at level ZERO - though in telegraphic notation this level is often referred to as level ONE. With stop/start transmissions, the first edge (the "start" edge) of a character is positive-going, and this edge should be used for resetting the element clock.

As mentioned above, in the method of this invention it is necessary initially to adjust the clock pulse source to deliver the preselected, constant number of pulses for each binary element period; that is to say, before assessing the bias distortion, the speed distortion in the assessing method must first be reduced to zero.In a preferred method of this invention, adjustment of the clock pulse source to deliver the preselected number of pulses per binary element period to the element clock is effected by: resetting the element clock at the arrival of a start edge of the binary waveform; detecting the first following edge of the same sense as the start edge; commencing the counting of clock pulses at the detection of the said first following edge; stopping the counting of clock pulses in the averaging counter when the element clock has counted a pre-set number of clock pulses following the completion of the counting of the immediately preceding element clock period; repeating the aforesaid averaging counter counting sequence for a chosen number of times; averaging the counts obtained thereby; and then displaying the averaged count as an indication of the error in the clock rate relative to the period of the binary waveform so that the clock pulse source rate may be adjusted dependent upon the displayed averaged count.

The effect of averaging the obtained count for a number of counting sequences serves to reduce the effect of fortuitous distortion, as well as to reduce the effects of receiving characters of different binary make-up, which will themselves prevent an accurate display of the precise error.

However, the display will show whether the clock pulse source is running too fast or too slow - and hence which way an adjustment should be made - and will show when the speed is exact.

Because bias distortion is regular in nature and each edge of the binary waveform of the same sense is subject to the same space bias distortion or the same mark bias distortion, the interval between any two edges of the same sense must be a multiple of the binary element period, ignoring the effect of any fortuitous distortion. In the above method, the speed of the clock pulse source is adjusted by operating on edges of the same sense, so that effects of bias distortion are eliminated.

The element clock is reset at the arrival of a start edge - that is the first positive-going edge of a stop-start transmission character.

With isochronous transmission, any edge may be regarded as a start edge - either a positive- or a negative-going edge. The first edge of the same sense as the start edge may arrive two binary element periods after the start edge at the soonest, though of course it may not arrive for some number of periods thereafter. Because the counted period error will increase as a function of the number of periods between the start and first detected edges, this method cannot give an accurate absolute indication of the speed error but can nevertheless show whether the clock pulse source is running slow or fast.

When operating onstop/start waveforms, it is advantageous for only one count to be made on each character. To this end, the detection of a start edge should be inhibited after the detection of one such edge for the duration of a number of successive binary element periods, the precise number depending upon the character alphabet chosen. This inhibiting may of course be used with isochronous waveforms, though there is no advantage in doing so.

As mentioned above, the clock pulse source is adjusted to supply said preselected number of clock pulses per binary element period to the element clock, the period of which is equal to one binary element period; that is to say, said preselected number of clock pulses drive the element clock through one period. In this way, an indication may be given by the element clock of the termination of one binary element period and the commencement of the next. In the preferred method of this invention, the bias distortion is assessed by stopping the counting of clock pulses at precisely the mid-point of the binary element period and thus when the element clock has, in effect, counted 50 clock pulses following the commencement of an element period, if there is a count of 100 pulses per element.To achieve this, the element clock preferably gives the required additional output half-way through its period, though a separate counter could be provided for this purpose. Said additional output can also be used to stop the counting of clock pulses when the speed of the clock pulse source is being adjusted, i.e. for an element period of 100 counts when the preset number of counts is 50. If the clock pulse source (and hence the element clock) is running at the correct rate, the detection of the said first edge will occur at precisely the correct moment as determined by the element clock and so the count of the clock pulses will be precisely 50 - always ignoring effects of fortuitous distortion.An averaged count of 50 should then be displayed as zero speed distortion, with counts of less than 50 indicating that the clock pulse rate is too quick and counts of more than 50 indicating that the clock pulse rate is too slow.

Preferably the same display arrangement is employed as has been described above for the display of bias distortions, the operator using a manual control for adjusting the rate of the clock pulse source, so as to illuminate the centre lamp of the linear display. For adjustment of the speed, the two lamps at the extremes of the display can be used to show that the speed error is so large that it is out of range of the other lamps, and the operator would have to adjust the control until both end lamps are extinguished; then one intermediate lamp will be illuminated and the operator can see which way further to adjust the control to bring the speed of the clock pulse source to the required value, as shown by the illumination of the central lamp.

It is preferred for the counting for clock speed to be effected 8 times and then averaged for display. It is found that the clock pulse source can be set with sufficient accuracy with a relatively low number of counts duly averaged, and again a power of two is chosen for it is easy to divide by such a number when working with digital signals.

If many more than 8 counts are used, the time required is rather long for low input speeds.

It is found that apparatus arranged to perform the method of assessing the bias distortion present in a binary waveform, as described above according to this invention, can be re-arranged relatively easily so as to allow the assessment of any total distortion present in the same waveform, including fortuitous distortion. Methods of assessing total distortion - comprising fortuitous distortion and bias distortion - are described and claimed in our co-pending Application No. 23574/77 (Serial No. 1584592).In its broadest aspect, such a method comprises: setting a clock pulse source as claimed in co-pending Application No. 23574/77 (Serial No. 1584592) to deliver to an element clock a preselected number of pulses per binary element period; re-setting the element clock on the arrival of an edge of the binary waveform; counting with the element clock delivered clock pulses; and at the arrival of each successive edge following the re-setting edge transferring the instantaneous count in the element clock to a display device, the early or late arrival of each successive edge relative to its expected arrival as determined by the element clock being displayed.

In this method of assessing total distortion no averaging takes place. Instead, an instantaneous display is provided of any total distortion (i.e. bias and fortuitous distortion) present in the arrival time of each edge, relative to its expected arrival time, from which any fortuitous distortion may be gauged. Preferably the display devices comprise a linear array of lamps, a particular instantaneous count illuminating a particular lamp of the array. Thus, total distortion will be shown as lamps flickering depending upon the amount of distortion present. If only bias distortion is present, a single lamp will remain lit, spaced from the sole lamp lit if no total distortion is present.In view of the random nature of fortuitous distortion, the lamps which are illuminated will also move around the display at random, and it is left to an operator to gauge the precise amount of fortuitous distortion by observing the display.

The method described above of adjusting the speed of a clock pulse source relative to a received binary waveform may of course also be employed in equipment in which a clock pulse source must run at a preselected number of clock pulses per binary element period, quite apart from the method of assessing bias and total distortion as described above. Such a method is also described and claimed in our co-pending A- plication No. 23574/77 (Serial No. 1584592.

This invention further extends to apparatus for performing the assessment of bias distortion in a binary waveform in accordance with the method described above.

Accordingly, a further aspect of this invention provides apparatus for assessing any bias distortion present in a binary waveform, which apparatus comprises: a clock pulse adjustable to provide a preselected number of clock pulses per binary element period; an element clock for counting clock pulses so as to be driven repeatedly through element periods, the element clock requiring said preselected number of pulses to be driven through each element period; first and second edge detectors for the binary waveform, the first edge detector detecting a positive-going edge and providing an output to reset the element clock and the second edge detector providing an output on detection of the next negative-going edge following an output from the first edge detector; an averaging counter arranged to start counting clock pulses on receipt -of an output from the second edge detector and to stop counting clock pulses when the element clock has reached a pre-set point in its period; and means to initiate a chosen number of counting sequences by the averaging counter and then to transfer the average of the counts to a display device.

Said means preferably comprises a counter in the form of a measurment clock (i.e. a counter with a preset period), which measurement clock counts the number of outputs from the first edge detector and at the completion of a period causes the total count in the -averaging counter to be averaged and transferred to a display device.

Advantageously, the period of the measurement clock is 64, and, for a binary count in the averaging counter, the averaging is effected by transferring the binary count except for the 6 least significant binary bits this having the effect of dividing the total count by 64.

The display device is preferably in the form of a linear array of lamps, and conveniently of light-emitting diodes, arranged such that the central lamp is lit for zero bias distortion, an appropriate lamp to one side or the other of centre being lit for mark bias distortion or space bias distortion respectively.

Preferably, the apparatus is in the form of a self-contained test-set, which can be connected to any desired binary waveform transmission line and used to measure any bias distortion present in the binary waveform being transmitted therealong. Advantageously, the apparatus also includes the means necessary for adjusting the clock pulse source to deliver the preselected number of pulses per binary element period, in accordance with the method described above. The test set may furthermore include means for assessing total distortion, in accordance with the method described and claimed in our copending Application No.

23574/77 (Serial No. 1584592).

A test-set employing the methods described above can be constructed so as to be especially simple to operate. With a display device in the form of a linear array of lamps - and preferably light-emitting diodes - and a three-position switch provided to adjust the apparatus respectively for the setting of the internal clock pulse source, the assessment of bias distortion and the guaging of fortuitous distortion by assessing the total distortion, the test-set is simple to operate and reliable results can be obtained. A further control has to be provided to allow the manual adjustment of the clock pulse source dependent upon the display indication when the switch is in the appropriate position, and re-set button can be provided for a peak total distortion hold display when the set is adjusted to assess total distortion.

By way of example only, one specific embodiment of a test-set arranged to perform the assessment of any bias distortion which may be present in a binary waveform, in accordance with the invention, will now be described with reference to the accompanying drawings, in which: Figure I is a block diagram of that part of the test-set employed to adjust the element clock thereof to the required element delivery rate for the waveform being tested; Figure 2 is a diagram showing the method of adjustment of the speed of the element clock; Figure 3 is a block diagram showing the arrangement of the test-set, for assessing bias distortion; and Figure 4 is a diagram showing the method of assessing bias distortion.

Referring initially to Figure 1, there is shown a block diagram of those parts of the test-set which are employed in the adjustment of the test-set so that its internal clock pulse source runs at the required rate; namely, 100 clock pulses per element period of the binary waveform being tested. In the figure there is shown a clock pulse source 10 having a manual control 11 to adjust the rate of delivery of clock pulses. The clock pulse source 10 drives an element clock C2 arranged to count from 0 to 99 and which then provides an output indicating the end of one element period and the commencement of the next, the element clock returning to 0 for the next count sequence.

Two edge detectors for the waveform to be tested are provided, detector 12 of which being arranged to detect a start edge (relevant to stop/start transmissions only) and detector 13 of which being arranged to detect the first positive-going edge following the detection of a start edge. Edges of the same sense are used to cancel out assessment errors due to any bias distortion present in the incoming waveform. For the case of an isochronous waveform, the start edge is any positive-going edge, though for stop/start transmissions, the start edge is the first edge of a character made up from typically - 7 binary elements. Detector 12 inhibits the action of detector 13 until a start edge has been detected; after this, the detector 13 serves to detect the next posi tive-going edge of the waveform being tested.The start edge detector 12 serves to re-set the element clock C2 to zero, the clock running at the clock pulse rate set by control 11 as soon as the test-set is turned on. Of course, the rate set by control 11 probably is not the required rate for the waveform being tested. The clock pulses from the source 10 are supplied to an averaging counter C1 which however is inoperative until it is started by detector 13 detecting the first positive-going edge following a start edge. As soon as this first positive-going edge is detected, the averaging counter C1 commences to count clock pulses from the source 10 and continues to count until averaging counter C1 is stopped by a signal from detector 14, this detector serving to detect when the element clock C2 has reached a count of 50.Detector 12 looks once more for a start edge. This cycle is repeated 7 more times, so that averaging counter C1 accumulates in it the result of 8 separate counting cycles. A measurement clock C3 serves to count the 8 cycles of operation and after the eighth cycle, causes the accumulated count in the averaging counter C1 to be transferred to an output device, after the count has been divided by eight in order to effect the averaging thereof.Because eight results are counted, it is especially easy to divide the count in averaging counter C1 by eight to obtain the average; all that is necessary is to "lose" the three least significant bits of the binary count in averaging counter C1 to effect the divide-by-eight, the remaining bits of the count in averaging counter C1 being transferred to an output device 16. If the averaged count is abnormally low or high, an 'out-of-range' indicator is operated. An averaged count of 50 implies frequency synchronism between the incoming data and the element clock.

Character clock 15 is associated with the element clock C2 and serves to inhibit the operation of the start edge detector 12 for 7 binary periods as defined by element clock C2 following the detection of a start edge.

This character clock has no significant function when operating on an isochronous binary transmission, though when operating in the stop/start mode the character clock 15 serves to prevent more than one result in the averaging counter Cl being made per character of the binary waveform. With a stop/start transmission having a different number of binary element periods per character, the character clock 15 should be adjusted accordingly, so as to prevent the detector 12 re-setting the element clock C2 more than once for each character.

The output device 16 comprises a decoder for the binary count of averaging counter C1, and a latch arrangement driving a linear display of 43 light-emitting diodes (LEDs).

The decoder and latch are arranged to illuminate the centre LED of the row thereof if the averaged count is 50, for in this case the clock is running at the correct speed. For counts of less than 50 an appropriate LED to one side of the centre is illuminated to show that the element clock (and hence the clock pulse source) is running too quickly relative to the received binary waveform, whereas for counts of greater than 50 an appropriate LED is illuminated to the other side of the centre, showing that the clock is running too slowly.

By observing which LED is illuminated, an operator may adjust control 11 to adjust the speed of the clock pulse source and hence the element clock until the centre LED of the array is illuminated.

The display is arranged so that the two extreme LEDs serve as the out-of-range indicator and when both LEDs are illuminated the operator adjusts the control 11 until both are extinguished; after this, the sense in which the control should be operated is indicated by the LED then illuminated and the operator can properly adjust the speed until the centre LED is illuminated.

Referring to Figure 2, there is shown an example of a character of a stop/start binary waveform with which the test-set element clock is to be synchronised. This character consists of seven binary elements, there being a positive-going start edge 20 at the commencement of the character. the waveform remaining at level ZERO (in telegraph notation) for three binary element periods before returning to level ONE - for one binary element period. Following the period at level ONE, the first positive-going edge 21 following the start edge occurs. with the waveform remaining at level ZERO for two binary element periods and then returning back to the rest state ONE at the end of the character. The test-set clock element counting is shown by the lower line 23 in the Figure.

The start edge 20 is detected by detector 12 and at this point the element clock C2 is re-set to zero as shown by the lower line 23.

The element clock is shown as running too fast, and in this case the clock has advanced from a count of 99 to 00 before the completion of the first binary element period. Here, the clock pulse source is shown as providing about 11 further pulses before the completion of the binarv waveform element period. By the time the first positive-going edge is detected bv detector 13, the element clock C2 has run through four completely binary element periods.

plus a further 45 counts. The arrival of the first-positive-going edge starts averaging counter C1 counting clock pulses until detector 14 stops the counting because the element clock has reached a count of 50: for this case therefore the averaging counter Cl will reach a count of only 5 before being stopped by detector 14.

This process is repeated for the next 7 characters, and the 8 results averaged to provide the output measurement to the display of LEDs. It will be appreciated that the count in averaging counter C1 will be dependent both upon the relative periods of the test-set clock and the binary waveform as well as on the number of binary element periods between the start edge and the first positive-going edge following the start edge.

The count will also be effected by any fortuitous distortion present. By averaging over 8 counts, the effects of fortuitous distortion can largely be eliminated, and provided that a succession of different data characters is being transmitted, the effect of the spacing between the start edge and the first following positive-going edge will also tend to average out. However, the test-set is not intended to provide a calibrated reading of the speed error but is instead simply intended to show whether the element clock is running too quickly or too slowly, so that the operator may suitably adjust the speed of the clock.When the element clock is synchronised properly, the centre LED of the display thereof will be illuminated, quite irrespective of the spacing between the start edge and the first following positive-going edge, though of course fortuitous distortion can still upset the counting.

Referring now to Figures 3 and 4, there is respectively shown the apparatus and method used to assess the amount of bias distortion present in a received binary waveform. The same clock pulse source 10 and associated element clock C2 are used as has been described above, and of course the clock pulse source must be synchronised in the manner just-described with the incoming waveform. The two detectors 12 and 13 are modified, so that the former serves to detect a positive-going edge of the waveform and the latter detects the next following negative-going edge of the waveform.

The output from detector 12 is used to re-set the element clock to zero, in the manner described above, and the detector 13 is used to start the averaging counter C1 which is also arranged to receive clock pulses from the clock pulse source. Again, a detector 14 detects when the element clock is precisely half-way through counting a binary element period and an output from this detector serves to stop the counting of clock pulses in averaging counter C1.

Measurement clock C3 is modified so as to count the number of positive edges detected by detector 12, and to provide an output to averaging counter C1 when 64 such edges have been detected. When 64 such edges have been detected, the accumulated count in averaging counter C1 is divided by 64 (by "losing" the 6 least significant bits of the binary count) and the averaged count is then transferred to the display device 16 for display of the assessed bias distortion.

The waveform shown in Figure 4 represents a data character of a stop/start transmission. This character has a positive-going start edge 25 followed by 3 binary element periods at level ZERO. Then, there is a negative- going edge 26 having mark bias distortion, followed by 2 binary element periods at level ONE, followed by a further binary element period at level ZERO, the negative-going edge 27 of which is also subjected to the same mark bias distortion as the first-mentioned negative- going edge 26. As is shown by the element clock line below the character, the element clock C2 has been synchronised so that there are 100 clock pulses per binary element period, ignoring the effects of bias distortion and any fortuitous distortion which may also be present. The element clock C2 is re-set to zero at the arrival of the start edge 25.This clock is only part-way through the counting of the third binary element period when the negative- going edge 26 following the start edge 25 is received. This is because this negative-going edge 26 is subjected to mark bias distortion and thus arrives relatively early. The arrival of this edge causes averaging counter C1 to start counting clock pulses until the element clock has run half-way through the next following binary element period - but by this time averaging counter C1 has accumulated a count of 85. If there were no bias (or fortuitous) distortion counter C1 would store 50 counts. The next positive-going edge 28 re-sets the element clock C1. Because bias distortion is constant in nature, the next counting sequence will add another count of 85 to counter C1.The count will only vary if fortuitous distortion is present, or might vary by very small amounts because of the probabilistic nature of the arrival time of each negative-going edge.

Typical amounts of bias distortion are +6%, though the test-set may measure up to 50% space bias distortion or up to 50% mark bias distortion. For these two extreme cases, the average count will be 99 and 0 respectively with a linear variation in distortion over this range.

Reference is directed to our copending Application No. 23574/77 (Serial No.

1584592), in which is described and claimed, as preferred aspects, modified forms of the above apparatus, for assessing total distortion to allow the gauging of any fortuitous distortion present in a received binary waveform.

WHAT WE CLAIM IS: 1. A method of assessing any bias distortion present in a received binary waveform, comprising: adjusting a clock pulse source to deliver to an element clock a

**WARNING** end of DESC field may overlap start of CLMS **.

Claims

**WARNING** start of CLMS field may overlap end of DESC **.

characters, and the 8 results averaged to provide the output measurement to the display of LEDs. It will be appreciated that the count in averaging counter C1 will be dependent both upon the relative periods of the test-set clock and the binary waveform as well as on the number of binary element periods between the start edge and the first positive-going edge following the start edge.

The count will also be effected by any fortuitous distortion present. By averaging over 8 counts, the effects of fortuitous distortion can largely be eliminated, and provided that a succession of different data characters is being transmitted, the effect of the spacing between the start edge and the first following positive-going edge will also tend to average out. However, the test-set is not intended to provide a calibrated reading of the speed error but is instead simply intended to show whether the element clock is running too quickly or too slowly, so that the operator may suitably adjust the speed of the clock.When the element clock is synchronised properly, the centre LED of the display thereof will be illuminated, quite irrespective of the spacing between the start edge and the first following positive-going edge, though of course fortuitous distortion can still upset the counting.

Referring now to Figures 3 and 4, there is respectively shown the apparatus and method used to assess the amount of bias distortion present in a received binary waveform. The same clock pulse source 10 and associated element clock C2 are used as has been described above, and of course the clock pulse source must be synchronised in the manner just-described with the incoming waveform. The two detectors 12 and 13 are modified, so that the former serves to detect a positive-going edge of the waveform and the latter detects the next following negative-going edge of the waveform.

The output from detector 12 is used to re-set the element clock to zero, in the manner described above, and the detector 13 is used to start the averaging counter C1 which is also arranged to receive clock pulses from the clock pulse source. Again, a detector 14 detects when the element clock is precisely half-way through counting a binary element period and an output from this detector serves to stop the counting of clock pulses in averaging counter C1.

Measurement clock C3 is modified so as to count the number of positive edges detected by detector 12, and to provide an output to averaging counter C1 when 64 such edges have been detected. When 64 such edges have been detected, the accumulated count in averaging counter C1 is divided by 64 (by "losing" the 6 least significant bits of the binary count) and the averaged count is then transferred to the display device 16 for display of the assessed bias distortion.

The waveform shown in Figure 4 represents a data character of a stop/start transmission. This character has a positive-going start edge 25 followed by 3 binary element periods at level ZERO. Then, there is a negative- going edge 26 having mark bias distortion, followed by 2 binary element periods at level ONE, followed by a further binary element period at level ZERO, the negative-going edge 27 of which is also subjected to the same mark bias distortion as the first-mentioned negative- going edge 26. As is shown by the element clock line below the character, the element clock C2 has been synchronised so that there are 100 clock pulses per binary element period, ignoring the effects of bias distortion and any fortuitous distortion which may also be present. The element clock C2 is re-set to zero at the arrival of the start edge 25.This clock is only part-way through the counting of the third binary element period when the negative- going edge 26 following the start edge 25 is received. This is because this negative-going edge 26 is subjected to mark bias distortion and thus arrives relatively early. The arrival of this edge causes averaging counter C1 to start counting clock pulses until the element clock has run half-way through the next following binary element period - but by this time averaging counter C1 has accumulated a count of 85. If there were no bias (or fortuitous) distortion counter C1 would store 50 counts. The next positive-going edge 28 re-sets the element clock C1. Because bias distortion is constant in nature, the next counting sequence will add another count of 85 to counter C1.The count will only vary if fortuitous distortion is present, or might vary by very small amounts because of the probabilistic nature of the arrival time of each negative-going edge.

Typical amounts of bias distortion are +6%, though the test-set may measure up to 50% space bias distortion or up to 50% mark bias distortion. For these two extreme cases, the average count will be 99 and 0 respectively with a linear variation in distortion over this range.

Reference is directed to our copending Application No. 23574/77 (Serial No.

1584592), in which is described and claimed, as preferred aspects, modified forms of the above apparatus, for assessing total distortion to allow the gauging of any fortuitous distortion present in a received binary waveform.

WHAT WE CLAIM IS: 1. A method of assessing any bias distortion present in a received binary waveform, comprising: adjusting a clock pulse source to deliver to an element clock a

preselected number of clock pulses for each binary element period, the element clock requiring said preselected number of pulses to be driven through each element period; re-setting the element clock on the arrival of one of a positive- or negative-going edge of the waveform; starting an averaging counter to count clock pulses on the arrival of the next following edge of the opposite sense; stopping the counting of clock pulses in the averaging counter when a pre-set number of clock pulses have been counted by the element clock following the point at which said next following edge should have arrived if no distortion is present; repeating the counting sequence by starting and stopping the averaging counter in the aforesaid manner for a chosen number of times; averaging the counts obtained thereby; and then displaying the averaged count to indicate the degree of any bias distortion present in the waveform.
2. A method according to claim 1, in which said counting sequence is repeated 64 times and then the 64 counts obtained thereby are averaged, in order to achieve an assessment of the amount of bias distortion present.
3. A method according to claim 1 or claim 2, in which the clock pulse is adjusted to deliver clock pulses at a rate of 100 for each binary element period.
4. A method according to any of the preceding claims, in which the present number of clock pulses which causes the stopping of counting of clock pulses by the averaging counter is precisely one half of the preselected number of pulses per binary period, as counted by the element clock.
5. A method according to any of the preceding claims, in which the averaged count is suitably encoded and then displayed on a linear array of lamps.
6. A method according to claim 5, in which each lamp of the array is a lightemitting diode.
7. A method according to claim 5 or claim 6, in which a measured count indicating no bias distortion illuminates the central lamp of the array with counts indicating mark bias distortion or space bias distortion illuminating an appropriate lamp of the display respectively to one side or the other of the central lamp.
8. A method according to any of claims 5 to 7. in which a separate lamp is provided for each 2% change in assessed bias distortion.
9. A method according to any of claims 5 to 8, in which the lamp array consists of 43 separate lamps, arranged to display assessed distortions of up to 40neo each side of zero assessed distortion.
10. A method according to any of the preceding claims, in which the element clock is reset at the arrival of a positivegoing edge, with the averaging counter being started on arrival of the next following negative-going edge.
11. A method according to any of the preceding claims, in which adjustment of the clock pulse source to deliver the preselected number of pulses per binary element period to the element clock is effected by: re-setting the element clock at the arrival of a start edge of the binary waveform; detecting the first following edge of the same sense as the start edge; commencing the counting of clock pulses at the detection of the said first following edge; stopping the counting of clock pulses in the averaging counter when the element clock has counted a pre-set number of clock pulses following the completion of the counting of the immediately preceding element clock period; repeating the aforesaid averaging counter counting sequence for a chosen number of times; averaging the counts obtained thereby; and then displaying the averaged count as an indication of the error in the clock rate relative to the period of the binary waveform so that the clock pulse source rate may be adjusted dependent upon the displayed averaged count.
12. A method according to claim 11 for use with stop/start waveforms, in which only one count is made on each character of the waveform.
13. A method according to claim 12, in which the detection of a start edge is inhibited after the detection of one such edge for a period of a number of successive binary element periods, the precise number depending upon the character alphabet chosen.
14. A method according to any of claims 11 to 14, in which the said pre-set number of clock pulses at which counting thereof is stopped is precisely one half of the said preselected number of clock pulses.
15. A method according to any of claims 11 to 14, in which the count for clock pulse source speed is displayed on a linear array of lamps.
16. A method according to any claims 11 to 15, in which the counting for clock speed is effected 8 times and then averaged for display.
17. A method according to any of the preceding claims, in which a manual control is provided for the adjustment of the clock pulse source rate.
18. A method according to claim 1 and substantially as hereinbefore described with reference to the accompanying drawings.
19. Apparatus for assessing any bias distortion present in a binary waveform, which apparatus comprises: a clock pulse source adjustable to provide a preselected number of clock pulses per binary element period; an element clock for counting clock pulses so as to be driven repeatedly through element periods, the element clock requiring said preselected number of pulses to be driven through each element period; first and second edge detectors for the binary waveform, the first edge detector detecting a positive-going edge and providing an output to reset the element clock and the second edge detector providing an output on detection of the next negative-going edge following an output from the first edge detector; an averaging counter arranged to start counting clock pulses on receipt of an output from the second edge detector and to stop counting clock pulses when the element clock has reached a pre-set point in its period; and means to initiate a chosen number of counting sequences by the averaging counter and then to transfer the average of the counts to a display device.
20. Apparatus according to claim 19, wherein said means comprises a counter in the form of a measurement clock which counts the number of outputs from the first edge detector and at the completion of a period causes the total count in the averaging counter to be averaged and transferred to a display device.
21. Apparatus according to claim 20, wherein the period of the measurement clock is 64.
22. Apparatus according to claim 19 or claim 20, wherein the display device is in the form of a linear array of lamps arranged such that the central lamp of the array is lit when there is no bias distortion, and a lamp to one side or the other of the central lamp is lit when mark bias distortion or space bias distortion respectively is present.
23. Apparatus according to claim 22, wherein each lamp of the display device is a light-emitting diode.
24. Apparatus according to claim 19 and substantially as hereinbefore described with reference to Figure 3 and Figure 4 of the accompanying drawings.
25. Apparatus according to any of claims 19 to 24, arranged in the form of a self-contained test-set, including means for adjusting the clock pulse source to deliver the preselected number of pulses per binary element period, which test-set is connectible to a binary waveform transmission line and used initially to set the clock pulse rate and then to measure any bias distortion present in the binary waveform transmitted along the line.