US2303909A - Transmission of electrical signals - Google Patents

Description

Dec. 1, 1942.

-A. D, BLUMLEIN TRANSMISSION OF ELECTRICAL SIGNALS Filed May 23, 1939 2 sheets-sheet 1 42 0 INVENTOR ALAN DOWER BLl/M E/A/ ATTORNEY Dec. 1, 1942 D BLUMLElN 2,303,909

TRANSMISSION OF ELECTRICAL SIGNALS Filed May 23, 1939 2 Sheets-Sheet 2 INVENTOR ALA/V DOWER BLUMLE/N yam ATTORNEY Patented Dec. 1 1942 UNITED TRANSMISSION or ELECTRICAL SIGNALS Alan Dower Blumlein, Ealing, London, England,

assignor to Electric & Musical Industries Limited, Hayes, Middlesex, England, a company of Great Britain Application May 23, 1939, Serial No. 275,169 In Great Britain May 30, 1938 6 Claims.

The present invention relates to systems for the transmission of electrical signals and is concerned with the correction of variations in the effective amplitude of signals representative of intelligence, such as may arise in the transmission of said signals as the result of the complete or partial loss of the D. C. component of said signals, the incorrect representation of that component, or varying attenuation of the signals.

In the specification of Willans Patent No. 2,252,746, issued Aug. 19, 1941 there is described a system wherein the D. C. component is re-inserted into electrical signals which periodically attain fixed amplitude values by the use of a device such as a diode valve in conjunction with a condenser and leak resistance. For example, in Figure 10 of that patent there is shown an arrangement for applying television signals to the control electrode of a cathode ray tube used for the reproduction of the television signals. A condenser is connected in the lead to the control electrode of the tube and a diode valve is connected from the control electrode of the tube to a suitable biasing arrangement, the cathode of the diode valve being connected to the control electrode. Television signals with the picture signals in a positive sense and with negative synchronising pulses are applied to the condenser,

through which the D. 0. component of the television signals is lost. The negative synchronising pulses cause the diode valve to conduct when the potential of the cathode becomes equal to the potential of the anode of the diode and the condenser is given a certain charge which biases the control electrode of the cathode ray tube. When the peaks of the synchronising pulses become more negative due to the loss of the D. C. component, the charge on the condenser will increase. Should, however, the peaks now become less negative owing to a change in the D. 0. level of the signal then the charge on the condenser will not be altered because the valvescwill be insulating and will thus be unable to harge the condenser to the new peak values. In order to avoid this difiiculty the leak resistance is connected across the diode valve so that between synchronising peaks the charge slowly leaks away, thus making it possible for the diode to charge the condenser during those synchronising pulses, the amplitudes of which are less than the amplitude of the maximum negative pulse.

Unfortunately, the connection of the leak resistance introduces a source of error into the D.

C. re-insertion in that the D. 0. level varies be tween synchronising peaks owing to the discharge of the condenser; in other words, a small sawtooth potential is added to the picture signal. The sense of the saw-tooth potential superimposed on the picture signal is such as to decrease the amplitude of the picture signals towards the end of the discharge period, for example, towards the end of a line scan.

A similar error is introduced by other re-insertion systems which involve the use of a uni-directional conducting device for charging a condenser and a leak resistance which is a necessary feature of such systems. For example, there is described in the Specification oi Browne Patent No. 2,190,752, issued Feb. 20, 1940, with reference to Figure 2 of that patent a system of D. C. reinsertion which is similar to that already discussed, in that a coupling condenser is used in connection with a leak resistance. The junction of the coupling condenser and the leak resistance is taken to the anode of a screen grid valve, to the control grid of which switching pulses are applied, the other end of the leak resistance being taken to a source of positive potential. As applied to D. C. re-insertion with respect to the peaks of the synchronising pulses, the operation is briefly as follows. Switch pulses of a duration slightly less than that of the synchronising pulses are applied to the control grid of the screen grid re-inserting valve and cause this valve to conduct during those periods when the synchronising pulses are applied to the coupling condenser. This has the eflec't of charging the coupling condenser to a value depending on the level of the synchronising pulses. The leak resistance serves a similar purpose to the leak resistance used in the method of re-insertion according to Patent No. 2,252,746, supra. In this case, however, the variation of the D. C. level during the picture signals due to the super-imposition of the saw-tooth shaped discharge potential is such that the sense of the saw-tooth potential tends to make the amplitude of the picture sig nals increase towards the end of the discharge period owing to .the leak resistance being taken to a positive source of potential.

The same difflculty of a change in signal level due to the discharge of a condenser between datum portions occurs in some methods of automatic volume control of television signals, for example, according to systems described in the specification of U. S. patent to Blumlein No. 2,224,134, issued on Dec. 10, 1940, and in the U. S. patent of Blumlein Ser. No. 275,823, filed May 26, 1939, which involves the use of leak resistances, or their equivalents,

It is the object of the present invention to provide an improved method of correcting electrical signals with a view to avoiding these difliculties.

According to the present invention there is provided apparatus for substantially correcting unwanted variations of arecurrent datum level in electrical signals such as may arise as a result of varying attenuation of said signals or the partial or complete loss of their D. C. and/or low upon the correction of the variation of said datum level is reduced.

According to one feature of the invention the observing devices are situated at successive points in the signal channel through which said signals pass.

According to a further feature of the invention a balanced, or push-pull signal channel is provided and an observing device is associated with each side of said balanced signal channel.

For the purpose of describing the invention more in detail reference will now be made to the accompanying drawings in which Figures 1, 2, 3

and 4 illustrate particular embodiments accord-' ing to the invention.

Referring to Figure 1, which illustrates one embodiment according to the invention of a method of D. C. re-insertion, television signals with, for example, a waveform of the Marconi/E. M. I. type such as is illustrated in the Wireless World of October 4, 1935, page 373, are

applied to the terminal Ill and thence through a coupling condenser i3 to the control grid of a cathode follower valve I2. The anode and screen grid of valve i2 are taken to suitable positive potentials and the cathode of valve I2 is taken via resistance l5 to a source of negative potential. The output from the cathode follower valve I2 is taken via coupling condenser l9 to the control electrode of an amplifier valve 2| which has an anode impedance 22 and output terminal 24. The sense of the television signals applied to terminal II are such that the synchronising pulses are in a negative sense. The purpose of the valve 5, which has its anode connected to the control grid of valve I2 is to re-insert the D. C. component at the grid of valve l2 according to the system described in the specification of U. S. Patent No. 2,190,753 supra. In other words, positive synchronising pulses (which may with advantage be slightly shortened in duration) are applied to the control grid of valve I6 through the coupling condenser I 8 and leak resistance I! in such a manner that the valve IE only conducts during the synchronising pulse periods. The condenser I3 is thus charged during the synchronising pulses to a datum level determined by the cathode potential of valve l6 so that the D. C. is re-inserted on the grid of valve l2. The resistance H, which is connected between the grid of valve l2 and a source of positive potential, tends to make the control grid of the valve more positive so that a saw-tooth shaped potential is super-imposed on the picture signals. The D. C. is again re-inserted at the grid of valve 2i by means of the diode 23 which operates in the manner described in the specification of U. S. Patent No. 2,252,746 supra. The battery 25 provides a suitable bias potential for the grid of the valve 2|. The leak resistance 20, which is taken to a source of negative poten- I the first place according to the method described in the specification of U. S. Patent No. 2,190,753 supra and secondly by the method described in the specification of U. S. Patent No. 2,252,746 supra, it will be obvious that the order can be reversed if desired. It is also not necessary for the two re-establishments to be situated at the same place, for example, re-insertion according to one method may be applied at one end of a cable or a radio link while the re-insertion ac-- cording to the reverse method may be applied at the other end of the cable or radio link. It is not necessary to associate only two re-establishment methods together, for example, the D. C. may bare-established twice by means of the method described in the specification of U. S.v

Patent No. 2,252,746 and subsequently re-inserted by an inverted method of re-establishment such that the rate of drift of the D. C. potential during the picture signal of the latter method of D. C. re-establishment neutralises the sum of the potential drifts on the first two re-establishments In order that the rates of potential drift may be adjusted to a nicety the grid leak resistances, for example, resistances l4 and 20, are preferably of high value and are fed from sources of high potential so that the rate of drift is substantially independent of the picture signal amplitude. Thus, for re-inserting the D. C. component when the signal amplitude is say 10 volts and the unidirectional conducting device tends to drive the condenser negative, the leak resistance may be taken to a positive potential of say 50 volts, the value of the leak resistance being related to the value of the coupling condenser so as to produce a rate of drift of the charge on the condenser greater than any probable drift of the picture signals.

In an alternative arrangement the method according to the invention may be applied to a push-pull or balanced transmission system. Figure 2 shows such an arrangement. The ampli- .fier 26, which may be of the type described in the specification of U. S. Patent to Blumlein No. 2,185,367, issued June 2, 1940, provides two outputs which are fed into similar condensers 21 and 28. The D. C. is re-established with regard to the recurrent datum level, such as the black datum level or the synchronising signal of a television wave, by means of the uni-directionally conductive devices 29 and 30 shown as pentode valves. These pentodes act as diodes as regards their anode-cathode circuits when the control grids are switched on. The valves 29 and 30 are switched on by positive pulses which are applied at terminal 42 to the control grids of valves 29 and 30, via coupling condensers 31 and 38 and leak resistances 35 and 36 respectively. The positive pulses applied to the grids of these valves cause grid current to flow, thus charging the grid condensers 31 and 38 and producing a negative bias on the grids of these valves in the absence' of switching pulses from 42, which bias is sufficient to cut oil the currents in valves 29 and 30. The production of such pulses is described, for

example, in the specifications of U. S. Patents Nos. 2,190,753 and 2,224,134. These pulses may, 1

period for a standard Marconi/E. M. I. television wave form. If the synchronising period is used the pentode re-inserting valve which has the syn-v chronising pulses applied to it in a positive sense may bereplaced by an un-switched diode valve, but this is not preferred as the different characteristics of the diode and th longer negative period (especially during long frame pulses) spoil the symmetry of the arrangement. In addition to the pentode 29 and 30, two similar leak resistances 3| and 32 are connected to the output side of each of the coupling condensers 21 and 23, that is, to the leads33 and 34. These resistances 3| and 32 are taken to a suitable source of positive potential and thus they tend to charge the condensers2l and 28 so that the potentials of lines 33 and 34 tend to drift positively during the intervals between successive datum periods.

During datum periods the valves 29 and 30 are made conductive and charge condensers 21 and 28 negatively so as to make the potential of lines 33 and'34 approximately earth potential during datum periods. As both lines 33 and 34 drift in potential, due to the leak resistances 3| and 32, in the same direction as regards push-pull signals, that is, the difference of potential of the lines 33 and 34, these potential drifts will cancel so that the only drift of correct level will be due to the drift of the incoming signal. The lines 33 and 34 are followed by an amplifier or transmission link 39' which is so arranged as to modify substantially only the push-pull input potentials, that is, the difierence of the potentials between lines 33 and 34, to the exclusion of the absolute average potentials of lines 3| and 32, that is, the push-pull potentials. A suitable amplifier arrangement for stage 39 is described, for example, in the specification of U. S. Patent No. 2,185,367 supra. The potential drifts due to the similar leak resistances 3| and 32 operating from a common large positive potential on similar condensers 21 and 28 will therefore cancel as regards the output of stage 39 which output may be balanced or unbalanced as-desired, the output being derived from terminals 40 and 4|. The tendency of the valves 29 and 30 to charge condensers 21 and 28 to a slightly different potential during the frame pulse (there are twice as many datum periods per line with the pulse generating methods described) due to the finite resistance of the valves, is also neutralised by the push-pull stage 39.

As an example, suppose a television wave is considered havingjan amplitude from black to white of say 10 volts on these leads 33v and 34, and suppose the D..C. component is lost through the operation of an efi'ective time constant of 0.01 second. If the picture changes from full black to full white, the black level will move nearly 10 volts relative to the A. C. mean value and the initial drift rate will be 10 volts in 0.01 second, which with a scanning frequency of 10,000 lines per second will be 0.1 volt per line. It is desirable that the drift rate due-to the leakage is greater than this so that the valves 29 and 30 conduct during each datum period,- that is, once per line. It is desirable that the condensers 21 and 23 should be large compared with any other stray capacities associated with this part of the circuit in order that a symmetrical arrangement be preserved. For this reason it is preferable that the output from amplifier stage 23 has a low impedance. The

capacity of these condensers should be small enough, however, as to be substantially charged through the resistance of the valves 29- and 30 and the output circuit of condensers 21 and 23 in series during one datum period. For example, making the resistance of a pentode 29 plus the output impedance of line 33, say, 1,500 ohms and the datum period at 4 microseconds, it should be arranged that the time constant, that is, capacity of the condenser 21 multiplied by 1,500 ohms should be less than, say 2 microseconds, that is. the capacity of condenser 21 may be 0.001 microfarad. The current through resistance 3| must be greater than that required to charge a 6011-. denser ofv 0.001 microfarad through the 0.1 volt in one line period (0.0001 second) that is, greater than 1 microampere. The potential from which resistance 3| is supplied should be large compared with 10 volts so that the rate of drift of potential is unaflected by the signal amplitude, say 50 volts. The leak resistance 3| must then be less than 50 megohms. In practice, very much higher rates of drift of incoming signal may be allowed for and the potential may be or 200 volts with leak resistance values of 10megohms.

The datum period used in the above method was repetitive once per line at least. Alternatively, a once per frame datum period may be used although of cource changes as rapid as that described above, that is, 0.01 second, cannot then be corrected.

Figure 3 shows an arrangement for observing the black level by charging two condensers which are provided with leak resistances, the purpose of the observation being either for D. C. re-insertion or for A. V. C. In Figure 2 the two condensers were connected in the signal leads so that their charge was added directly to the signal. In the present instance the condensers are situated in a separate path from the signal path and the potentials across them may be used either to add to the signal afterwards, to feed back into the input circuits, or to control the gain of an amplifier. Referring to Figure 3, 45 is a push-pull, or balanced, transmission link such as an amplifier supplying a push-pull output signal to leads 46 and 41, which are connected to output terminals 63 and 64. The lead 46 is connected to the control grid of a valve 5|, the cathode of which is connected to the cathode of valve 53 and to a suitable cathode load resistance 55. The cathodes of valves 5| and 53 are connected also to the cathode of a diode valve 43, the anode of which is connected via resistance 59 to a condenser 51 and its associated discharging resistance 6|. Similarly, the lead 41 is connected to the grid of valve 50, the cathode of which is connected to the cathode of valve 52 which is provided with a suitable cathode impedance 54 and is also connected to the cathode of a diode valve 48. The anode of this diode 48 is connected to another charging condenser 56 via resistance 58, the condenser 56 being connected to the discharging resistance 60 which is similar to resistance'tl. The two discharging resistances 30 and 6| are taken to a source of positive potential as are the anodes of the valves 50, 5|, 52 and 53. The valves 48, 49, 50, 5|, 52 and 53 act as two uni-directional conducting switches as described in co-pending U. S.

patent to Blumlein No. 2,258,732, issued Oct. 14, 1941. Negative switching pulses lying totally within the datum period are applied to terminal 62 which is connected to the control grids of valves 52 and 5!. Normally, terminal 62 is maintained positive but during the datum periods the negative pulse applied to terminal 62 switches off the triodes 52 and 53 so that the cathode potentials of these valves fall to values determined by lines 46 and 41 so rendering valves 48 and 49 conducting and charging condensers 56 and 51 to the cathode potentials ,of valves 50 and 5| respectively, in other words, the condensers are charged to potentials which are representative of the datum levels on lines 41 and 46 respectively. The

condensers 56 and 51 are preferably of equal value and are charged positively by the equal leak resistances 60 and lil from a source of positive potential. If. the condensers are not equal in value the time constants of the condensers with their associated leak resistances divided by the effective voltages associated with the leak resistances should be made equal in order to give an equal rate of drift to the two condensers.

The potentials on the two condensers 56 and 51 will be set during each datum period to a value depending on the datum potentials on lines 41 and 46, and the diflerence of potential will be equal to the difference of potential of the lines 46 and 41 during datum periods. After the datum period the potentials on the condensers will drift positively but since the rates of drift are equal the potential difference between the lines 56a and 51a will not be affected. The difference of potential across the condensers 56 and 51 which is obtained from the terminals 56a and 5111 may be added to the potentials of lines 46 and 41 in reverse phase so producing in effect a D. C. re-establishment. The difference of potential on the condensers is th same as if the clamp system of D. C. reinsertion had been employed. By a clamp system is meant a system, for example, as described in the specification of U. S. Patent No. 2,190,753 supra, wherein a 'bi-directionally conductive switch is employed so that the charging condenser is not discharged between datum periods. Alternatively, the potentials may be fed back in a negative sense into the push-pull input of stage 45, thus serving to re-establish D. C. on lines 46 and 41. In order for this to function correctly stage 45 must be a D. C. coupled push-pull amplifier and is preferably of the type described in the specification of U. S. Patent No. 2,185,367, so that it is substantially unaffected by the push-pull drift of the potentials on condensers 56 and 51 during the intervals between datum periods. The operation will then be similar to the clamp feedback system described in the U. S. patent application of Blumlein Ser. No. 275,168, filed May 23, 1939. In order that high speed correction can be obtained, the amplifier 45 is provided with two special input valves for the introduction of the feedback potentials. These input valves may have their anodes connected to the anodes of the input valves receiving the incoming signals. The resistances 58 and 59 may be necessary in order to prevent oscillation around the loop circuit as described in the patent application Ser. No. 275,168 supra.

The arrangement described above may with advantage be used to re-establish D. C. as described in the U. S'. patent application of Blumlein, Ser. No. 275,367, filed May 24, 1939, wherein the blacked out signals from a cathode ray type of television transmitting tube are used for the re-cstablishment of the D. C. component of the picture signals. In this case the terminal 62 of Figure 3 will have negative shortened blackout pulses applied to it. 5 In an alternative modification the potential difference across terminals 56a and 51a may be used to control the gain of an amplifier stage preceding the stage 45 and may thus control the potential difference between thclines 46 and 41. This is in effect a feedback arrangement so that if quick control is required the resistances 5B and 59 may be necessary in order to prevent oscillation. With such an arrangement carrier frequency signals may be amplified in a suitable amplifier and the output applied to two detector valves which may apply push-pull signals to the input of stage 45. Alternatively 45 may be an amplifier of the type described in the specification of U. S. Patent No. 2,185,367, supra which will convert the output from a single detector into a push-pull signal. The gain of the carrier frequency amplifier is thus adjusted with reference to the difference of potentials on terminals 56a and 51a, thus operating as an automatic volume control, the sense of control being arranged so that any difference of potentialon lines 45 and 41 during the datum periods tends to alter the amplifier gain so as to suppress such potential differences. Suppose, for example, B. Marconi/E. M. I. television signal is being transmitted with picture signals in the positive sense at line 46 and in the negative sense at line 41, the desired black levels being zero, then an. increase of signal strength will make the black datum level positive at line 46 and negative at line 41, which will tend to charge condenser 51 more positively than condenser 56. If the control is such that condenser 51 being positive relative to condenser 56 reduces the amplifier gain, the conditions described above will reduce the amplifier gain, thus tending to suppress any change in the push-pull signals at terminals 63 and 64.

Figure 4 shows a circuit arrangement for controlling the gain of an amplifier stage in accordtentials as derived from terminals 56a and 51a of Figure 3. Carrier frequency signals are applied to the primary of transformer 10, the secondary of which is connected to the grid of valve 1| wh ch represents a carrier frequency amplifier valve. An output transformer 16 is connected in the anode circuit of amplifier valve 1|. The

cathode of valve 1! is connected to the cathode of another screen grid valve 12, the anode of which is connected to a suitable positive potential. Associated with the cathodes of both valves 1| and 12 is a cathode resistance 13, which is taken to a source of negative potential. The cathodes are lay-passed to earth for radio frequencies by means cf a tuned circuit comprising inductance 14 and condenser 15. The control potentials from termnals 56a and 51a are applied to the terminals 11 and 18 which are connected to the control electrodes of the valves 1| and 12. Equal potential changes of both terminals 11 and 18 will not materially. affect the currents in the two valves. since impedance 13 is taken to a comparatively large negative potential. Hence the small potential drifts on condensers 56 and 51 of Figure 3 will not appreciably affect the gain of valve 1|.

tween terminals 11 and 18 the current of one valve will increase at the expense of the current in the other valve thus altering the gain of valve H. In the case considered above it was required ance with the difference of the two control po- If, however, there is a difference of potential bethat when condenser 51 became positive with respect to the potential on condenser 56 the gain of the amplifier should be reduced. Such a change would make terminal 18 more positive than terminal 11, and diverting current through valve 12 at the expense of current flowing in valve ll, thus lowering the mutual conductance of valve H and therefore the gain of the amplifier. In such an arrangement the charges on condensers 56 and 51 may be arranged to control a plurality of amplifier stages so as to give a large range of control.

In the examples described it has been assumed that the reference potential is always earth potential, for example, in Figure 2 where the cathodes of valves 29 and 30 are connected .to earth. Any other reference potentials may of course be used by inserting suitable sources of potential such as batteries at the appropriate places.

The invention has been described above as applied to television signals although it will be understood that this invention is not limited in its use to television systems. The datum period may be inherent in the signals when generated or may be inserted therein after the signals are generated or by passing the signals through a valve and switching ofi the valve at intervals so that pulses of datum level are interspersed through the signals.

I claim:

1. Apparatus for developing a substantially undistorted corrective signal from received recurring signals having predetermined voltage levels as compared to a fixed value, said corrective signal being proportional to the voltage level of the recurring signals, comprising means for storing energy during the recurring signal intervals in proportion to the voltage levels of the received signals, the value of said stored energy being subject to change in a predetermined polarity direction during at least a portion of the time between successive signal intervals. a second means for storing energy during the recurring signal intervals in proportion to the voltage levels of said received signals, said second stored energy being subject to change in the opposite polarity direction during at least a portion of the time between successive signal intervals, and means for combining both of said sets of stored energy so as to compensate for the changes in energy stored between the recurring signal intervals.

2. Apparatus for developing a substantially undistorted corrective signal from a received signal having a recurring datum value, said corrective signal being proportional to the value of the recurring signal, comprising a first thermionic tube having anode, cathode and at least one control electrode, a second thermionic tube having anode, cathode and at least one control electrode. means for impressing the received datum signals onto the anode-cathode path of each of said thermionic tubes in balanced relationship, a first capacitive means connected in a current carrying electrode-cathode path of said first thermionic tube, a second capacitive means connected in a current carrying electrode-cathode path of said second thermionic tube, a leakage path connected substantially across each of said capacitive means, means electrically connecting the first of said capacitive means to the second of said capacitive means, and output terminal means connected to said capacitive means and the said cathode whereby the variations in charge on one of said capacitive means during the interval be tween recurring datum portions of the received signal is substantially compensated by the variations in charge on the other of said capacitive means. 3. Apparatus in accordance with claim 2, wherein the received signal comprises a composite television signal, and wherein said recurring datum portion bears a definite relationship to the background level of the image transmitted, and wherein said combined stored energies represent the corrective signal for're-establishing the direct current component.

4. Apparatus for developing a substantially undistorted corrective signal from a received signal having a recurring datum value, said corrective signal being proportional to the value of the recurring signal, comprising a first thermionic tube having anode, cathode and at least one control electrode, a second thermionic tube having anode, cathode and at least one control electrode, means for impressing the recurring datum portion of said received signal onto the control electrode-cathode paths of both of said thermionic tubes in balanced relationship, first and second uni-directional conductors, means for connecting each of said uni-directional conductors into the anode-cathode path of one of the thermionic tubes, a first capacitive means, a second capacitive means, means for storing energy in said first and second capacitive means during the occurrence of the recurring signal having a datum value, a leakage path connected substantially in parallel with each of said capacitive means, and output terminal means for removing a corrective signal substantially free of changes in stored energy occurring during the time interval between the recurring signals having a datum value.

5. Apparatus in accordance with claim 4, wherein the received signal comprises a, composite television signal, and wherein said recurring datum portion bears a definite relationship to the background level of the image transmitted, and wherein said combined stored energies represent the corrective signal for re-establishing the direct current component.

6. In a corrective signal developing means wherein recurring signals are received having predetermined voltage levels as compared to a fixed value, and wherein the developed corrective signal must bear a definite relationship to the value of said recurring voltage level signals during intervals between occurrences of said signals, the method of developing said corrective signal which comprises the steps oi storing electrical energy of a value proportions to the value of the received signal during its time of occurrence, the stored electrical energy being subject to change in a predetermined polarity direction between recurring signal intervals, combining a signal indicative of the stored energy with the composite signal, storing a second quantity of energy dependent upon the same signal interval, the stored electrical energy being subject to change in the opposite polarity direction between received signal intervals, and combining a signal indicative of the value of said latter stored energy with the composite signal whereby a corrective signal will be produced that does not vary between voltage level signal intervals.

ALAN DOWER BLUMLEIN.

Images