US 3552383 A
Abstract available in
Claims available in
Description (OCR text may contain errors)
United States Patent  inventors Dean T. E. Krueger Rochester, Minn.;
Louis C. Sheppard, Birmingham, Ala. [21 1 App]. No. 789,778
 Filed Jan. 8, 1969 [45 Patented Jan. 5, 1971  Assignee lnterhational Business Machines Corporation Armouk, N.Y.
a corporation of New York  METHOD AND SYSTEM FOR ESTIMATION OF ARTERIAL PRESSURE 30 Claims, 5 Drawing Figs.
 U.S.Cl 128/2.05  lnt.Cl A6lb 5/02  Field of Search l28/205AS,
205M, 205MS  References Cited UNITED STATES PATENTS 2,756,741 7/1956 Campanella 128/2.05 2,827,040 3/ 1 958 Gilford l28/2.05 3,101 ,082 8/1963 Steen et a1.. l28/2.05 3,126,886 3/1964 Karsh l28/2.05
Primary Examiner-William E. Kamm Attorney-Sughrue, Rothwell, Mion, Zinn and MacPeak ABSTRACT: A method and apparatus for automatically following and monitoring systolic and diastolic blood pressure in an animal organism is disclosed. The pressure of an occluding cuff on the organism is controlled by decisional logic to increase or decrease in fixed increments the pressure in the cuff. The decision to increase or decrease depends upon the existence and amplitude of the Korotkofi' sounds detected during a sample period which may comprise a preset number of heartbeats. The Korotkoff signals are analyzed for the purpose of indicating whether the cuff pressure is just above or just below the systolic (or diastolic) blood pressure and the actual cuff pressure is changed accordingly. Multiple samples are made for systolic blood pressure and those samples for which an acceptable Korotkoff pulse is detected are averaged. Then multiple samples are made for diastolic blood pressure. The method and apparatus is based on the typical Korotkoff sound envelope seen when going from above systolic pressure to below diastolic pressure.
PATENTED JAN 5 |97I SHEET 2 OF 3 New cEEEES METHOD AND SYSTEM FOR ESTIMATION OF ARTERIAL PRESSURE BACKGROUND OF THE INVENTION ltis known in the prior art to provide continuous monitoring of systolic and diastolic blood pressure in animal organisms. This is necessary in many cases, particularly where a patient has had cardiovascular surgery. Only continuous monitoring can detect the blood pressure variations at the time they occur. In general, the prior art methods were based on the use of an occluding arm cuff and a microphone or other detector and employ a ramp method to identify the pressures at which Korotkofi sounds occur. These methods are usually designated for isolated measurements on aigeneral patient population. Special problems are encountered when indirect measurements of blood pressure must be made over many hours or days in patients with cardiovascular disease.
In one known system the systolic pressure is followed and monitored by raising the occluding cuff pressure above the systolic blood pressure (or what is believed to be the systolic pressure) each time an arterial pulse is detected by a transducer attached to the organism. The cuff pressure is then allowed to leak below the systolic pressure until the next arterial pulse is detected. For following and monitoring the diastolic pressure, the Korotkoff sounds are detected and the cuff pressure is controlled in a reverse manner to that described above. That is, the detection of a Korotkoffsound initiates adecrease in the cuff pressure to a value believed to be below the diastolic pressure and the cuff is then allowed to increase in pressure to above the diastolic pressure until the next Korotkoff sound is detected.
The latter described system operates to only either decrease or increase cuff pressure in response to detectedindications, the opposite form of pressure variation being left to a leakage mechanism. As a result, the cuff pressure is never held at a substantially constant rate and this magnifies the hunting effect and degrades the accuracy of the output. Furthermore, in the above system the operation is based on a, sound or nosound proposition and no provision is made for taking the sound variations into account.
SUMMARY or THE INVENTION In accordance with the present invention blood pressure is monitored by controlling cuff pressure of an occluding cuff and detecting and analyzing Korotkoff soundsover'a sample period. The Korotkoff sounds are analyzed to decide whether the cuff pressure is above or below-the systolic (or diastolic) blood pressure. Following the decision, the cuff pressure is in creased or decreased accordingly by a predetermined increment. The cuff pressure, analysis, and decision may be. digitally accomplished in a computerized. mechanism, the digital pressure output being converted to an analogue quantity for controlling a cuff pressure servosystem.
During the sample period the cuff pressure is held constant and the sample period is long enough to cover a plurality of heartbeats thereby allowing for detection'of multiple Korotkoff sounds. The detected sounds are classified by an amplitude parameter into acceptable and loud sound groups or classes. Acceptable sounds are between'first and second limits above the system noise level, and loud sounds are above the second limit. The classification of sounds'is based on a typ ical Korotkoff sound envelope obtained when cuff pressure is lowered from above systolic to below diastolic level. The magnitude of the sounds increases at first as the cuff pressure is decreased. It reaches a maximum when the cuff pressure is between systolic and diastolic levels and decreases as the arm cuff pressure approaches diastolic level. This phenomenon results from the opening of the artery with each pulse when the arterial pressure exceeds the cuff pressure. The method of pressures at which weak sounds at the ends of the sound envelope are detected.
BRIEF DESCRIPTION OF THE DRAWINGS FIGS. 1, 2, 3, and 4, taken together is a block diagram of a I DETAILED DESCRIPTION OF THE DRAWINGS That portion of the preferred embodiment shown in FIG. 1
operates to analyze the Korotkoff sounds, classify these sounds into acceptable and loud sounds, accumulate characteristic measures of the acceptable and loud sounds over a sampling period, and control the sampling period. A typical microphone, now shown, adapted for use in detecting K'orotkoff sounds may be placed on the organism such as on the brachial artery of a human. As is well known, the microphone converts the sounds into electrical impulses which are applied through an amplifier 101 and a signal processor 102 to a pair of voltage comparators 103 and 104. The specific form of the signal processor 102 is not critical to the present invention but it is designed for eliminating signals which could not result from Korotkoff sounds. For example, the signal processor could be a filter which filters out low frequency signals or it may be adapted to count the zero crossovers between the maximum and minimum peaks of any given signal. For example, in one specific embodiment acceptable sounds were defined by the following criteria: (1) the sound must be between 200 .and 2,000 digital converter unitsabove the background noise; (2) the sound must be within plus or minus 42 milliseconds of a predetermined average lag time after the occurence of a QRS complex in an EKG waveform; (3) the signal resulting from the sound must have one to three zero crossings between its maximal positive and maximal negative peaks. The zero crossings could be detected by the signal processor 102 thereby eliminating any signals not having the proper number of zero crossings. As will be more fully understood below, the amplitude limits and the time window mentioned above are controlled by the apparatus of FIG. 1. For related sets of conditions, see application Ser. No. 608,130, filed Jan. 9, l967 now Pat. No. 3,450,131, by J. R. Vogt and assigned to the assignee herein.
Each of the voltage comparator'units 103 and 104 has two inputs and two outputs. To one input of each of the voltage comparator units there is applied the output from the signal processor 102. A reference voltage V, is applied to the other input of comparator 103 and a reference voltage V is applied to the voltage comparator 104. The output terminals of the voltage comparators are indicated by the designations 1 and 4 and these outputs are always complementary. 1n comparator 103, there is an'output at the 1 terminal whenever the input from the signal processor is greater than the reference voltage V There is an output at the 1 terminal of voltage comparator 104 if the signal from signal processor 102 is greater than the reference voltage V whereas there will be an output from the (1 terminal of comparator 104 when the signal from the signal processor 102 is equal to or less than the reference voltage V The reference voltages V, and V;, are selected as the limits of acceptable sounds. That is, a signal which is greater than V but less than V is an acceptable signal. When such a signal occurs, the 1 output of comparator 103 and the 1 output of comparator 104 will be energized."
The latter two output terminals are connected to an AND gate 105. The '1 output terminal from comparator 104, in-
dicating a loud" sound is applied to AND gate 106. The
thereby creating electrical signals corresponding to heartbeats. vThe heartbeat signals are amplified in amplifier 111, processed in signal processor 112, and applied to the delay means 113. The signal processor 112 may be a filter for filtering out undesirable signals. For a pulse coming out of signal processor 102 to be considered an acceptable Korotkoff pulse, it should occur within a certain time window (lag time) after the QRS complex of the electrocardiogram is detected by the EKG electrodes and associated apparatus. (The QRS complex is a portion of an EKG waveform well known in the art; see e.g., Dorlunds Illustrated Medical Dictionary, 23rd Ed. (1957), page 433.) This time lag is provided for by the delay means 113 which delays the "heartbeat" signal and applies it to the AND gates 105 and 106. As pointed out above, in a specific example, the time window may be plus or minus 42 milliseconds. 1
An output from AND gate 105 means that an acceptable Korotkoff sound has occurred, whereas an output from AND gate 106 means that a loud Korotkoff sound has occurred. The acceptable and loud Korotkoff sounds are counted in a pair of counters 107 and 108, respectively. The counters accumulate the acceptable and loud sounds during the sample period which may be a predetermined number'of seconds, or as in the case of the embodiment illustrated, may be a predetermined number of heartbeats. A typical example would be for the sample period to be five heartbeats in duration, thereby allowing a maximum of five Korotkoff sounds during a sample period. The sample period is controlled by a preset counter 114 and a single shot 115. The preset counter 114 counts the heartbeats out of the signal processor 112. After a predetermined number of heartbeats has been counted, the single shot 115 is energized thereby providing a pulse output which is applied to a pair of AND gates 109 and 110. When the AND gates 109 and 110 are energized, the contents of counters 107 and 108 are passed therethrough'to-the analyzing and decisional logic system which is shown in FIG. 2. I
The apparatus shown in FIG. 2 carries out the decisional logic illustrated in FIG. 5. This decisional logic is designed to increase or decrease the pressure in the occluding cuff 601 (FIG. 4) in predetermined small increments to cause the pressure in the cuff to substantially follow the systolic or diastolic pressure, depending upon which is being measured. The decision is based entirely upon the number of acceptable sounds and loud sounds detected during the sampling period. It will be noted that during any sampling period the pressure in the occluding cuff 601 is held constant.
A few examples will serve to illustrate the decisional logic of FIG. 5. Referring to row 1 of the table, if the system is measuring the systolic blood pressure and there are no acceptable sounds and no loud sounds detected during the sample period, this probably means that the cuff pressure is above the systolic blood pressure and therefore a down signal is to be generated. If the diastolic blood pressure is being measured then the absence of both acceptable and loud signals indicates that the cuff pressure is probably below the diastolic blood pressure and therefore an up signal should be generated. Looking now at the third row in FIG. 5, if there are no acceptable sounds detected during the sampling period and there are more than one loud sounds detected during the sample period, this probably means that the cuff pressure is in between the systolic and diastolic blood pressures because that is where the Korotkofi sounds are greatest. Consequently, if systolic blood pressure is being measured, an up" signal should be generated whereas if diastolic pressure is being measured, a down" signal should be generated. The word continue" in rows 5 and of the columns headed systolic and diastolic means that if the conditions of acceptable and loud sounds are ,met, the command signal should be the same as the command signal generated from the prior sampling period. In other words, if the last command signal was up, then the present command signal should also be up provided the conditions indicated in row 5 or 10 of FIG. 5 are met. v
The apparatus for perfonning the decisional logic of FIG. 5 is shown in FIG; 2 and comprises a pair of digital comparators 201 and 202, latches 203--208, and 226, AND gates 209- 217, 222-225, and 227228, and OR gates 218221, and 229-231. All of the elements operate in the conventional manner well known in the art. Digital comparator 201 receives the number of loud sounds fromcounter 108 in FIG. 1. The digital comparator 201 has three outputs indicating respectively no loud sounds, one loud sound, and more than one loud sound. These outputs areapplied respectively to the set input terminals of latches 203. 205, and 204. If any of the latches are set, an output appears at its upper output terminal. The digital comparator 202 is the same as digital comparator 201 and it receives as its input the number of acceptable" sounds from counter 107. The outputs from digital comparator 202 are applied to the set input terminals of latches 206, 207, and 208. The only remaining inputs to the apparatus of FIG. 2 are the diastolic and systolic inputs to AND gates 222- 225. The diastolic input is energized when the diastolic blood pressure is being measured, whereas the systolic input is energized when the systolic blood pressure is being measured. The latter two logic inputs are derived from the apparatus shown in FIG. 3 as will be explained more fully hereafter. The total logic of FIG. 2 is self-explanatory but a couple of examples will be described to demonstrate 'the' correspondence between the apparatus shown and the decisional logic of FIG. 5.
For the first example, assume that during the sampling period there are no acceptable sounds and there are no loud sounds. Digital comparator 202 will provide an output which sets latch 206 and digital comparator 201 willprovide an output which sets latch 203. The outputs from latches 203 and 206 are applied to the AND gate 209 whose output passes through OR gate 219 to AND gates 222 and 223. If the systolic logic signal is energized, the output from AND gate 222 passes through OR gate 229, thereby causing a down command signal to be generated. If the diastolic logic input is energized, the output from AND gate 223 will, pass through OR gate 230 thereby causing an up command signal to be generated. These command signals correspond to the ones listed in the first row of FIG. 5.
As a second example, consider row 5 of FIG. 5 wherein there is one acceptable sound and one loud sound. The digital comparators 201 and 202 will set latches 205 and 208, respective ly.- The outputsfrom latches 205 and 208 are applied to AND gate 216 whose output is applied through OR gate 221 to the AND gates 227 and 228. The other inputs to AND gates 227 and 228 are from the l and I outputs of latch 226. Latch 226 is set by a down command signal and reset by an up command signal. Consequently, if the previous command signal was a down command signal, the latch will be reset and the AND gate 227 will be fully energized. The output from AND gate 227 passes through OR gate 229 to generate another down command signal. If the prior command signal had been an up command signal, the AND gate 228 will be fully energized thereby providing an output which'passes through OR gate 230 to generate another up command signal. Thus, the apparatus shown generates the proper command signals as required by the decisional logic of table of FIG. 5.
It will be noted that both the up and down command signals pass through OR gate 231 and are applied to reset counters 107, 108, and 114 in FIG. 1. The resetting of the latter counters starts a new sampling period.
The apparatus of FIG. 2 also providesan output from OR gate 218 which is applied to the apparatus of FIG. 3. The output from OR gate 218, labeled 2E, occurs whenever at least one acceptable sound is detected during the sample period. This is accomplished by connecting the one and more than one outputs from digital comparator 202 to the set input ter- -minals of latches 208 and 207, respectively, and connecting the outputs from latches 207 and 208 to the OR gate 218. Consequently, whenever a pulse appears at the output of OR gate 218, it indicates that at least one acceptable sound has been detected during that sample period.
The apparatus shown in FIG. 3 operates to generate a digital cuff pressure signal, decrease or increase the cuff pressure signal in response to the propercom mand, and average all of the cuff pressure signals which cause at least one acceptable Korotkoff sound to occur. The latch 401 is provided with the start signals which are indicated respectively as a systolic and diastolic start signals. The apparatus may be controlled by external means (not shown) so that these signals are alternated periodically to thereby cause the overall system to alternately monitor and track the systolic blood pressure and the diastolic blood pressure. When the latch 40] is set, indicating that systolic blood pressure is being monitored, the output therefrom is applied in the fonn of a logic signal to the AND gates 222 and 224 of FIG. 2. The output registers 503 and 507 contain the averages for the diastolic and systolic blood pressures previously obtained. The systolic start signal energizes the bank of AND gates 402 to pass the contents of register 507 through the bank of OR gates 308 to the register 309 thereby inserting into register 309 a digital number corresponding to the prior average systolic blood pressure. The bank of AND gates 403 is energized by the diastolic start signal to insert the prior average diastolic blood pressure into register 309.
The digital contents of register 309 is converted into an analogue quantity by the digital-to-analog converter 310 and the output therefrom is applied to the occluding cuff servo (FIG. 6) for controlling the pressure of the occluding cuff. Thus, the pressure in the cuff corresponds to the digital number contained in register 309. The pressure is increased or decreased by small incremental amounts in response to the up and down command signals respectively. This is accomplished by digital subtractor 304, digital adder 305, preset register 303, and AND gates 306 and 307. When an up command signal occurs, it is applied to a delay signal shot 301 which energizes the bank of AND gate 306. The other inputs to AND gate bank 306 is from digital subtractor 304 whichoperates to subtract the contents of the preset register 303 from the contents of the pressure register 309. The value held in preset register 303 may be equivalent to an incremental pressure having a constant magnitude of 4 millimeters of mercury as a specific example. Thus, the new digital value inserted into register 309 amounts to a decrease corresponding to 4 millimeters of mercury. When an up command signal is received, the delay single shot 302 is energized thereby energizing the bank of AND gates 307 to pass the output from adder 305 into register 309. The digital adder 305 operates to add the incremental amount from register 303 to the amount held in register 309 thereby increasing the pressure by the incremental value of 4 millimeters of mercury.
Each digital quantity held in register 309 is a measure of the occluding cuff pressure for the instant sample period. An accurate reading of the systolic or diastolic blood pressure is obtained by averaging the cuff pressures over a plurality of sample periods. However, only those cuff. pressures are averages which result in at least one acceptable Korotkoff sound. The averaging is controlled by a counter 502, adders 506 and 510, registers 505 and 509, dividers 504 and 508, and output registers 503 and 507. The apparatus for obtaining the average systolic pressure is the same as that for obtaining the average diastolic pressure so only the apparatus for obtaining the systolic pressure will be explained. Assuming that systolic pressure is being monitored and followed by the system, latch 40! will provide one input to the bank of AND gates 404. If the present cuff pressure, which corresponds to the digital number held in register 309, results in at least one acceptable Kortkoff sound an input pulse 2E, from OR gate 218 in FIG. 2, will fully energize the bank of AND gates 404 thereby passing the contents of register 309 into the adder 510 of the systolic averaging circuitry. The pulse 2E also is applied to the counter 502 which contains a number corresponding to the number of sample periods during which at least one acceptable Korotkoff sound is detected. A second input to the adder 510 is from the register 509, and the output from the adder 510 is entered into register 509. As a result of the adder-register combination, the sum of the cuff pressures is placed in register 509 and is updated for each new cuff pressure that results in a sound pattern containing at least one acceptable Korotkoff sound. The sum in register 509 is divided in divider 508 by the number of sample periods during which acceptable Korotkoff sound was detected. This is accomplished by applying the contents of counter 502 to the divider 508. The output from divider 508 is applied to output register 507 and represents. in digital form, the systolic pressure of the patient on whom the occluding cuff is placed.
The occluding cuff and the analogue control mechanism for inflating and deflating the occluding cuff is illustrated generally in FIG-4. The occluding cuff is illustrated as an arm cuff 601 having air duct connections to an inflate valve 602 and a deflate valve 603. The valves 602 and 603 may be solenoid controlled valves which operate in response to signals from a conventional analogue servodevice 605. The actual pressure in arm cuff 601 is measured by a conventional strain gauge 604 whose output, representing in electrical form the cuff pressure, is applied to the pressure servo 605. The other input to the pressure servo 605 comes from the digital to analogue converter 310 of FIG. 3. The servo operates in a conventional manner to maintain a zero difference between its two input signals. Consequently, if there is an incremental increase in the cuff pressure signal-3A from the converter 310, the pressure servo will energize solenoid 602 long enough to bring the strain gauge output up to the level of the cuff pressure input signal.
1. A method for tracking a blood pressure of an animal organism, comprising the steps of:
a. applying a first pressure to a member of said organism;
b. sensing a selected parameter of Korotkoff sounds resulting from'said first pressure;
c. placing each of said sounds into at least one of a plurality of groups based upon specified ranges of said parameter;
d. generating, for a determinate time interval, a characteristic measure for each of a plurality of said groups;
e. combining said plurality of measures with each other to form a pressure-correction representation; and
f. applying a second pressure to said member in response to said representation.
2. The method of claim 1 wherein said parameter is the intensity of each of said sounds.
3. The method of claim-2 wherein each of said measures indicates the number of said sounds placed within its associated group during said interval.
4. The method of claim 3 wherein said pressure-correction representation is further formed in response to the selection of one of a plurality of blood pressure types to be tracked.
5. The method of claim 3 wherein said pressure-correction representation is further formed in response to a pressure-correction representation formed for a preceding time interval.
6. The method of claim 3 wherein a first of said groups has an intensity range extending from a first threshold to a second threshold, and wherein a second of said groups has an intensity range exceeding said second threshold.
7. The method of claim 6 wherein each of said measures as sumes a first value when said number is less than unity, a second value when said number is unity, and a third value when said number is greater than unity.
8. The method of claim 7 wherein said pressure-correction representation has a first magnitude when the measure of said second group assumes said first value, and when the measure of said second group assumes said second value unless the measure of said first group also assumes said second value.
9. The method of claim 7 wherein said representation has a second magnitude when the measure of said second group assumes said third value unless the measure of said first group also assumes said third value.
10. The method of claim 7 wherein said representation has the same magnitude as that of a pressure-correction representation formed for a preceding interval when said first and second measures both assume saidsecond value, and when said first and second measures both assume said third value.
11. The method of claim 8 wherein said first magnitude is positive and said second magnitude is negative for a first blood pressure type, and wherein said first magnitude is negative and said second magnitude is positive for a second blood pressure type.
12. The method of claim 2 wherein the duration of said time interval is established by counting a predetermined number of heartbeats of said organism.
13. The method of claim 2 wherein said interval is delayed with respect to the time of occurrence of said heartbeats.
14. The method of claim 2 wherein said pressure-correction representation has a substantially constant absolute magnitude and a variable sign.
15. The method of claim 2 wherein said second pressure is applied after a predetermined time lag from the time at which said representation is formed.
16. The method of claim 2 further comprising the step of generating blood pressure indication from at least one of said applied pressures.
17. The method of claim 16 wherein said indication substantially equals said first applied pressure.
18. The method of claim 16 wherein said indication substantially equals said second applied pressure.
19. The method of claim 16 further comprising the step of detecting the existence of a preselected pattern of said measures, said blood pressure indication being made substantially equal to said second applied pressure when said pattern is detected during said interval.
20. The method of claim 19 wherein said indication is made substantially equal to said first applied pressure when said pattern is not detected during said interval.
21. The method of claim 19 wherein said pattern represents the presence of at least one of said sounds within at least one of said groups during said interval.
22. A method of tracking a blood pressure in an animal organism comprising the steps of:
a. applying constant pressure for a predetermined sample period to said organism;
b. detecting the Korotkoff sounds during said period resulting from said applied pressure;
c. analyzing said sounds on the basis of number and intensity to determine a probable direction of deviation between said constant pressure and the blood pressure being tracked; and
d. altering said applied pressure by a predetermined incremental amount in a direction opposite to said probable direction of deviation.
23. The method as claimed in claim 22 further comprising the steps of:
a. repeating (a) through (d) of claim 22 for a plurality of sample periods, said pressure being altered for each sample period; and
b. averaging all said constant pressures which cause at least one Korotkoff sound which meets a predetermined set of criteria.
24. The method as claimed in claim 23 wherein the step of analyzing comprises:
a. classifying said sounds into classes of acceptable and loud sounds according to their intensities relative to a predetermined set of criteria;
b. counting the loud and acceptable sounds occurring during a given sample period; and
c. generating an up or down pressure command signal dependent upon the number of loud and acceptable sounds occurring during said sample period and the particular blood pressure being tracked.
25. The method as claimed in claim 24 further comprising:
a. detecting the heartbeats of said organism; and
b. controlling the duration of said sample periods in accordance with a preset number of detected heartbeats.
26.. The method as claimed in claim 25 wherein the blood pressure being tracked is the systolic blood pressure.
27. The method as claimed in claim 25 wherein the blood pressure being tracked is the diastolic blood ressure.
28. Apparatus for automatically tracking t e blood pressure of a human organism comprising:
a. pressure applying means adapted to be connected to said organism for applying pressure thereto;
b. detecting means for detecting Korotkofi" sounds resulting from certain pressure levels being applied to said organism; and x c. means for controlling incremental up and down variations in pressure applied by said pressure applying means in accordance with-the number and a selected parameter of Korotkoff sounds detected over a predetermined sample period.
29. Apparatus as claimed in claim 28 wherein said means for controlling incremental up and down variation comprises:
a. voltage comparator means responsive to said detected Korotkoff sounds for generating two sets of signals corresponding to acceptable and loud sounds, respectively, whereby the classification ofthe sounds into acceptable and loud is a function of the sound intensity;
b. means for accumulating each of said two sets of signals for said predetermined sample period;
c. decisional logic means responsive to said accumulated two sets of signals for generating one of two pressure change command signals; and
d. means responsive to said generated command signal for changing by a predetermined incremental amount the pressure in said pressure applying means.
30. Apparatus as claimed in claim 29 wherein said two command signals are up and down command signals and said means for changing the pressure comprises;
a. a pressure servosystem connected to said pressure applying means for controlling the pressure in said pressure applying means in accordance with an analogue electrical quantity applied to a control input thereof;
b. a digital storage means for storing a digital quantity;
c. means responsive to said command up signal for increasing by an incremental amount the digital quantity stored in said digital storage means;
d. means responsive to said command down signal for decreasing by said incremental amount the digital quantity stored in said digital storage means; and
e. digital to analogue converter connected between said digital storage means and said pressure servo means for converting the digital quantity stored in said storage means into an analogue electrical quantity and applying said analogue electrical quantity to said control input.