Digitally Programmable Phase Shifter

The present invention relates to digitally programmable phase shifters and, more particularly, to a novel programmable phase shifter having a phase-shifted output at which a signal appears with synchronously switched phase, with respect to a reference-phase output.

It is often required that a signal having a referenced phase be phase shifted to provide a second signal of identical frequency and having a phase different from the first signal, which phase difference, or phase shift, can be not only programmably selected in a "random" fashion, but also wherein switching between various amounts of phase shift can be achieved in synchronous manner whereby erroneous transients are not generated in the phase-shifted output signal.

A number of techniques are known in the prior art for providing an output signal having a phase shifted from the phase of a reference signal. In particular, digital signals may be phase-shifted by means of a count-down chain wherein the count is either inhibited or advanced by one count in order to change the phase respectively in a lagging or leading direction. In this count-down-chain phase-shifter, a constant clock frequency train of pulses at the input to the circuit has an extra clock pulse "stuffed" into the clock pulse train in order to advance the count, whereby the maximum possible average clock rate usable with this form of phase shifter is reduced to about one-half the clock rate of the counter utilized for counting the clock pulse train. This reduction in maximum average clock rate is required so that the "stuffed" pulse will still satisfy the timing specifications of the clock pulse counter. This reduction in maximum average clock rate limits the number of phase increments which may be achieved at a given output frequency. If the "stuffed" pulse technique is not utilized, phase changes are limited to changes in the lagging direction only. Phase shifters using a count-down chain require that the total phase change be created by summation of the required number of incremental phase changes whereby the time required to realize a desired phase change, equal to K incremental phase changes, is equal to the time required to implement the minimum phase change K times, and further limits the maximum frequency of phase-shift operations. It is therefore desirable to implement a phase shifter of the programmable type, which does not require the addition, or "stuffing", of one or more extra clock pulses and subsequently allows the phase-shift circuitry to operate at the maximum specified frequency of the logic components thereof, thereby achieving a minimum possible phase step increment, and hence the maximum number of programmable phase increments, for a given output frequency. It is also desirable that the amount of phase shift can be programmed in a random manner so that any desirable phase relationship can be programmed with a single change in the phase shift control data to the circuitry. It is also desirable that the operation of the phase shift circuit be synchronous with a common phase-shift circuitry clock signal, whereby problems with so-called "gating slivers" may be avoided.

In accordance with the invention, a digitally programmable phase shifter comprises a multi-output phase-shift means, such as a digital tapped delay line, receiving input square-wave signals at the desired output frequency and having a total delay equal to (1/T) of the input square-wave. The phase-shift means, e.g. the tapped delay line with a total of K taps, has K outputs each at the same frequency but with a time delay (t.sub.d) between outputs of (t.sub.d =T/K). Each of the outputs of the phase-shift means, such as a tap of the delay line, is fed to one input of a 1-of-K data selector. The data selector control inputs are programmed to cause one of the K phase-shifted square-waves at the selector inputs, from the tapped delay line, to be coupled to the selector output where it appears as a square-wave having the input square-wave frequency, and having a phase that is phase-shifted from the phase of the input square-wave in programmable increments of (2.pi./K) radians.

In one preferred embodiment of the present invention, the multi-output phase-shifting means is a digital shift register, having K stages, that is clocked at a rate K times the input square-wave. A digital latch is utilized between the one-of-K selector and the input data lines upon which phase control data appear, with the data being clocked into the digital latch synchronously by the shift register clock signal, of frequency K times the output frequency. Changes in the plase of the output square-wave therefore occur synchronously with respect to an edge of the input square-wave, thereby avoiding "gating slivers".

Accordingly, it is an object of the present invention to provide a digitally programmable phase shifter whose phase shifted output does not contain "gating slivers" and whose dwell time at a specific phase setting is an integer multiple of t.sub.d.

This and other objects of the present invention will become apparent upon consideration of the following detailed description, taken in conjunction with the drawings.

FIG. 1 is a block diagram illustrating the principles of the present invention;

FIG. 2 is a block diagram illustrating the present invention for use in obtaining phase-shift increments of 45.degree.; and

FIG. 3 is a schematic block diagram of a particular preferred implementation of the 45.degree.-increment phase shifter of FIG. 2.

Referring initially to FIG. 1, a digitally programmable phase shifter 10 receives a square-wave input signal at a desired output frequency f.sub.0, on line 11. The digital square-wave signal is fed into means 12 having a plurality K of outputs each at the input frequency but having progressively larger phase shifts .phi. with respect to the phase of the input signal; the phase .phi. of the H-th output, where 1.ltoreq.H.ltoreq.K, is .phi..sub.H =360/K.times.H, if the phase shift means has a total delay of one period, from input to the last (Kth) output, at the input frequency. Preferably, the phase-shift means is a tapped delay line 12 having a plurality K of outputs 12a, 12b, . . . 12k, each having a progressively greater delay with respect to the delay line input 12n, and having essentially equal delays between each of the input and the first tap 12a and between each pair of the K taps. If the total delay between input 12n and the furthest tap 12k therefrom is set equal to (1/T) of the input square-wave, where T is the period thereof, then the time delay t.sub.d between each tap is (t.sub.d =T/K). The phase increment, i.e. the difference between similar edges of the square waves from adjacent taps, is (2.pi./K) radians or (360/K) degrees.

The K taps 12a-12k from tapped delay line 12 are individually coupled to like-ordered inputs 14a-14k of a one-of-K data selector 14. A selected one of the phase-shifted square-waves appearing at inputs 14a-14k is coupled to data selector output 14n responsive to selection information appearing at data selector control inputs 15, which may be a plurality of input lines having the phase-selection information coded thereon. Advantageously, K is set equal to an integer (N) power of two, whereby binary logic can be utilized to implement data selector 14. Thus, K=2.sup.N, and, if N=3, K=2.sup.3 =8, whereby a total of N=3 data selector lines 15a, 15b and 15c would be required to select that one-of-eight taps 12a-12k to be coupled to data selector output 14n, at which appears a square wave of frequency f.sub.0 and having some phase angle .theta. in increments of (360/K).degree., or 45.degree. for K=8, with respect to the phase of the input signal on line 11, which signal can be taken as a reference-phase first output signal.

The binary-coded information for the data selector 14 may be strobed in a N-bit address latch 18 (where N=3 bits for K=8) which latch receives and stores the phase shift control information on lines 20, as available from suitable external circuitry (not shown) determining the desired phase shift, and which latch 18 provides a steady N bit output to data selector 14. This is especially useful where the external apparatus supplying phase shift control information has additional tasks to complete and does not constantly provide phase shift control information. Thus, by allowing new phase control information to be provided to the data selector at times uniquely related to the transitions of the input square wave, "gating slivers" are eliminated and the duration of a given output phase is uniquely controlled. Accordingly, latch 18 also includes a strobe input 18a which, when enabled, allows latch 18 to acquire the phase shift control information then on input lines 20 and to provide that information to data selector 14 thereafter, and to disregard any additional information on lines 20 when input 18a is not energized.

Thus, outputs from phase shifter 10 are at the input frequency f.sub.o and at a selectable phase of (2.pi.N/K) radians where K is the number of taps (or phase increments) and N is an integer between zero and K.

Referring now to FIG. 2, digitally programmable phase shifter 10' utilizes a K-stage shift register 12' in place of the tapped delay line 12. Shift register 12' has a data input 12'n, receiving the reference-phase square-wave signal at frequency f and establishing a reference phase angle of 0.degree.; shift register 12' also receives clock pulses at a second input 12m', which clock pulses serve to shift the square-wave data signals at data input 12n' sequentially through the K (=8) stages of the shift register and provide delayed square-wave signals at each of the K shift register outputs 30a-30h. If the clock pulses on clock line 32 to clock input 12m' occur at K tims the data frequency, i.e. the clock frequency F=Kf (or F=8f, for a K=8 stage shift register 12'), the total delay in clocking the input square-wave through the K stage shift register is equal to (1/.sub.f) of the input square-wave, whereby the delay t.sub.d between a pair of adjacent shift register outputs, e.g. 30i and 30h, is t.sub.d =(T/K) and is a phase increment of essentially (2.pi./K) radians or (360/K).degree.. The input square wave, at input 12n', is itself derived from the clock frequency F, by division thereof by the number K; thus, a divider 35 has its input 35a receiving clock frequency F and provides an output 35b, coupled to data input 12n' of the shift register, at the square-wave frequency f=F/K.

Each of the K shift register outputs 30a-30h is coupled to a like-designated one of the K inputs 14a-14h of selector 14. A selected one of the K inputs is coupled to selector output 14n in accordance with the binary coded data made available on data selector input lines 15a-15c, from storage in latch 18, which latch receives its phase shift control information on lines C.sub.1, C.sub.2 and C.sub.3. The strobe input 18a' is advantageously coupled to the clock line 32 and is enabled to receive new phase shift control data inputs at the clock frequency rate, whereby any change in the phase increment, as selected via selector 14, is in response to a change in binary-coded information, on lines 15, which information is only clocked into latch 18' at an edge of the clock frequency waveform on line 32. The proper one of the rising or falling edges of the clock waveform is thus used for clocking new data into latch 18 and also for accomplishing the phase shift of the waveform selected to be coupled to output 14n; by the addition of an inverter (not shown) between clock line 32 and clock data input enable input 18a, the opposite edge of the waveform may be used as the clocking edge. In this manner the phase-increment-determining information is stored in the latch at the same time that the input frequency square-wave information is shifted into the shift register, so that a new phase is programmed concurrent with level transitions in the output square-wave and a full half-square-wave-interval is gated out for the first level change at the new phase increment, without the appearance of a gated "sliver" of a square-waveform.

Referring now to FIG. 3, a phase shifter providing essentially equal phase increments of .pi./4 radians, or 45.degree., is illustrated. The 8-stage shift register 12' is implemented with a pair of 4-stage shift registers 51 ad 52, in series connection. Each of the 4-stage shift registers is advantageously implemented with integrated circuitry, whereby, if Schottky-type TTL logic is utilized, the shift registers may be of the 74S175 and the like type. Each of the shift registers has a clock pulse input CP, connected in electrical parallel connection to clock line 32, having the clock frequency F, equal to K times the output frequency (8f.sub.0) thereon. Clock pulse line 32 is also coupled to the strobe input 18a'of the latch 18', which may also be a 74S175 and the like type integrated circuit. The clock-divide-by-K stage 35 is implemented with a divide-by-8 counter 35 which may be of the 74S161 type, wherein clock pulses received at clock input 35a enable a square-wave to appear at the third output Q.sub.3 of the counter with frequency f.sub.0 at a reference phase of zero degrees, for coupling to the data input 12n' of the K-stage shift register 12'. Data input 12n' of the shift register is the data-1 input D.sub.1 of the first shift register stage 51. Each of the first, second and third stage outputs, e.g. Q.sub.1, Q.sub.2 and Q.sub.3 , respectively, of each shift register integrated circuit are coupled not only to the associated input, e.g. 14a, 14b and 14c, respectively, of the one-of-K data selector 14 (which may be a 74S151 and the like type integrated circuit) but are also coupled to the data inputs of the next-successive shift register stage. Thus, the output Q.sub.1 of the first shift register stage is coupled to the data input D.sub.2 of the second shift register stage, while the second stage output Q.sub.2 is coupled to the third stage input D.sub.3 and the third stage output Q.sub.3 is coupled to the fourth stage data input D.sub.4. The output Q.sub.4 of the fourth stage is coupled to the associated data selector input 14d and also to the input D.sub.1 of the first stage of the next successive shift register integrated circuit 52. The data inputs an outputs of each shift register integrated circuit are connected similar to that of the inputs and outputs of the first integrated circuit 51, with the exception that the data output Q.sub.4 of the last shift register stage is connected only to the last remaining input 14-14L of data selector 14.

As previously explained hereinabove, the phase shift control data input lines 20 are coupled to selector 14 via latch 18' and lines 20 connect directly only to the data inputs of latch 18. As there are K=8=2.sup.3 possible phase increments which can be programmed, only three control lines C.sub.1, C.sub.2 and C.sub.3 are required.

The clock rate capability of the Schottky-TTL logic described hereinabove is on the order of 60MHz., whereby the maximum output frequency, of the output waveform appearing at data selector output 14n, is equal to the maximum clock frequency divided by the number of phase increments K, or about 7.5MHz. for K=8. If an additional pair of 4-stage shift registers are added in sequence after shift register 52, and if selector 14 is a one-of-sixteen data selector and divider 35 is a divide-by-16, then the maximum output frequency would be 60/16 or 3.75 megacycles, but with a phase increment of (2.pi./16) radians, or 360/16=22.5 degrees, programmable. Higher frequency performance may be obtained by using integrated circuit logic having inherently higher speed/frequency capability, such as emitter-coupled logic (ECL) integrated circuits. Similarly, if operation at lesser output frequencies is required, other forms of discrete and/or integrated logic circuitry, having lower maximum clock frequencies, may be utilized.

While the present invention has been described herein with reference to one presently preferred embodiment thereof, many variations and modifications will now occur to those skilled in the art. It is my intent, therefore, to be limited only by the scope of the appending claims and not by the specific embodiments described herein.