NO790496L - METHOD AND DEVICE FOR DEMODULATING SIGNALS IN A BURGING LOGGING SYSTEM - Google Patents

Description

"Method and apparatus for demodulating signals in a system for logging while drilling"

The present invention relates to communication systems and more particularly to an apparatus and a method for receiving and interpreting data signals which have been transmitted to the earth's surface in a system for logging during drilling.

Logging during drilling entails the transmission to the ground surface of measurements taken down the borehole while drilling is in progress, rather than the measurements which are usually carried out using instruments mounted just behind the drill bit. It is a clear advantage to be able to provide information continuously during drilling with the entire string in place. Nevertheless, systems for logging while drilling have not achieved widespread commercial recognition, mainly due to problems associated with transmitting the measurement information through the noisy and unfavorable environment of a borehole. Various proposals have been made for achieving the transfer of measurement information to the earth's surface. One proposed technique, for example, involves transmitting log measurements using insulated electrical conductors that extend through the drill string. However, this method requires adaptation of the drill string pipes > including the provision of electrical connections at the pipe connections. Another proposed method uses an acoustic wave that propagates upward through the metal drill string#but the high levels of disturbing noise in a drill string^ is a problem with this technique. Another technique that seems particularly promising uses a drilling fluid in the borehole as a transmission medium for acoustic waves modulated with the measurement information. Drilling fluid or "mud" is typically circulated downward through the drill string and drill bit and upward through the annular space defined by the portion of the borehole surrounding the drill string. This is done conventionally to remove "drilling chips" and to maintain a desired hydrostatic pressure in the borehole. In the technique referred to# there is an acoustic transmitter down the borehole, known as a rotary valve,., or "mud siren" 1 which repeatedly interrupts the flow of drilling fluid, and this causes an acoustic carrier wave signal to be produced in the drilling fluid at a frequency which depends on the speed of the interruptions. The acoustic carrier wave is modulated as a function of digital logging data down the borehole. In a phase shift keying ("PSK") modulation technique, the acoustic carrier wave is modulated between two (or more) phase states. Different coding methods are possible when using PSK modulation. In a non-return to zero coding method, a change in phase represents a particular binary state (for example, a logic "1") while the absence of a phase change represents the other binary state ( for example a logical "0"). The phase changes are achieved mechanically by temporarily modifying the interruption frequency of the mud siren to a higher or lower frequency until a desired phase shift (positive or negative) is achieved, and then returning the mud siren to its nominal frequency. For example, if the nominal frequency of the mud siren is 12 Hz, a phase rotation of 180°\ can be achieved by temporarily

lower the frequency of the mud siren to 8 Hz for 125 milliseconds

(which is one period at 8 Hz and one and a half periods at 12Hz)

and then return the mud siren frequency to 12 Hz. It is easy to

see that a phase turn of 180° can also be achieved by transiently increasing the mud siren frequency for a suitable period of time (ie to achieve a desired advanced phase) and then returning it to the nominal frequency.

In conventional communications using phase shift keying, the carrier wave phase is conventionally changed in alternating directions (i.e. alternating between negative and positive phase rotation) so that the net change in the carrier wave phase over a long period of time is close to zero. In a system for logging during drilling where an electromechanical device/such as a mud siren is used to apply acoustic waves to the drilling fluid, it is preferred to carry out all phase changes in the same direction (ie all positive or all negative), which means that the technique for operation of the sludge siren becomes more efficient and simple. (If, for example, all the phase changes are achieved by transiently lowering the frequency, it is never necessary to increase the frequency above the nominal frequency, and the mud siren needs less driving power. The control circuits can also be less complex.) The term "one-way" PSK modulation means

this type of modulation where all phase changes are in the same direction.

The modulated acoustic signal is received at the top of the hole by means of one or more transducers which convert the acoustic signal into an electrical signal. It is therefore necessary to recover the digital information contained in the modulation of the received signal. This is briefly explained by first processing the received signals to extract the carrier signal. The reconstructed carrier wave is then used to synchronously demodulate the modeled electrical signal.

In the conventional type of system described, a bandpass filter is typically used at the receiver, the filter having a bandpass spectrum centered at the nominal carrier frequency and being used to detect the modulated carrier wave. However, as part of the present invention, it has been discovered that using a filter centered at the nominal carrier frequency results in less than optimal performance. In particular, the one-way nature of the modulation results in the average carrier frequency being different from the nominal carrier frequency. It has also been realized that a further problem arises in using conventional filters in phase shift keying systems of the type described. A typical conventional filter construction seeks to achieve a symmetrical spectral characteristic around the filter's center frequency. However, the one-way nature of the modulation results in a symmetrical filter characteristic that is less than optimally adapted to the frequency characteristic of the transmitted signal.

It is an object and aspect of the present invention to provide an improved filter for use in detection in a phase shift keying transmission system of the type where the modulation is achieved by transient one-way modification of the carrier frequency.

In the known type of system described, a convergence loop is usually used for the carrier wave at the receiver, the purpose of the convergence loop being to lock it onto the carrier wave of the received signals and to produce time signals that can be used in the demodulation process. It is desirable to achieve a lock to the carrier wave as quickly as possible to avoid possible loss of information. It is also desirable, once locking has been achieved, to have a confluence loop that is relatively stable, i.e. that it is not adversely affected by short-term error signal components in the loop at different frequencies. These two purposes are somewhat contradictory, since relatively fast achievement of locking requires a relatively large loop width, while loop stability will usually require a relatively small loop width. It is known that the loop width can be varied manually as soon as locking is achieved, but this technique is not particularly appropriate. In the type of apparatus for logging during drilling described above, and where relatively low-frequency acoustic signals are used, practical problems also arise when attempting to vary the loop width of the confluence loop for the carrier frequency. In particular, the variation of the loop width usually entails the connection of different capacitors in the loop filter circuit and a simultaneous modification of the loop's amplification factor. At the frequencies that are of interest-

the capacitors in the circuit usually have relatively large capacitances without the large dimensions that are typical for non-electrolytic capacitors. When a previously inactive capacitor is connected to the circuit, a problem arises due to the introduction of an offset voltage which is the result of the previous voltage across the new capacitor not matching the voltage applied to it as soon as it is connected. the circuit*..

It is an object and another aspect of the present invention to provide an improved carrier confluence loop with variable loop width which can overcome the problems of the prior art as mentioned above.

In the known type of systems described, the carrier wave is usually recovered by using a common loop circuit for the carrier wave. The convergence loop is a phase-locked loop comprising a voltage-controlled oscillator (VCO*) which responds to error signals resulting from the difference between the phase of the signal derived from the voltage-controlled oscillator and the phase of the carrier signal. In accordance with the present invention, however, it has been discovered by the one-way supply f the unit to. the phase modulation in the type of systems described above tends to cause a problem in operating the phase lock loop. Since in particular the phase changes are achieved by transient variation of the frequency, error pulses are produced in the phase-lock loop every time a data transition occurs. Since the PSK modulation is one-way (ie, transient frequency modification is always to a lower frequency or always to a higher frequency), these error pulses always have the same polarity. These error pulses can tend to cause frequency deviations in the confluence loop for the „ , ^ carrier wave.

It is a further aspect and object of the present invention to provide an improved carrier confluence loop for use in detection in a phase shift keying transmission system of the type where modulation is achieved by transient unidirectional modification of the carrier frequency.

One aspect of the present invention is directed to a method using a PSK signal which has been modulated with a digital information by transiently unidirectionally lowering or raising the nominal frequency of a carrier signal as a function of the digital information to effect a phase shift, comprising the following steps: filtering the modulated carrier signal

with a filter having a bandpass center frequency shifted from the nominal carrier frequency in the direction of said one-way lowering or raising of the frequency, and recovering the digital information from the filtered signal.

Another aspect of the present invention is directed to a method for utilizing a PSK signal that has been modulated with digital information by transient one-way

either lowering or raising the nominal frequency of a carrier signal as a function of the digital information to effect a phase change, comprising the following steps: filtering the modulated carrier signal with a filter having a bandpass characteristic which is asymmetrical in the direction of said one-way lowering or increasing the frequency, and recovering the digital information from the filtered signal.

A further aspect of the present invention is directed to a method for reducing transient signals resulting from the connection of a different capacitor in the circuit for use in an electronic system comprising: a pair of clamps, a plurality of capacitors, each having a of its plates connected to one of the terminals, coupling devices for connecting the other plate to a selected one of the capacitors of the second terminal, and devices with variable gain synchronized with the coupling devices for influencing the potential at the other terminal recharacterized by continuously producing a reference voltage associated each of a series of capacitors not currently connected between said pair of terminals, each produced reference voltage being a function of the voltage that would appear across its associated capacitor in the event that ^ this capacitor was instantaneously connected by the coupling devices between said pair of terminals and continuous true application of each produced reference voltage across the associated capacitor.

Yet another aspect of the invention is addressed

on a method of stabilizing a carrier confluence loop used in connection with an apparatus receiving a PSK signal modulated with digital information by transiently unidirectionally lowering or raising the nominal frequency of a carrier signal as a function of the digital the information to effect a phase change, said apparatus comprising a carrier lock-in loop for lock-in with the carrier of the received signal, said lock-in loop comprising a controlled oscillator with a control terminal, the frequency of the oscillator being determined by a signal supplying the control terminal, a comparator for generating of a control signal by comparing the phase of a signal derived from the received PSK-modulated signal with the phase of a signal derived from the output of the controlled oscillator, as well as means for supplying the control signal to the control terminal of the oscillator, which method comprises the following steps: generation of compensation pulses in response to transition ger in the received signal, and applying said compensation pulses to the control terminal to account for the difference between the nominal frequency of the carrier and the average frequency of the received signal resulting from the one-way nature of the carrier modulation.

The present invention also includes another aspect directed at an apparatus which receives a PSK signal modulated with digital information and which is effective for recovering the digital information from this, said PSK-modulated signal having been modulated with the digital information by transient unidirectional lowering or raising of the nominal frequency of a carrier signal as a function of the digital information to effect a phase change, the apparatus comprising a filter for use in selectively filtering the modulated carrier signal, characterized in that the filter has a bandpass center frequency which is shifted from the nominal carrier frequency in the direction of said one-way lowering or raising of the frequency.

The present invention comprises a further aspect directed to an apparatus for demodulating a digital signal and which receives a PSK signal modulated with digital information and _._.. recovers the digital information therefrom, which PSK modulated signal has been modulated with the the digital information by temporarily one-way lowering or increasing the nominal frequency of a carrier signal as a function of the digital information to effect a phase change, and then returning the carrier to the nominal frequency as soon as the phase change is completed, the apparatus comprising a filter for use at selective filtering of the modulated carrier wave signal, characterized in that the filter has a bandpass characteristic which is biased in the direction of said one-way lowering or raising of the frequency.

A further aspect of the present invention is directed to an electronic system comprising: a pair of clamps,

a series of capacitors each having one of its plates connected to one of the terminals, coupling means for coupling the second plate on a selected one of the capacitors to the second terminal, and variable gain means synchronized with the coupling device and affecting the potential of the second terminal, characterized wood circuit for reduction of transient signals originating from the connection of a different capacitor between said pair of terminals, comprising: Means for continuously producing one reference voltage in connection with each of the capacitors which have not yet been connected between the pair of terminals, each produced reference voltage being a function of the voltage that would appear across its associated capacitor in the event that the associated capacitor was instantaneously connected between the terminal pair by means of the coupling device, and means for continuously applying each produced reference voltage across its associated capacitor.

Another aspect of the present invention is directed to a variable loop width carrier confluence loop

for locking to the carrier of an input signal, comprising: a phase-locked loop with an oscillator having a control input, means for generating an error signal as a function of the phase difference between a signal derived from the oscillator and the input signal, and a variable filter having a variety of bandwidths for connecting the output of devices that produce an error signal to the control input, characterized by: a loop width control device connected to the variable filter to automatically change the loop width of the filter as a function of the input

the walking signal. Figure 1 is a simplified schematic diagram of an apparatus for logging during drilling which includes the present invention. Figure 2 comprises curves showing conventional PSK modulation and one-way PSK modulation used in the present invention. Figure 3 is a block diagram of the receiver system above the borehole in the apparatus of Figure 1. Figure 4 shows idealized waveforms useful in understanding the nature of the signals appearing at various locations in the receiver circuit of Figure 3. Figure 5 shows the nature of a phase change which

performed in accordance with the PSK modulation of Figure 2.

Figure 6 illustrates the nature of the frequency spectrum of a conventional PSK-modulated signal compared to the spectrum of a one-way PSK-modulated signal. Figure 7 shows an example of a filter that is useful in

the present invention.

Figure 8 is a block diagram of a phase lock loop with variable loop width in accordance with an embodiment of the invention.

Figure 9 shows a basic loop filter.

Figure 10 illustrates a variable loop width filter

according to an embodiment of the invention.

Figure 11 shows a previously known carrier wave confluence loop. Figure 12 shows an embodiment of an improved confluence loop in accordance with one aspect of the invention. Figure 13 illustrates the type of waveform that is produced and which is an input to the loop filter in the carrier confluence loop of Figure 11.

Reference is now made to Figure 1 where a simplified diagram of an apparatus for logging during drilling according to an embodiment of the present invention is shown used in connection with a conventional drilling apparatus. A platform and drawbar

10 is arranged over a borehole 11 which is formed in the ground by drill drilling. A drill string 12 is suspended in the borehole and comprises a drill bit 15 at the lower end. The drill string 12 and the drill bit 15 attached to it are turned by a turntable 16 (driven by means not shown) which engages with a driver rod 17 at the upper end of the drill string. The drill string hangs in a hook 18 attached to a runner block (not shown). The driver rod

is connected to the hook through one swivel 19 which allows rotation of the drill string in relation to the hook. Drilling fluid or mud 2 6 is held in a pit 27 in the ground. A pump 29 pumps the drilling fluid into the drill string via an opening in the swivel 19 so that it flows downwards through the drill string 12. The drilling fluid exits the drill string via openings in the drill bit 15 and is then circulated upwards in the area between the outside of the drill string and the periphery of the drill hole. As is well known in the field, the drilling fluid thereby carries with it drilling residues from the formation to the ground surface, and the drilling mud is returned to the pit 27 for recycling. The small arrows in Figure 1 illustrate the normal direction of drilling fluid flow.

Mounted inside the drill string 12, preferably near the drill bit 15, is a subsystem 50 for sensing and sending. The subsystem 50 comprises a measuring device 55 which can measure any desired condition downhole, for example specific resistance, gamma radiation, the weight of the drill bit, the cutting angle of the drill, etc. However, it will be understood that the measuring device 55 can be used to measure any useful downhole parameter in the hole. The transmitter part of the sub-system down the hole comprises an acoustic transmitter 56 which generates an acoustic signal in the drilling fluid which is representative of the measured conditions down the hole. A suitable type of acoustic transmitter, which is known in the art, uses a device known as a "mud siren" comprising a slotted stator and a slotted rotor which rotates and repeatedly interrupts the flow of drilling fluid to establish a desired acoustic wave signal in the drilling fluid. The transmitter 56 is controlled by control and drive electronics 57 which comprise analogue-to-digital circuits (A/D circuits) which transform the signals which are representative of the conditions down in the borehole into digital form. The control and drive electronics 57 also comprise a phase shift keying modulator (PSK modulator) which

provides drive signals for supply to the transmitter 56.

In conventional phase shift keying (PSK) communications, the phase of the carrier signal is changed in accordance with a digital data signal having two or more levels to provide a modulated carrier wave having two or more phases. The carrier wave phase is conventionally changed in alternating directions (that is, alternating in positive and negative directions) so that the net change in the carrier wave phase over a long period of time is close to zero. In a system for logging while drilling where an electromagnetic device such as a mud siren is used to add acoustic waves to the drilling fluid, it is preferred to carry out all phase changes in the same direction, which results in the technique of operating the mud sirens becoming more efficient and simple. The term "one-way" PSK modulation used here is intended to mean this type of modulation where all phase changes are in the same direction. Techniques for operating a mud siren to obtain a PSK-modulated acoustic carrier wave in drilling fluid, and to obtain one-way PSK modulation thereof, are described, for example, in US patents no. 3,789,355 and 3,820,063. However, one would understand that any suitable device may be used to achieve the types of unidirectional PSK modulation described herein. Figure 2 illustrates the difference between conventional PSK modulation and the one-way PSK modulation used in a system for logging while drilling. Curve 2A illustrates an unmodulated carrier signal with a period of T/4 where T is the bit period of the modulating

the information. An example of a bit pattern is shown in curve 2B, with transitions from "0" to "1" occurring at times 2T and 5T, and transitions from "1" to "0" at times T, 4T and 6T. If a conventional "differential coded PSK" coding technique is used, a phase change at the bit time slot (T, 2T, 3T, 4T...) indicates an "l" bit, while the absence of a phase change at the bit time slot indicates a "0" bit. However, it will be understood that the opposite convention can be used, or that any coding technique that can be adapted to the present invention can be used. Accordingly, on curve 2C where conventional PSK modulation is illustrated, a phase change of 0 is effected each time the next bit is "l" meaning that phase changes are effected at times 2T, 3T and 5T. Curve 2C thus shows phase changes that are effected at these times, which phase changes occur in alternating directions. Curve 2D illustrates the nature of the PSK modulation in a one-way PSK modulation used here. One sees that the phase changes are carried out in the same places, but in this example all phase changes are negative (ie they result in a sag of the phase) and one sees that the phase changes accumulate.

Reference is again made to Figure 1 where the generated acoustic wave (ie the primary component of this to be received) propagates upwards in the fluid through the center of the drill string at the speed of sound in the fluid. The acoustic wave is received at the earth's surface by means of transducers represented by reference number 31.

The transducers, which can for example be piezo-electric transdu..

received, converts the/acoustic signals into electrical signals. The output from the transducers 31 is connected to the receiver subsystem 100 on the surface which works to demodulate the transmitted signals and display the information measured down the hole on a display device and/or recording device 500.

Reference is now made to Figure 3 which shows a block diagram of the surface receiver system comprising an improved filter according to one aspect of the invention. The waveforms of Figure 4 showing an example of a "1011" bit pattern will be referred to from time to time to illustrate the operation. The acoustic signals in the borehole fluid are sensed by transducers 31 (figure 1) which in the present embodiment comprise transducers

31A and 31B. In the present embodiment, this transducer pair is used in connection with a differential detection arrangement comprising delay circuit 103 and differential amplifier 104. The output from transducer 31B is connected via buffer amplifier 102 and delay 103 to the negative input terminal of differential amplifier 104. Transducer 31A is connected via buffer amplifier 101 to the positive input terminal of differential amplifier 104. This differential detector arrangement . is used for the purpose of suppressing noise that propagates in the direction opposite to the primary acoustic carrier wave. If, for example, the distance between the transducers 31A and 31B is chosen equal to a quarter wavelength at the carrier frequency and the pre-sink 103 is also set to a quarter wavelength at the carrier frequency, acoustic waves propagating in the direction of the primary signal (arrow A) will undergo a phase lag of a total of half a wavelength. When the output from the delay circuit 103 is subtracted from the undelayed signal from transducer 31A, it is seen that the signals propagating in the direction of arrow A are added in phase. However, acoustic signals propagating in the opposite direction (arrow B) will result in inputs to the differential amplifier 104 being in phase, resulting in cancellation of these signals. This is easy to see when one realizes that in such a case the input to the positive terminal of the differential amplifier 104 is subjected to a delay of a quarter wavelength due to the distance between the transducers, while the input to the negative terminal of the differential amplifier 104 is subjected to a delay of a quarter wavelength due to the electrical delay circuit 103.

The output from the differential amplifier 104 is connected to

a bandpass filter 110 having a filter characteristic in accordance with the principles according to an aspect of the invention and as described in more detail below.

The output from filter 110 is connected to an amplifier

115 with automatic steering control (AGC) which is provided with a fast engagement and slow release characteristic. The fast engagement is useful when it comes to achieving stability and sync locking in minimal time, and the slow release maintains gain during transient signal loss or signal level changes. The output of AGC amplifier 115 (shown in idealized form on curve 4A) is connected to both a synchronous demodulator 130 and a carrier confluence loop 120 with variable loop width. The confluence loop 120 may comprise a phase locking loop and is an important aspect of the invention. The variable loop width can be regulated manually or automatically. With manual control, the carrier confluence loop will operate at a particular fixed loop width (for example wide, medium or narrow) in accordance with the operator's choice.

These loop widths can be, for example, 0.3 Hz, 0.1 Hz and 0.03

Hz so that a range from ten to one is covered. The wide or medium loop width will typically be used when seeking to achieve locking, and the narrow loop width will be engaged as soon as locking is achieved to strengthen the stability of the loop. In automatic regulation, the loop will initially achieve synchronization using the widest loop width (or the medium loop width if desired in certain circumstances). After synchronization is achieved, the loop width is switched to a narrower value. When a signal loss occurs as indicated by an output from a signal loss detector in circuit 120, the loop width is switched to the widest setting. In both manual and automatic operation, the carrier wave confluence loop with variable loop width can be provided with circuits for pre-charging certain capacitors which are switched in and out of operation when the loop widths are switched. This technique rned f precharge is advantageous to prevent possible loss of locking when, for example, switching to a narrower loop width, which can be caused by transient voltages•due to the voltages across the capacitors when they are connected to the circuits. Further details regarding the variable loop width confluence loop are given below.

As discussed in more detail below, the output from the carry-

the wave confluence loop with variable loop width derived from the output of a voltage-controlled oscillator in the circuit's phase-locked loop. This oscillator usually operates at a multiple of the nominal carrier frequency. A clock generator comprising a frequency divider therefore derives a clock signal from the output of this oscillator, the derived clock signal (illustrated by curve 4B) being at the carrier frequency and in a form suitable for use in demodulating the filtered input signal. The clock generator in circuit 120 may comprise clock correction circuits according to a further feature of the invention. As described in detail below, the unidirectional nature of the PSK modulated carrier signal results in a build-up of error signal components in the carrier feedforward loop. If this build-up of error component signals is not taken into account, using, for example, the clock correction circuit described below, it can cause an unwanted drive in the voltage-controlled oscillator in the common loop. This unwanted buildup of error components can be eliminated by producing offset pulses that tend to cancel out the error signals that would otherwise accumulate. Since the type of error signals considered here occur at each bit transition, the output of a bit transition detector 150 (to be described in more detail below) is used to regulate the generation of correction pulses.

The output of the carrier convergence loop 120 (curve 4B) is coupled to the synchronous demodulator 130 which, as noted above, receives as its second input the output of the AGC amplifier 115

to be demodulated. The synchronous demodulator can, for example, be ... an analog multiplication circuit. Its demodulated output is illustrated by curve 4C. The output from the synchronous demodulator 130 is connected

to a matched filter 140. The filter 140 is matched to a square pulse at the bit rate. As known in the area, a custom works

filter so that upon a data transition at its input, it integrates in a time interval equal to one bit period. Accordingly, at the end of each bit period, the output of the matched filter has an extreme positive or negative value (the waveform of curve 4D) at which sampling can be performed most efficiently. Sampling of the output from the matched filter 140 is performed by means of a sample and hold circuit 160 whose output is connected to an analog/digital converter 170 which produces a signal in digital form. (The output of the matched filter 140 is also coupled to the bit transition detector 150 which may include a zero-crossing detector that follows the zero-

passes on the output of the matched filter to produce output pulses having a phase synchronized with the bit transitions. Use of the transition detector output is referenced below. for.) The signal used to trigger sampling by the sample-hold circuit 160 and to define the conversion period of the analog-to-digital converter 170 is produced by a strobe signal detector 180. The sampling signal produced by the strobe generator (the waveform on curve 4F ), you see is at the bit or symbol frequency. To achieve this relatively accurate signal at the bit rate,

a symbol confluence loop 190 can be used in connection with the carrier wave. This symbol confluence loop is described in US Patent Application No. 684,604. Briefly described, the circuit 190 is a quadrature-type phase-locked loop comprising a voltage-controlled oscillator and a frequency divider in the loop. In this respect, the circuit is similar to a conventional bit synchronization circuit. As described in the above-mentioned US patent application, however, the output from the carrier confluence loop becomes a 120, in addition to receiving time information when a

transition is detected in the received signal (ie, the output of the bit transition detector 150 in Figure 3), also used to assist symbol concurrency loop 190 (the output illustrated by curve 4E) during the periods where the symbol transitions are absent. This is made possible by the coherent relationship between the carrier wave and the bithastigens. If after a number of bit periods there are no bit transitions, a signal derived from the carrier is used to maintain synchronization.

The output bit pattern from the analog/digital converter 170, for this example, is shown at curve 4G, and as can be seen is a result of sampling the matched filter output (curve 4D) with the strobe signal (curve 4F) and subsequent A/D conversion.

Since the data was originally encoded in conventional "differential coded PSK form" (as described above), a differential decoder 199 is used to recover the data in its original form. Since a change in phase indicated a "1" in the code, in particular, a bit change in the output of the A/D converter 170 (curve 4G) is interpreted as a "1" by the differential decoder 199. The absence of a bit change in the A/D converter output is conversely interpreted as a "0". Accordingly, and as known in the art, the differential encoder comprises an Exclusive-OR gate which acts on successively received bits and produces a "1" output when successive bits are different and a "0" output when successive bits are the same. The output from the differential encoder 199 is illustrated in Figure 4H for the present example.

It will be understood that on curve 4A of Figure 4, the PSK modulation was illustrated in idealized form, with "instantaneous" phase changes to facilitate understanding of the operation of the system of Figure 3. The actual phase changes are performed in the manner illustrated.in connection with curve 2D. Figure 5 illustrates such a phase change, caused by a temporary lowering of the carrier frequency until the desired phase rotation is achieved. The dashed line shows what the carrier waveform would look like without the frequency modification.

In one embodiment of the present invention, the carrier frequency is 12 Hz and the bit frequency is 1.5 Hz. Unidirectional PSK modulation is performed by transiently lowering the carrier frequency to 8 Hz until a phase shift of 180° is achieved, and then returning the carrier frequency to the nominal value of 12 Hz. (The desired sag is half the period of the nominal carrier frequency. A frequency of 8 Hz has a frequency that is lh times the period of

the nominal carrier frequency. Consequently, after a full period at 8 Hz (125 milliseconds), the desired phase lag will be achieved.

This is easy to see from Figure 5 where the waveform with the solid line changes to 8 Hz for a whole period, while the dotted waveform illustrates the continuation of a frequency of 12 Hz. However, since it takes a finite time to perform the change between the two frequencies and the average frequency is less than 12 Hz during the transition, the actual time at 8 Hz is slightly less than 125 milliseconds.) Techniques for operating the mud siren in this way are known, e.g. from them. ■ the above-mentioned US patents 3,789,355 and 3,820,063.

When the total receiver system has now been described, certain aspects of this according to the invention shall be described in more detail. According to one side of the invention, it is noted

ket that unidirectional phase rotation of the carrier wave means that the modulated signal spectrum is shifted in frequency from the nominal carrier frequency. The frequency shift, or shift, is followed by an asymmetry, or shift, in the spectrum. Figure 6B illustrates the nature of the one-way PSK frequency spectrum, and can be compared to the frequency spectrum of a conventional PSK modulated signal having the same nominal carrier frequency ' f o . The use of a bandpass filter (e.g., filter 110 in Figure 3) which takes into account

this shift and the asymmetry of the frequency spectrum for it__

modulated signal, is advantageous in that it more effectively separates the signal from the noise and minimizes the filter's signal distortion. The exact degree of spectrum shift and asymmetry depends on the data pattern of the modulation. An alternating pattern between "1" and "0" would, for example, result in a shift of the order of magnitude equal to the bit rate. Any other data pattern would result in an offset somewhat less than the bit rate.

If the data pattern is not known in advance (which is usually the case), one can assume that it is a random data pattern, and such a pattern results in an offset from the carrier frequency that is approximately half the bit frequency. For example, in the present embodiment where the carrier frequency is 12 Hz, the bit frequency is 1.5 Hz, and when the PSK modulation is achieved by one-way transient lowering of the frequency, the preferred filter center frequency

until the bandpass filter is at 11.25 Hz, i.e. the nominal carrier frequency minus half the bit frequency. (It will be understood that if the phase rotation is achieved by a one-way transient increase of the frequency, the shift will be towards higher frequencies and lie at 12.75 Hz in such a case.)

There are several ways to construct a filter that meets the principles of the present invention. The bandwidth of the filter is chosen to pass the modulated signal with the least possible distortion while suppressing unwanted noise and interference. The minimum bandwidth (-3 dB to -3 dB) for filtering in a PSK system is usually equal to the bit rate, although a somewhat larger bandwidth, such as 1.5 times the bit rate, is usually recommended. When constructing the band-pass filter, the following steps can be followed: first a prototype of a low-pass filter is selected and it is scaled to have a bandwidth equal to half of the desired bandwidth for the band-pass filter. The low-pass filter design is then transferred to a band-pass filter centered at a frequency offset from the carrier frequency in accordance with the rules mentioned above. The bandpass filter's transmission zeros are then chosen to produce the -desired filter symmetry-* (or asymmetry-) characteristics. A particular filter configuration is then selected and the values of the filter components are calculated. Since the details of how such a construction can be carried out are assumed to be known to those skilled in the art, for the sake of brevity such details shall not be given here.

An example of a filter can be realized by cascading two active biquadratic RC filter sections. A feedforward circuit configuration as described in "Design Formulas for Biquad Active Filters Using Three Operational Amplifiers", by Fleischer & Tow, Proe. of the IEEE, May 1973, may be used. The final filter can be composed of the two cascaded active biquadratic RC sections, as represented by the transfer functions of Figure 7, with

for the first biquadratic filter section. The construction formula of Fleischer & Tow can be used to calculate the values of filter-'■; the components. For example, for the first filter six ion, R8 and C, = C2 are chosen by the constructor

K^ = K~chosen by the designer.

So:

The foregoing is a non-limiting example of how to construct a bandpass filter that is usable in connection with the principles of the invention, and various alternative construction techniques can be used.

Reference is now made to Figure 8, where an embodiment of the carrier wave confluence loop 120 with variable bandwidth (Figure 3) in accordance with the invention is shown. A squaring circuit 201 receives the output signal from the AGC amplifier 115 (Figure 3), i.e. the filtered, gain-controlled PSK modulated signal. The squaring operation mainly serves to remove the modulation from the carrier wave, and in this process the frequency of the carrier wave is also doubled. The output from the squaring circuit 201 is an input to a phase detector 202. The other input to the phase detector 202 is the output from a frequency divider (or clock divider) 203. The output from the phase detector 202 is connected to a new filter 300 with variable bandwidth, which will be described in detail below. The output from filter 300 is connected to the voltage controlled oscillator (VCO) 204, and the output from the oscillator 204 is then connected to the clock divider 203.

The bandwidth of the filter 300 can be adjusted either manually or automatically under the control of the bandwidth control unit 205. In automatic operation, the control unit 205 receives the output from the signal loss detector 206. The signal loss detector 206 comprises a comparator which detects loss of locking in the loop by comparing the input signal (from the AGC -amplifier 115) with an adjustable threshold level. When the input signal is less than the threshold level, it indicates loss of lock. The bandwidth control unit 205 responds to a signal loss indication by effecting a bandwidth modification of the variable filter 300 to a larger bandwidth. When locking has again been achieved, or for example after a predetermined period of time when there is a high probability that locking has again been achieved, the control unit 205 effects a modification of the bandwidth of the variable filter 300 to a narrower bandwidth. In the case of manual regulation, the switching is under manual control by means of a switching device 205A.

The loop width (or bandwidth) of the phase-locked loop usually determines the acquisition time (or "lock time") of the loop and also the loop's stability, i.e. its ability to maintain lock to noisy input signals. As noted above, a larger loop width is advantageous in achieving fast locking, but once locking is achieved, the large bandwidth is disadvantageous in that it results in lower stability than a phase-locked loop having a smaller loop width. It is therefore advantageous to use a large bandwidth to achieve locking, and then switch to a narrower bandwidth after locking is achieved to improve loop stability. According to the present invention, modifications of the loop width can be performed automatically. An important feature of the invention prevents coupling between different bandwidths due to the introduction of bias voltages into the loop which would cause a loss of locking.

To better understand the invention, it is useful to first consider the basic loop width filter as illustrated in Figure 9. The output of the phase detector 202 (Figure 8) is an input to the positive input terminal of an operational amplifier 401. The negative input terminal of the operational amplifier 401 becomes ^ fed back from the output of the amplifier via a capacitor C. The output of the operational amplifier 401 is also connected, via a gain control resistor circuit 402 (shown in dashed lines), to the positive input terminal of another operational amplifier 405. The gain control circuit in this simplified embodiment comprises a series -resistance denoted and a resistance R, which is connected to the ground reference potential. The output from operational amplifier 405 is fed back to its negative input terminal. The output from the operational amplifier 405 is also connected via a voltage divider, which consists of series resistors 99R and R,

to the ground reference potential. The connection point between the resistors in the voltage divider is connected back to the negative input terminal of the operational amplifier 401. The transfer function of the loop width filter in Figure 9 is

The closed-loop transfer function, when integrated with the phase-locked loop of Figure 8, can be expressed as

where A is a gain factor less than or equal to one, as regulated by unit 402, and K is a loop gain constant that varies with the frequency of the voltage controlled oscillator. It is easy to show that the loop width can be changed, without affecting the loop's damping factor, if A and either R or C are varied inversely proportional to each other. Usually, A and C can be varied in discrete steps. As noted in the description introduction, however, switching the loop width during operation can result in loss of data due to the loss of locking due to an offset voltage in the loop width filter when the loop width is switched. Suppose, for example, that in figure 9 there exists a particular voltage across the capacitor'C in the loop filter. To change the loop width, another capacitor will usually be connected in the filter circuit (instead of C) and at the same time the gain factor of the loop filter will be changed. When this is done, a different voltage will be applied across the "new" capacitor. If the first voltage applied across the new capacitor does not have a suitable value,

the change in gain factor can result in an unwanted f.ei.l signal in the loop causing loss of locking.

Reference is now made to Figure 10 where an embodiment of a variable loop width filter is shown which includes a feature of the present invention whereby the capacitors are pre-charged to prevent loss of locking when switching to a different bandwidth. The operational amplifiers 401 and 405 and the resistors 99R and R are the same as in Figure 9. Resistor R^ in the gain control circuit A of Figure 9 is replaced by three resistors connected to ground through a three-pole portion 480A of a switch 480. Depending on the switch position, one of three resistors Rj_i*R12°^R13bubbled between the positive input terminal of amplifier 405 and the ground reference potential. The capacitors Cj_1/<C>12°^C13 can be visualized as a replacement for the capacitor C in figure 6. When operating the switch parts 480B, 480C and 480D of switch 480, it is seen that one of these capacitors is connected between the negative input terminal of operational amplifier 401 and a point which is a fixed voltage across the output of operational amplifier 401. This fixed voltage can be, for example, 5.1 volts using the zener diode 412 and current sources 415 and 416. The positions of the various parts of switch 480 in the embodiment of Fig. 10, is under joint management. The three positions of the switch are designated w (wide), m (medium) and n (narrow) denoting the available loop width settings of the circuit in this embodiment. The control of the switch can be either manual or automatically effected by the loop width control circuit 205 (figure 8). It can be seen that when the switch control is in the w- (wide) position, resistance R^ and capacitor C^^ are in the loop, when the switch control is in the m- (medium) position" resistance R.^0<3 is capacitor C, in the loop, and when the switch control is in the n (narrow) position, resistance°9capacitor C^3 is in the loop. At relatively low operating frequencies used in logging while drilling of the type described here, relatively high capacitance values are used . For example, C-^1»C^2 and C^^ can have values of 10, 33 and 100 microfarads respectively. In order to avoid excessively large physical capacitor dimensions, it is practical to use capacitors of the electrolyte type. These capacitors requires a bias voltage which is provided in the circuit of Figure 10 by the bias current diodes 415 and 416 and the zener diode 412. A filter capacitor 413 having a large value, such as 220 microfarads, is connected in parallel with the zener diode 412. The individual resistors R-q, R12 and R13 can have value-?” as infinite resistances respectively (open circuit), 3.86 kohm and 1 00 kohm and resistor 414 can have a value of 9 09 kohm.

Based on the part of the circuit in Figure 10 that has been described so far, it can be assumed that the variable loop width filter works at the wide loop width, that is, with resistance

(open circuit) and capacitor in the circuit. If the output of operational amplifier 401 is at a voltage V^, and since the input impedance of operational amplifier 405 is very high, the voltage at the input of operational amplifier 405 is also approximately V^.<*>

Now assume that the loop width switch control of switch 480 is switched to the medium loop width position. The resistor R^2 will now. form a voltage divider with resistor 414. Since R-^2 is only three-tenths of the total resistance of resistor 414 plus R^2, the voltage to operational amplifier 405 would drop to a value of about (0.3) V^. The output from operational amplifier 405 would therefore be instantly reduced to three-tenths of its previous value.

This jump can itself cause a loss of lock since the output of amplifier 405 is coupled to the loop oscillator VCO (Figure 5). The positive side of the capacitor C^2, which will be connected into the circuit, is 5.1 volts above voltage V, (as is the positive side of the capacitor, which will be disconnected from the circuit). To avoid a sudden jump at the output of amplifier 405, the original voltage across C^2 should be greater than the voltage across C^^ by a factor of 10/3. Accordingly, as described below, the present invention provides for pre-charging of those capacitors which are not currently active in the circuit. However, another consideration must be taken into account, which is as follows: Two signal components are usually present in the loop filter circuit, namely an AC signal component and a DC or very low frequency error voltage. Since the positive side of

all three capacitors C^-^, C^2and C,.~ are connected to a common point (ie 5.1 volts above the output voltage from operational amplifier 401), care must be taken not to precharge the inactive con-| densators (ie those temporarily out of the circuit) to a fixed gain times both components, since the AC component is a common signal that should remain the same regardless of the chosen loop width.

In the circuit of Figure 10, a voltage representative of the voltage across the capacitor currently in the circuit is applied to each of a series of gain control amplifiers - • 421, 423 and 425. In particular, the voltage which is 5.1 volts below the voltage of the positive side of the capacitor in the circuit, applied to the positive input terminal of each of these amplifiers 421, 423, and 425, and the voltage at the negative input terminal of operational amplifier 401 (which is also the voltage at the negative side of the capacitor that is currently in the circuit) is fed to the negative input terminal of each of amplifier legs 421, 423 and 425. Three additional sections of switch 480, designated 480E, 480P and 480G, may act to supply one of three gain control inputs to a gain control terminal of each of the respective amplifiers 421, 423 and 425. In the present embodiment, the gain control multipliers applied to amplifier 421 are for switch p the positions w, m and n, respectively 3.3, 1.0 and 0.33. The gain control multipliers that are supplied

the gain 425 for switch positions w, m and n is 10, 3.0 and 1.0 respectively. It will be understood that the gain control multipliers applied to the gain control amplifiers 421, 423 and 425, via the switch sections 480E, 480F and 480G, can be provided by any means known in the art, such as by connecting suitable weighing resistors (not shown) into voltage divider circuits for

... to achieve the desired gain multipliers.

The outputs of amplifiers 421, 423 and 425 are respectively connected to the negative input terminals of operational amplifiers 422, 424 and 426. The positive input terminals of these amplifiers are each connected to the output of operational amplifier 401, so that they each receive a signal which is 5 .1 volt below the voltage on the positive side of the capacitor currently in the circuit. The outputs of the amplifiers 422, 424 and 426 are respectively connected to two poles of the respective switch parts 480B, 480C and 480D. The three switch parts can be seen arranged so that the negative terminals of the capacitors currently inactive in the loop filter circuit are connected to the output of their respective amplifiers (422, 424 or 426). In particular, capacitor C-^

connected to the output of amplifier 422 for the m and n switch positions, capacitor C^2 is connected to the output of amplifier 424 for the w and n switch positions, and the capacitor is connected to the output of amplifier 426 for the w and m switch positions.

During operation it is seen that the switch 480 causes switching of the filter loop width by simultaneously switching on the appropriate gain factor (resistor R-q, Rj_2°9R13^ together with its corresponding capacitor (C"11# C^ 2 or c13). The switch parts 480B, 480C and 480D also serve to supply the desired precharge voltages to those capacitors not currently in the circuit, this is achieved by means of amplifiers 421 to 426. In particular, the positive terminals of these six amplifiers are connected to a potential of 5.1 volts below the voltage on the positive plates of each of the three capacitors C-^l/C- ±2°^C13"The ne9ative input terminal of amplifiers 422, 424 and 426 is connected to the potential on the negative plate of the special capacitor (cj_i what?

or C^^) which is currently in the circuit. Since the outputs of amplifiers 421, 423 and 425 are respectively connected to the negative input terminals of amplifiers 422, 424 and 426, it is seen that

the common AC signal component is canceled in the output signal from amplifiers 422, 424 and 426, and not applied as a precharge voltage.

An example of how it works is as follows: Assume once again that the circuit works at the wide loop width, that is, with R,, (open circuit) and capacitor in the circuit. As described

above, a switch to medium loop width will require an initial voltage across C^2 (the "new" capacitor in the circuit) which is 10/3 (=3.3) times the value that was applied across C^^ just before the switch,, It can be seen that in this situation a gain control actuator of 3.3 is supplied to amplifier 423 via switch part 480F.

If the switching was instead to the narrow loop width, the resistor Rj^ which is connected into the circuit will itself cause the input voltage of amplifier 505 to drop to 1/10 of its value just before the switching. Consequently, the gain control factor applied to amplifier 425 (and affects the precharge of capacitor C-|^ which will be connected in this situation) has a value of 10. The other gain control factors for amplifiers 421, 422 and 423 can also be easily seen to have the correct values for each situation.

A further feature of the improved carrier convergence loop (eg, block 120 of Figure 3) according to another aspect of the present invention will now be described. Figure 11 illustrates a conventional prior art circuit for a carrier converging loop. It will be seen that it corresponds to the circuit illustrated in figure 8, but without signal loss detector 206, loop width control 205A and the variable loop width filter 300. Accordingly, the same numbers are used for the same components. The modulated carrier wave is first squared in a squaring circuit 201 to destroy the modulation information on it. The output of the squaring circuit 201 is a signal at about twice the carrier frequency and is an input to a phase comparator 202. The output of the phase comparator is connected to a loop filter 203 whose output is in turn connected to a loop filter 203 whose output is in turn connected to the control input terminal of a voltage controlled oscillator (VCO) 204. The output of this oscillator is connected via a frequency divider (or clock divider) 205, to the other input of the phase comparator 202.

During operation and as is well known in the field, as soon as convergence or locking is achieved, the phase lock loop in Figure 11 remains locked on the carrier wave since the phase difference between the produced clock signals (output signal from clock parts 205) and the received carrier wave produces an error signal which tends to regulate the oscillator frequency to correct any sensed "errors". However, as noted in the introduction to the description, the one-way nature of the phase modulation in the type of system described here tends to cause a problem in operation of the phase-locked loop. In particular, since changes are introduced (at data transitions) by transient variation of the frequency (to a lower frequency in the present embodiment), error pulses are produced at the output of the phase comparator each time a data transition occurs. Since the PSK modulation is unidirectional (ie transient frequency modification is always to a lower frequency -- as here -- or always to a higher frequency) these error pulses always have the same polarity. The inventors have noted that these error pulses can tend to pull the carrier convergence loop off in frequency.

Figure 12 shows an improved circuit over a carrier convergence loop in which means are provided which respond to transitions in the received signal to compensate the signal applied to the control terminal of the voltage controlled oscillator to account for

the difference between the nominal frequency of the carrier wave and the actual average frequency of the received signal. In Figure 12, the squaring circuit, the phase comparator, the loop filter, the voltage controlled oscillator and the clock divider all have the same reference numbers as in Figure 11.

In the embodiment in Figure 12, the output from the phase comparator 202 is fed to the loop filter and the voltage-controlled oscillator via a summing circuit 210. The other input to the summing circuit 210 receives compensation pulses from a pulse generator 220, which can be a monostable multivibrator, and which is triggered by the output of a bit transition detector 150 (Figure 3) via conductor 222 and produces a short compensation pulse each time a data transition occurs. In this way, the effects of the previously described error pulses do not accumulate and cause a frequency drive in the phase-locking loop. Figure 13 shows the waveform that is output from the summing circuit 210. Error pulses 1, 2 and 3

■ occurring at data transitions is compensated for by the pulses 1', 2<*>and 3' produced by the pulse generator 220. The net input to the voltage controlled oscillator, which results from the frequency-modifying nature of the phase modulation, is therefore essentially zero. '

Claims (26)

1. Method using a PSK signal (phase shift keyed signal) that has been modulated with digital information by transiently one-way either lowering or increasing the nominal frequency of a carrier signal as a function of the digital information to effect a phase change, characterized by the following steps: filtering the modulated carrier signal with a filter having a bandpass s~. center frequency which is shifted from the nominal carrier frequency in the direction of said one-way lowering or raising of the frequency, and recovering the digital information from the filtered signal. 2. Method according to claim 1, characterized by displacement of the center frequency from the nominal carrier frequency of said digital information. 3. Method according to claim 2, characterized in that the amount of displacement is equal to half the bit frequency of said digital information. 4. Method according to claim 1, 2 or 3, characterized by biasing the bandpass characteristics of the filter in the same direction as said displacement. 5. Procedure when using a PSK signal that has been modulated with digital information by transiently one-way either lowering j ning or increasing the nominal frequency of a carrier wave sig- . nal as a function of the digital information, to effect a phase change, characterized by the following steps: filtering the modulated carrier signal with a filter having a bandpass characteristic biased in the direction of said one-way lowering or raising the frequency, and recovering the digital information from the filtered signal. 6. , Method for reducing transient signals resulting from connecting a different capacitor in a circuit for use in an electronic system comprising: a pair of terminals, several capacitors each having one of its plates connected to one. of the terminals, coupling devices for connecting the second plate of a selected capacitor of the aforementioned to the second terminal, and variable gain devices which are synchronized with the coupling devices and affect the potential of the second terminal, characterized by the following steps: continuously generating a reference voltage associated with each of said capacitors not currently connected between said terminals, each generated reference voltage being a function of the voltage that would appear across its associated capacitor in the event that said associated capacitor were instantaneously connected between the terminals by means of the coupling devices, and continuously applying each produced reference voltage across the associated capacitor. 7 . Method of stabilizing a carrier confluence loop used in connection with an apparatus receiving a PSK signal modulated with digital information by transiently unidirectionally either lowering or increasing the nominal frequency of a carrier signal as a function of the digital information for to effect a phase change, which apparatus comprises a carrier overlap loop for overlap with the carrier wave of the received signal, which carrier convergence loop comprises a controlled oscillator with a control terminal, the frequency of the oscillator being determined by a signal applied to said control terminal, a comparator for producing a control signal by comparing the phase of a signal derived from the received PSK modulated signal with the phase of a signal derived from the output of the controlled oscillator, as well as means for supplying the control signal to the control terminal of the oscillator, characterized by the following steps: generating compensation pulses in response to transitions in the received signal, and supplying the compensation pulses to said control terminal to account for the difference between the nominal frequency of the carrier signal and the average frequency of the received signal resulting from the one-way nature of the carrier modulation. 8. Apparatus which receives a PSK signal modulated with digital information and acts to recover the digital information from this, the PSK modulated signal having been modulated with the digital information by transiently unidirectionally either lowering or...^ increasing the nominal frequency of a carrier signal as a function of the digital information to effect a phase change, said apparatus comprising a filter for use in selectively filtering said modulated carrier signal, characterized by that the filter has a bandpass center frequency which is shifted from the nominal carrier frequency in the direction of said 'one-way lowering or raising of the frequency. 9. Apparatus according to claim 8, characterized in that the center frequency is shifted from the nominal carrier frequency by an amount which is a function of the bit frequency of the digital information. 10. Apparatus according to claim 9, characterized in that the center frequency is shifted from the nominal carrier frequency by an amount equal to half the bit frequency of said digital information. 11. Apparatus according to claim 8, 9 or 10, characterized in that the bandpass characteristic of the filter is biased in the same direction as said displacement. 12. Apparatus according to claim 8, wherein the apparatus comprises an apparatus for logging while drilling for obtaining subsurface measurements while drilling in a fluid-filled borehole and for transmitting the measurements to the earth's surface, comprising: a downhole arranged subsystem for sensing and sending comprising means mounted on a drill string for obtaining measurement information, means for generating acoustic carrier waves at a nominal frequency in the borehole fluid, means for PSK modulation of the generated acoustic carrier waves in accordance with digital data representing the measurements, by transient unidirectional either lowering or increasing of the frequency of the acoustic carrier signal, and a receiver subsystem on the surface comprising a filter and transducer means for converting the modulated acoustic carrier waves into electronic signals, and characterized in that the center frequency of the filter is shifted from the nominal carrier frequency by an amount that is a function of bit rate one to said digital information, and means for recovering the digital data from the filtered electronic signals. 13. Apparatus for demodulation of a digital signal which receives a PSK signal modulated with digital information and causes the recovery of the digital information from this, the PSK signal having been modulated with the digital information by transient one-way either lowering or increasing the nominal frequency of a carrier wave signal as a function of the digital information to effect a phase change, and then returning the carrier wave to the nominal frequency as soon as the phase change has been carried out, which apparatus comprises a filter for use in selectively filtering the modulated carrier wave signal, characterized in that the filter has a bandpass characteristic which is biased in the direction of said one-way lowering or raising of the frequency. 14. Electronic system comprising: a pair of terminals, an array of capacitors each having one of its plates connected to one of the terminals, coupling means for connecting the other plate of a selected capacitor to the other terminal, and variable gain means synchronized with the connecting devices and which affect the potential on the other terminal, characterized by a circuit for reducing transient signals which are the result of connecting a different capacitor between said pair of terminals, comprising: means for continuously producing a reference voltage in connection with said capacitors which are not connected between the pair of terminals, each produced reference voltage being a function of the voltage which would act across its associated capacitor in the event that said associated capacitor was instantaneously connected between the terminal pair of the coupling devices, and means for continuously applying each produced voltage across the associated capacitor.. 15. System according to claim 14, characterized in that said means for producing a reference voltage reacts to the voltage that is currently across the terminal pair and also reacts to a ratio between gain factors for the means with variable gain. 16. System according to claim 15, characterized in that the means for producing a reference voltage comprise several stronger devices, each of which reacts to the current voltage between the pair of clamps and is gain-controlled in accordance with a ratio between said gain factors. 17. System according to claim 14, 15 or 16, characterized in that the positive plate of each of the capacitors is connected to one of the terminals and that this one terminal is held at a positive voltage. 18. Carrier confluence loop with variable loop width for locking onto the carrier of an input signal, comprising: a phase locked loop with an oscillator having a control input, means for generating an error signal as a function of the phase difference between a signal derived from the oscillator and the input signal, and a variable filter with several different bandwidths for connecting the output of the means that produce error signals to the control input of the oscillator, characterized by loop width control means connected to the variable filter for automatically changing the filter's loop width as a function of the input signal. 19. Confluence loop according to claim 18, characterized in that the variable filter comprises several capacitors which can read into and out of the filter under control of the loop width control means, and wherein the variable filter comprises means for continuously pre-charging those of the capacitors not currently active in the filter to prevent loss of lock in the phase lock loop when switching is effected by the loop width control means. 20. Confluence loop according to claim 18, where the variable filter includes: a first amplifier having a first and a second input terminal, which first input terminal is arranged to receive an input signal, a second amplifier, variable gain control means for connecting the output of the first amplifier to an input of the second amplifier, the variable gain control means having at least first and second different gain factors, variable capacitance means comprising at least first and second capacitors of which is coupled in accordance with the variable gain control means for capacitively coupling the output of the first amplifier to the second input terminal of the first amplifier, the kerer (characterized by means for generating a reference voltage in connection with the capacitor which is currently inactive in the variable capacitance means, the produced reference voltage being a function of the voltage that would appear across the capacitor which is currently inactive if it were momentarily switched on, and means for continuously applying the produced reference voltage across the capacitor which is not currently active in the variable capacitance means. 21. Confluence loop according to claim 20, characterized in that the means for producing a reference voltage react to the voltage across the capacitor which is currently active in the variable capacitance means and also to a relationship between the amplification factors. 22. Convergence loop according to claim 20, characterized in that the means for generating a reference voltage comprise first and second amplifier devices respectively connected to said first and second capacitor, the amplifier device which is connected to the capacitor which is not currently active, reacts to the voltage above the currently active capacitor. 23. Confluence loop according to claim 22, characterized in that each of the amplifier devices is gain-controlled in accordance with a different ratio between gain factors. 2 4. Apparatus which receives a PSK signal modulated with digital information and which causes recovery of digital information from this, the PSK-modulated signal having been modulated with the digital information by transient one-way either lowering or increasing the nominal frequency to a carrier signal as a function of the digital information to effect a phase change, the apparatus comprising a carrier confluence loop comprising: a controlled oscillator with a control terminal, the frequency of the oscillator being determined by a signal applied to the control terminal, a comparator device for generating a control signal by comparing the phase of a signal derived from the received PSK modulated signal and the phase of a signal derived from the output of the controlled oscillator, means for supplying the control signal to the control terminal of the oscillator, characterized by means which react to transitions in the received signal to compensate the signal supplied to the control terminal. so that account is taken of the difference between the nominal frequency of the carrier and the average frequency of the received signal resulting from the one-way nature of the carrier modulation. 25. Circuit according to claim 24, characterized in that the compensating means comprise means for generating a pulse at each data transition in the received signal, and means for adding the generated pulses to the control signal. 26. Apparatus according to claim 24 comprising an apparatus for logging during drilling for obtaining subsoil measurements during drilling ring in a fluid-filled borehole and for transmitting the measurements to the earth's surface, comprising: a downhole subsystem for sensing and transmitting, including means mountable on a drill string for obtaining measurement information, means for generating acoustic carrier waves at a nominal frequency in the borehole fluid, means for PSK modulation of the generated acoustic carrier waves i ., .correspondence with digital data representative of the measurements, by transient one-way either lowering or increasing the frequency of the acoustic carrier wave signal, dg a receiver subsystem on the surface comprising: transducer means for converting the modulated acoustic carrier waves into an electronic input signal, and a carrier wave confluence loop comprising the controlled oscillator, the comparator device, the supply means and the compensating means, characterized in that the compensating means comprise means for generating a pulse at each data transition in the received signal, and means for adding the generated pulses to the control signals, as well as means for recovering the digital data by then nodulating the input signal with a signal derived from the output of the controlled oscillator.