GB2116388A

GB2116388A - Variable circuit characteristics

Info

Publication number: GB2116388A
Application number: GB08306589A
Authority: GB
Inventors: Keith Hutchings
Original assignee: COLE ELECTRONICS Ltd
Current assignee: COLE ELECTRONICS Ltd
Priority date: 1982-03-10
Filing date: 1983-03-10
Publication date: 1983-09-21

Abstract

The characteristics of circuits are varied by varying the mark/space ratio of a signal controlling a switch which switches components in and out, either stepwise or continuously variable control can be achieved. In the preferred embodiments frequency and/or phase characteristics of a network comprising resistors and/or capacitors connected in multi-pole or multi-stage passive or active circuits are varied or variable. In these circuits several switches operate in response to a single control signal thereby achieving accurate tracking between stages. <IMAGE>

Description

SPECIFICATION Variable circuit characteristics The invention relates to circuits especially but not exclusively filter circuits or the like, and more particularly to circuits the characteristics of which are variable. In its simplest form a circuit according to this invention may be equivalent to a variable resistor but a particular object of the invention is to provide a variable filter.

The invention arose in connection with the following problem: when analysing signals of a short duration, such as transient signals, a socalled transient recorder is used to convert the analogue signal into a digital form convenient for storage and subsequent and/or repeated analysis.

Analogue to digital conversion invoives sampling the analogue signal at intervals, and it will be understood that the sampling rate cannot be equal to or less than the frequency of the signal of interest. In fact, the minimum acceptable sampling rate is twice the frequency of interest if aliasing (i.e. generation of interfering frequencies) is to be avoided. For this reason an input filter is required to exclude frequencies outside the range of interest.

It is desirable to provide for the sampling rate of a transient recorder to be varied and if this is to be achieved, then it is necessary to change the cut-off frequency of the input filter. This can be done only by changing the filter by switching or by using a variable filter.

Variable filters are known but these tend to be complex with many components and in the case of multi-pole filters ganged potentiometers or the like are required, so ensuring that the resistances for the various filter poles are tracked together to vary the filter characteristics. It has been proposed to use an FET or the like connected in series with the resistance to enable continuous variation of the filter frequency of a single pole filter. This arrangement is, in practice, unsuitable for multi-pole filters due to the difficulty in ensuring that the FET's are "tracked" accurately together This invention aims to enable variation of selected characteristics of an electrical or electronic circuit, and in particular multi-pole or multi-stage networks, such as filters, amplifiers etc., and furthermore to achieve such variations by remote control.

In accordance with the present invention, we propose controlling the characteristics of a branched circuit, which may comprise active and/or passive components depending upon the characteristics to be varied, by varying the mark/space ratio of a signal controlling switching means operable to switch in and out a selected branch or branches of the circuit.

The invention also includes a circuit having variable characteristics and comprising one or more branches, means for generating a control signal having a variable mark/space ratio, and switching means operable to switch in and out a selected branch or branches of the circuit in response to the control signal.

In general the switch may be of any kind having a sufficiently fast response but we prefer to use an electronic switch and in many applications a bi-lateral switch is advantageous.

One important advantage of the present invention is that any number of stages or poles in a network can be controlled in response to a single control signal and since the branched circuits are simply switched in and out, in contrast to the previous proposal to vary effective resistance by means of series connected FET's, the need to ensure accurate tracking of their resistances, is entirely avoided.

Hence the invention may be applied, for example, to control the gain of a number of amplifiers connected either in tandem or in cascade and to passive or active multi-pole filter networks. In the particular case of the filter network of a transient recorder such as described above, or more generally, any device in which the (filter) characteristics are to be varied in response to variations in another signal (sampling rate), there is another advantage in that adjustment of the mark/space ratio of the control signal may be effected automatically.

A further advantage is that the control signal may be an analogue derived signal, for example having a continuously variable mark/space ratio or a true digital signal derived from clock pulses and combination logic circuits.

Embodiments of the invention will now be described by way of example and with reference to the accompanying drawings of which: Figure 1 shows schematically a simple single stage variable filter; Figure 2 is a circuit diagram of a 5-pole variable low-pass filter; Figure 3 and 4 are characteristic curves showing the frequency response of a 5-pole variable low-pass filter for different switch control signal mark/space ratios; Figure 5 is a circuit diagram of a variable band pass filter; and Figure 6 are characteristic curves showing the frequency response of the variable band pass filter of Figure 5, for different switch control signal markispace ratios.

The principle underlying this invention can be understood with reference to Figure 1. In this simple single pole filter a branch circuit containing a resistance R2 connected in series with a bilateral switch S1 is connected in parallel with another resistance R,. Capacitor C is connected between the output and earth. The control electrode of the switch S, is connected to a control signal the mark/space ratio of which is adjustable to vary the effective resistance in the R-C network as will be understood from the following: When the switch S, is closed the effective resistance of the circuit is R,R2/(R,+R2) and with the switch open this value becomes R,.At 50% mark/space ratio of the signal applied to the control electrode of the switch the effective resistance is R,(2R2)/(R1+2R2). Hence by varying the markispace ratio of the control signal, the resistance and consequently the cut-off frequency of the filter is scaled between the limits set by R, and R,R2/(R,+R2)- Any conventional filter circuit employing resistance-capacitance elements to control the amplitude and phase response may be converted to a variable filter circuit in accordance with the present invention, by applying the principle described above. Fine tuning for critical adjustment of the centre frequency of a notch or band pass filter is feasible or coarse adjustment over one or more decades in frequency may be achieved. Furthermore, circuit Q may be independently controlled using this technique.

This allows variable bandwidth and variable centre frequency filters to be realised.

As stated above the method of control may be analogue or digital so that, for example, a 5-pole low pass Chebyshev filter may be designed to have a cut-off frequency continuously variable from 5 kHz to 30 kHz controlled by a single potentiometer.

The cut-off frequency may be stepped between the same limits by for example, an 8-bit binary coded decimal signal giving 256 steps between minimum and maximum.

The pulse rate of the control signal should be higher than the maximum frequency of interest to prevent aliasing effects.

A practicai embodiment of variable filter circuit according to this invention is shown in Figure 2, the filter being a 5-pole low-pass Chebyshev filter, the cut-off frequency of which is adjustable between 5 kHz and 30 kHz. Each of the five poles includes a bilateral switch S1, S2, S3, S4 and S5 and a series connected resistor r1, r2, r3, r4 and r5, connected in parallel with a second resistor R1, R2, R3, R4 and R5. The control electrode of the switches S1-S5 are all connected to a common control line and are operated in synchronism by a multivibrator M producing a pulsed output signal the mark/space ratio of which is adjustable by means of a potentIometer P.

Figure 3 illustrates the frequency response of the 5-pole Chebyshev filter shown in Figure 2, at three different mark space ratios (namely 1 OU/o, 50%, 95%). Figure 4 shows similar response curves but with an expanded y scale to show the 0.5 dB ripple characteristic.

Figure 5 shows the circuit of a 4-pole band pass filter according to this invention and in which the digital control of the frequency is carried out by means of a BCD switch. Conveniently the filter circuit is constructed using, two AF100 chips.

These are state variable active filter networks and include two RC networks within the chip. Each chip can be programmed by external resistors to produce second order functions (i.e. two poles).

Low pass, high pass and band pass functions are available simultaneously at separate outputs.

Independent control of circuit Q, and hence band width, is achieved by a separate variable mark to space ratio multivibrator (not shown) fed to the VARI-O input for controlling a bilateral switch Sq and hence varying the effective resistance of a branched circuit incorporating resistor rq in series with the switch Sq and resistor Rq in parallel with rq and Sq. A separate branched circuit (Sq, rq and Rq) is required at input to each AF100 chip as shown. This arrangement enables independent control of circuit 0 and centre frequency. Alternatively, however, the bandwidth and Q-factor may be ganged (i.e. variable in response to the same control signal) to produce a constant bandwidth variable filter. The digital control of frequency is carried out by means of a DRM chip (digital rate multiplier) with BCD (Binary coded decimal) input from a rotary switch.The DRM is clocked from an external pulse generator (not shown) set to 50% M/S ratio.

With the BCD switch set to 0 the four DRM inputs A, B, C and D are low and there are no output pulses. With the BCD switch set to 1, one pulse appears for every 10 clock pulses. With the switch set at 9 then 9 clock pulses are transmitted giving approximately 45% MIS ratio.

(100% down to 55% M/S ratio is obtained from IC1 pin 5).

Output pulses from the DRM are applied to four bi-lateral switches, S1, S2, S3, and S4. In preference to separate switches S1-S4, a "quad" arrangement of switches in one package is used to give a better matching of on-resistance value. The mark/space ratio determines the effective resistance of the externally connected branched circuit (S, r and R) of each pole, the principle of the control method is the same as already described.

No frequency determining capacitors, at least not externally connected, are required so enabling an economic and compact design. The only capacitors used are for decoupling the supply rails.

The performance of the circuit has been plotted as a family of curves, one curve for each position of the BCD switch.

Since the mark/space ratio of the CLOCK pulse affects the output markispace ratio, the CLOCK pulse should be accurately defined for best stability. However, the CLOCK mark/space may be varied to form a fine control of the center frequency of the filter, with coarse steps derived by logic signals to the DRM. Alternatively, extra DRM's may be added in cascade to achieve greater resolution.

The stability of the response depends on the accuracy of the M/S ratio in addition to component drift. To minimise the effect of M/S ratio shift in the CLOCK a divide by two circuit driven by a generator of twice the clock frequency ensures an accurate 50% clock pulse.

The versatility of the method of controlling filter response by digital methods open the field of control to microprocessor generated response compensation. This in turn may lead to 'frequency agile" filters in a similar manner to the frequency agile synthesized transmitters and receivers popular in high security military systems.

It is preferred to use integrated circuits which have the following advantages: 1. Four bi-lateral switches per pack result in matched switch on states resistances and improved temperature tracking; 2. Extremely high control input impedance isolation between control and signal (1012 ohms typical); 3. Matched control input to signal output capacitance resulting in reduced output signal transients; 4. The use of 4 or more switches per pack is more cost effective than the use of discrete switching networks; 5. The use of the M/S ratio switching technique provides an accurately defined ganged variable resistance with low cross talk (or good isolation) between elements. This feature is of particular advantage in multipole filter networks as the number of poles is not restricted; 6. The circuit we described defines both the minimum and maximum value of resistance.

In the above-described embodiments parallel resistors are switched in and out to vary the circuit characteristics but we also envisage switching capacitors in the frequency determining networks using variable mark/space ratio. This may be useful in very low frequency filters such as those required for Seismic survey instruments, underwater sonar scanning and sub-audio vibration measurements.

Another area of application which may prove worthwhile is in temperature compensated filters.

A temperature sensing element such as thermistor could be used to control the mark to space ratio of a multivibrator in such a way that the temperature coefficient of the capacitors and resistors is eliminated. The temperature stability of filters tends to be a problem at the high frequencies. The upper frequency of filters investigated to date is 30 kHz. One of the features which makes the bi-lateral switch attractive is the fast response and the wide band width of analogue signal handling.

Claims

1. A method of varying the characteristics of a branched circuit which may comprise active and/or passive components depending upon the characteristics to be varied, by varying the mark/space ratio of a signal controlling switching means operable to switch in and out a selected branch or branches of the circuit.

2. A method according to claim 1 wherein the circuit is divided into two or more poles or stages, each pole or stage including switching means operable to switch in and out a branch or branches of that pole or stage of the circuit.

3. A method according to claim 2 wherein the switching means in the said two or more stages are operable in synchronism in response to a common control signal.

4. A method according to any one of claims 1 to 3 wherein the markispace ratio of the control signal is variable continuously.

5. A method according to any one of claims 1 to 3 wherein the mark/space ratio of the control signal is variable in a stepwise fashion by adjusting a digital input to the control signal generating means.

6. A method according to any one of claims 1 to 5 wherein the characteristics to be varied are the amplitude frequency and/or phase characteristics of a network comprising resistors and/or capacitors connected in multipole or multistage passive or active circuits.

7. A method of varying the characteristics of a branched circuit substantially as hereinbefore described with reference to the accompanying drawings.

8. A circuit having variable characterstics and comprising one or more branches, means for generating a control signal having a variable mark/space ratio, and switching means operable to switch in and out a selected branch or branches of the circuit in response to the control signal.

9. A circuit according to claim 8 and which is divided into two or more poles or stages each comprising a branched circuit and including switching means operable to switch in and out a branch or branches of that pole or stage, the switching means of the two or more poles or stages preferably being incorporated in one integrated circuit.

1 0. A circuit according to claim 9 wherein the switching means in the two or more stages are operable in response to a common control signal.

11. A circuit according to any one of the preceding claims 8 to 10 wherein the switching means comprises a bi-lateral switch.

12. A circuit according to any one of claims 8 to 11 wherein each pole or stage of the circuit comprises a bi-lateral switch and a series connected resistor connected in parallel with a second resistor, the control electrodes of the bilateral switches being connected to a common control line and operated in synchronism by a pulsed output from the control signal generating means.

1 3. A circuit according to any one of claims 8 to 1 2 wherein the control signal generating means comprises a multi-vibrator, the time constants of which are adjustable by a potentiometer.

14. A circuit according to any one of claims 8 to 1 3 wherein the control signal generating means comprises a digital rate multiplier clocked by a pulse generator and controlled by a binary coded decimal switch to adjust the mark/space ratio of the control signal.

1 5. A circuit according to any one of claims 8 to 14, and which comprises resistors and/or capacitors connected in multi-pole or multi-stage variable filter network.

1 6. A circuit according to any one of claims 8 to 1 5 and comprising means for varying the mark/space ratio of the control signal automatically in response to variations in another selected parameter of the circuit.

17. A circuit according to claim 1 5 or claim 16 forming part of a transient recorder and wherein the generating means is operable automatically in response to variations in the sampling rate of the recorder.

18. A variable characteristic circuit constructed and arranged substantially as hereinbefore described with reference to and as illustrated in the accompanying drawings.

19. A method of varying an amplitude frequency and/or phase characteristic of an electrical circuit wherein the instantaneous impedance of at least one branch of the circuit affecting the characteristic is repetitively varied between a first impedance value and a second impedance value at a rate higher than the normal operating frequency range of the circuit so that the effective (or average) impedance of the branch in the operating frequency range lies between the first and second impedance values.