Super-regenerative Receiver

U.L. Rohde and A.K. Poddar Synergy Microwave Corp., Paterson, NJ

This article describes the super-regenerative receiver (SRR), a device widely used in short distance telemetry and remote control applications. The difficulty in modeling the regenerative behavior served as the basis for cut and try methodology based on results obtained from the simplified model using a conventional approach. This work presents the SRR behavior to practical implementations by dynamically controlling the injection-mode levels of the super-regenerative oscillator (SRO) for optimum sensitivity and dynamic ranges. Integrated Circuit (IC) designers have the luxury of taking for granted that the incremental cost of a transistor is essentially zero, which has given liberty to the large device count circuits that are prevailing today. This is a recent development, but it really was not all that long ago when the economics of circuit design were fundamentally based on the number of device counts, and often the designer was restricted by the relatively expensive active device to try to get blood (or at least rectification) from a stone.1-13 It is amazing that in the early 1920s, Edwin Armstrong devised a super-regenerative receiver circuit using few components that trade log of gain for bandwidth, contrary to the conventional wisdom that gain and bandwidth should trade off more or less directly. The reduction of the number of the device components is not only cost-effective, but also improves reliability. The characteristics of the super-regenerative receiver to generate large-signal gain at very low bias currents and the ability to operate above the cut-off frequency (fT) of the RF device make it attractive and the preferred architecture for integrated ultra-low power wireless receivers. Figure 1 depicts a 1940s 500 MHz SRR, which is bulky and require

manual calibration, but the circuit operates above the device fT, which was an astonishing accomplishments in those days.12-13

Fig.1. A 1940s 500 MHz Super-regenerative Receiver (vintage two-tube superregenerative detector)12

The phenomenon of super-regenerative detection has been the main focus of the research, and it is still an open issue despite significant gains in practical experience and modern CAD tools for design. The SRR circuit uses just a few components, and its basic design is simple, but detailed analysis is complex, due to the time varying and nonlinear characteristics of the receiver circuit. Super-regenerative receivers have been used for many decades, and are still manufactured in large quantities for short-distance data exchanges. Although an SRR has advantages of high gain, simplicity, low cost, low power consumption and constant demodulated output over a wide range of input signal levels, it has also drawbacks of inherent frequency instability. Figure 2 shows the typical block diagram of a super-regenerative receiver, which consists of a matching network, an isolation amplifier, an amplifier with time varying loop gain and a bandpass feedback network forming a regenerative oscillator.13

Ante nna

Band Pass Positive Fee dback Ne twork

Passi ve Matching Ne twork

Buffe r Ampl ifie r

A(t)
Ti me varying loop gain

S RO

Figure 2. A typical bloc diagram of super-regenerative receiver (SRR) The buffer amplifier between the antenna and the SRO (super-regenerative oscillator) performs the following functions: It reduces the RF leakage of the oscillation signal to the antenna, it provides an input match to the antenna via the passive matching network, and it injects the RF input signal current into the oscillator tank without adding significant loading to the SRO. The time varying nature of the loop gain is designed such that the SRO transconductance periodically exceeds the critical values of the transistor transconductance gm necessary to induce instability. Consequently, the SRO periodically starts up and shuts off. The periodic shutdown of the SRO is called quenching. The start-up time of a SRO (the time from enabling the oscillator until it reaches its saturation voltage VSRO,) can be described by:
V trise = rise log SRO v2 n

(1)

where rise is the time constant of the exponentially increasing oscillation envelope, VSRO is the zero-peak RF voltage of the saturated oscillator, and vn is thermal noise generated due to resonator tank and active devices. From Equation 1, the start-up time of the SRO is dependent on the instantaneous noise in the resonator, when the SRO loop gain is unity (at the onset of oscillation).

Figure 3 exhibits the startup of the SRO in presence of thermal noise. Observing that the startup time is very sensitive to low-level signals in the SRO tank, it becomes obvious that this mechanism can be used to amplify any RF input signal.

Figure 3. A typical startup transient of SRO circuit in presence of thermal noise. Although the conceptual block diagram of the SRR is of a compact size, powerefficient and uses small device counts, the design methodology is not well defined, which leads to a cut and try approach, based on results obtained from the simplified model (see Equation 1). Moreover, the compact size of the SRR implies a small antenna, with poor efficiency that degrades the sensitivity of the receiver. To achieve reasonably good performance, a high sensitivity is required that conflicts with the circuit design requirements for low power consumption.3 Figures 4 and 5 show the schematic of the super-regenerative front-end and a photograph of the SRR die, which overcome the above problems and offers improved sensitivity, but at the cost of an expensive 1.9 GHz BAW resonator.13 The BAW resonator exhibits a thermal drift; therefore, additional compensation circuitry is required for integrated solutions.

Figure 4. A typical schematic of the super-regenerative front-end [13].

Figure 5. Photograph of the SRR die13 This article describes a cost-effective methodology to improve the sensitivity of the SRR with practical implementations by using mode-injection locking for optimum sensitivity and dynamic range.

Theory Figure 6 shows a typical SRR circuit that consists of the following basic modules: Antenna, LNA (low noise amplifier), SRO (super-regenerative oscillator), ED (envelope detector), AGC (automatic gain control), QO (Quench oscillator) and BBF

(baseband filter). The working principle is based on a repeated build-up and decay of the self-oscillations in an SRO whose frequency is near or equal to the RF signal frequency intercepted by antenna.

VRF(t)

RF Signal iAGC iq
AGC QO

Antenna Resonator
Cc
Active device with noise sources & feedback Capacitors

v0(t): Output of SRO

vbn
B

rb

B' Noise-Free 2-Port Bipolar

LNA

iRF

C1 Vbe

ibn

icn
C V0

VEnv VEnv BPF

V

VF

VRF

v(t) inr

L C
C2

v(t): Input to SRO

'

SRO

ED

Figure 6. A typical super-regenerative receiver (SRR) circuit.

As illustrated, the RF signal VRF is intercepted by the antenna and injected through an LNA into the SRO module. The main objective of the LNA module is to facilitate the impedance matching and the injection of the RF signal intercepted by antenna into the SRO module. The LNA module optimizes the noise figure and prevents the impedance variations in the antenna, caused by changes in the nearby environments, from pulling the SRO frequency, thereby resulting in a loss of sensitivity. In addition, the LNA module provides isolation to minimize the power injected back into the antenna due to the building-up of self-oscillations by SRO module. The AGC port is driven by QO, which stabilizes the gain of the LNA module so that the SRO operating point can be dynamically controlled for build-up and dampening of the self-oscillations. The start-up time of the SROs self-oscillations is proportional to

the amplitude and frequency of the input RF signal received through the antenna. The gain of the SRO is periodically varied by QO signal (fq) for sampling the RF signal at the point of maximum sensitivity. The demodulation is performed by rectifying the selfoscillation voltage V0 and then filtering the envelope voltage VEnv by the BPF.

The AGC loop is used to stabilize the amplitude of the VEnv by adjusting the bias current, thereby enhancing the receivers input dynamic range.

Design Basis Figure 7 shows a typical block diagram of the SRR circuit for formulating the basic relationship and describing its regenerative behavior and sensitivity.8 As depicted, the SRO module can be modeled as a selective network (SN), integrated with the quench oscillator (QO) and a feedback through an amplifier for stable and sustained regenerative actions.
v RF(t) Antenna SRO v RF(t) LNA v(t) ED v Env LPF vF

v s(t)

SN

v 0

Feedback Amplifier

v a(t)

Ka(t) f q(t)

QO

Quench Oscillator

(t )

Figure 7. Simplified block diagram of the SRR.

The feedback amplifier module shows a variable gain Ka(t), which is controlled by the quench signal ( fq) that forces the system to be alternatively stable and unstable. The typical input signal v(t) of the SRO can be given in a single quench interval as
v(t ) = Vm pc (t ) cos[ 0 t + ] + n(t )

(2)

where Vm is the peak amplitude, pc(t) is normalized envelope of the SRO output, n(t) is the additive Gaussian noise generated by the active devices of the receiver circuit (described by the autocorrelation function Rn ( ) = [ / 2] (t ) with the power spectral density
Gn ( f ) = / 2 ).

The SRO output v0(t) can be described by4-9

& v 0 (t ) = 2 0 K 0 K s p c (t ) v ( )s ( ) sin 0 (t ) d
ta t

(3) where 0 is the quiescent damping factor, s(t) is the sensitivity curve, K0 is the maximum amplification and Ks is the super-regenerative gain. The noise output n0(t) can be given by
n o ( t ) = 2 0 K 0 K s p c ( t ) n ( )s ( ) cos 0 ( t ) d
ta t

(4) From Equations 2 to 5, the signal to noise ratio (S/N) at the SRO output can be given as8-9

 S [ SNR ] o = N output

Ec

tb p c ( ) s ( ) d ta
tb

( ) s 2 ( ) d
ta

tb

ta

(5) where Ec is the average chip energy, which can be described by

V 2 2
tb

Ec =

2 c

( ) d

ta

(6) From Equation 5, the SNR of the SRO can be optimized using Schwarzs inequality as
tb p c ( ) s ( ) d ta
tb 2

( ) s ( ) d
2 ta

tb

ta

(7) From Equations 5 and 7, [SNR]o, opt can be given by

[ SNR]o
opt

Ec

for

pc (t ) = s(t )

(8) The transfer function of the selective network (SN) can be described by
2 0 0 s G ( s) = K 0 2 2 s + 2 0 0 s + 0

(9) From Equation 9, the characteristic differential equation of the SN can be given by
2 v0 (t ) v (t ) v (t ) + 2 0 0 0 + 02 v 0 (t ) = 2 K 0 0 0 s 2 t t t

(10)

where 0 is the quiescent damping factor and K0 is the maximum amplification. Figure 8 shows the typical parallel and series equivalent representation of the SRO.

Block Diagram Parallel RLC Series RLC

v i v

v0 v0 i0

K0 1/G0 1/R0

0 G0/(2C0) R0/(2L0) R0

0
1/LC 1/LC -Ra(t) i0(t)

Ka(t) Ga(t) Ra(t) L

i(t)
Injected Current

G0

-Ga(t) L C

0(t)

v(t) +

(a) Parallel RLC Circuit

(b) Series RLC Circuit

Figure 8. (a) Parallel and (b) Series equivalent representation of typical SRO module.

The feedback network can be described as vs (t) = v(t) + Ka (t)v0(t) , which can be represented by the linear, time-variant, second-order differential equation
2v0 (t) v (t) 2 v(t) v (t) K (t) + 200 0 + 0 v0 (t) = 2K000[ + Ka (t) 0 + v0 (t) a ] t 2 t t t t

(11)
2v0 (t) v (t) 2 v(t) + 2 (t)0 0 +0 v0 (t) = 2K000 t 2 t t

for

K a (t ) 0 t

(12) where (t ) is the instantaneous damping factor of the closed-loop system that can be
characterized as (see Figure 9)
(t ) = 0 [1 K 0 K a (t )]
(13)

(t) 0 dc ta o tb t

Pulse ge ne rated in the pre vious que nch cycle

Pulse ge ne rate d in the curre nt que nch cycle

v 0(t)
se nsitive pe riod

ta

( o)
Build-up starts he re

tb

Figure 9. Typical plot of the damping function and envelope of the SRO output as depicted in Figure 6.

The instantaneous quality factor Q(t) is6-8

Q(t ) = 1 2 (t )

(14) From Equations 9 to13, the general solution of the homogenous differential equation is given by
voh (t ) = Vh pc (t ) cos( 0 t + h )

(15) where Vh and h are the magnitude and the phase, respectively. The particular solution is

& v op (t ) = 2 0 K 0 K s p c (t ) v ( )s ( ) sin 0 (t ) d
ta

(16)
K r = 0 0 pc ( ) s ( ) d
ta tb

(17)
tb

Ks = e

0 ( ) d
0

(18)
0 ( ) d

pc (t ) = e

tb

(19)
s (t ) = e
0 ( ) d
0

(20) where Kr is the regenerative gain, Ks is the super-regenerative gain, pc(t) is the normalized envelope of the SRO output and s(t) is the plot of the sensitivity curve.

Figure 10 shows the plot of the s(t) and p(t) under sinusoidal quench, their
maximum values is unity at t = 0 for s(t) and t = tb for p(t).7-11

(t)

A+ ta o Atb
Tq

s(t)

p(t)

ta

tsa o tsb

tb

Tq

Figure 10. Typical plot of the damping function, sensitivity curve and normalized envelope

Super-regenerative Modes
There are two basic modes of operation: the logarithmic mode and linear mode [12]. Figure 11 shows the typical modes of SRR operation: (a) detector output in the linear mode, (b) detector output in the logarithmic mode and (c) RF input signal.

Figure 11. Modes of SRR operation: (a) Linear, (b) Logarithmic, and (c) RF Input

The linear mode is described by a linear relationship between the amplitude of the RF input signal and the amplitude of the demodulated output. In this mode, SRO does not reach its steady state during the quench period, due to a high quench frequency and/or a low start-up current. However, the typical operation is made of a combination of both modes; hence, it is difficult to predict an exact mode of SRO operation.3 In the linear mode, the oscillation amplitude is measured before the oscillator reaches saturation, providing a high independent gain.13 As depicted in the linear mode, the sampled envelope is much larger in the presence of the RF input signal. Notice that, in this mode, the oscillator is disabled after the amplitude is measured. Thus, in the linear mode, the oscillator never reaches saturation. A logarithmic mode is described by a logarithmic relationship between the amplitude of the RF input signal and the amplitude of the demodulated output characterizes the logarithmic mode. In this mode the SRO reaches its steady-state amplitude at each quench cycle, due to a low quench frequency and/or a high start-up current; therefore, a reduction in the dynamic range of the demodulated output is due to the logarithmic compression of the signal. In the logarithmic mode, however, the oscillator is allowed to saturate during each cycle. The detector circuitry senses the area under the oscillation envelope, providing a signal dependent gain. As shown for the logarithmic mode, the area under the saturated oscillation envelope is increased in the presence of an RF input, resulting from the decreased oscillator start-up time in this condition. Due to severe fading anticipated in dense indoor

sensor networks, a very wide dynamic range is required from the receiver. The logarithmic mode provides an inherent automatic gain control, making its use preferable for this application.

Design Example
For validation purpose, a typical ISM band SRR at 433.92 MHz is taken for design evaluation as described in the Design Basis, with the following specifications: SRO O/P : 1V, Q0 of SN : 50, Q 5Q0 (assuming average Q is 5 times the quiescent Q of the selective network), Gain of SN at resonance : 0 dB, Gain of LNA : 10 dB, RF O/P of Antenna : -1 dBm, Bit rate: 10 kbps, RX modulation: OOK (ON-OFF Keying).8 From Equations 2 to 20, the design parameters can be evaluated as: Oscillation frequency f0: 433.92 MHz. Modulation bandwidth: 10 kHz. Quench frequency fq (fq 10 times modulation bandwidth): 100 kHz Quiescent damping factor 0 :
1 1 10 2Q 0
2

Average damping factor dc : SN gain K0 at resonance:0 dB

1 = 2 10 2 2Q

Regenerative gain K r : 00 pc ( )s( )d 30dB

ta
tb

tb

Super regenerative gain K s : e

t

0 ( ) d
0

40 dB

Normalized envelope pc (t) : e

0 () d
tb

, pc (t) 17dB

Peak gain K : K 0 K r K s 70dB

Effective gain K eff : K pc (t ) 53dB 3dB bandwidth f 3 dB 456 kHz

2.4 GHz SRR (Injection Mode-coupled SRO)

As depicted in the typical super-generative receiver circuit, the SRO module consists of a lumped LC resonator network (Q 50 @ f = 2.4 GHz), which has a low Q and inherent frequency instability due to temperature changes and mechanical shocks that can degrade the accuracy and sensitivity of the SRR. This problem can be overcome by means of a stable frequency reference, such as high Q factor SAW/BAW resonator devices. However, SAW/BAW resonators are expensive, sensitive to microphonics, and also not easily amenable to IC fabrication. The circuit shown in Figure 12 responds to the above problem by making use of a planar injection mode-coupled resonator, which improves the rate of change of phase and group delay of the SRO, thereby improving sensitivity and dynamic ranges. This novel configuration of SRO can be a cost-effective and power-efficient (3V, 6 mA) alternative of SAW/BAW device SRR. The typical schematic shows the 2.4 GHz SRR using injection mode-coupled SRO, where the printed mode-couple resonator network works as a Q multiplier to support an optimum group delay during the active mode of operation (Q

412 @ f = 2.4 GHz), and the circuit is easily amenable for currently available IC
fabrication technology for mass commercial production.

NOISE FEED BACK DC BIAS

Antenna

Mode-Coupled

ED

Self-Injection Coupling

LNA

SRO

Figure 12. Typical SRR using injection mode-coupled SRO.

From Equations 14 to 20, the performance parameters can be evaluated for comparative analysis. The circuit is operated at a chip rate of 1.5 106 chips/sec for a sequence of 15 chip lengths, resulting a bit rate of 1 103 b/sec. Figure 13 depicts the typical SRO output envelope. The overall sensitivity is calculated to be approximately 86 dB for a periodic quench using a Gaussian chip envelope. The drawbacks of injection mode-coupled SRR is frequency instability in planar coupled resonator implementations, due to temperature variations. This problem can be overcome by incorporating a phase locked loop (PLL) device (see Figure 14), thereby improving performance and being easily amenable for currently available IC fabrication

D e te c to r O u tpu t

vRF(t)

technology for mass commercial productions. Additionally, a frequency synthesizer-based approach is configurable for other operating frequency band; hence, a user defined SRR module.

Figure 13. Typical CAD simulated SRO output envelope of the SRR.

fREF vRF(t)

PFD

Charge Pump

Mode-Coupled Planar Resonator Frequency Synthesizer

A ntenna

Tuning Diode

Divider

Output

gm (t)
ED

LNA

RFin LNA
Quench Signal

DSP

Figure 14. A typical high sensitivity injection mode-coupled PLL based SRR

Conclusion

This article describes a planar injection mode-coupled SRR, which is a costeffective alternative to expensive high Q BAW/SAW resonator SROs for SRR applications.

References
1 E.H. Armstrong, Some Recent Developments of Regenerative Circuits,

Proceedings of the IRE, Vol. 10, August 1922, pp. 244-260.

2 A. Vouilloz, M. Declercq and C. Dehollain, A Low-power CMOS Superregenerative Receiver at 1 GHz, IEEE Journal of Solid-State Circuits, Vol. 36, No. 3, March 2001, pp. 440-451. 3 P. Favre, N. Joehl, A. Vouilloz, P. Deval, C. Dehollain and M.J. Declercq, A 2 V 600 A, 1 GHz BiCMOS Super-regenerative Receiver for ISM Applications, IEEE Journal of Solid State Circuits, Vol. 33, No. 12, December 1998, pp. 2186-2196. 4 N.B. Buchanan, V.F. Fusco and J.A.C. Steward, A Ka-band MMIC Superregenerative Detector, 2000 IEEE MTT-S International Microwave

Symposium Digest, Vol. 3, pp. 1585-1558.

5 W.G. McCoy, Design of a Super-regenerative Receiver for Solar Powered Applications, IEEE Transactions on Consumer Electronics, Vol. 38, November 1992, pp. 869-873. 6 D.L. Ash, A Low Cost Superregenerative SAW Stabilized Receiver,

IEEE Transactions on Consumer Electronics, Vol. 3, Aug. 1987, pp. 395403.

7 N. Joehl, C. Dehollain, P. Favre, P. Deval and M. Declercq, A Low-power 1 GHz Super-regenerative Transreceiver with Time-shared PLL Control,

IEEE Journal of Solid State Circuits, Vol. 36, No. 7, July 2001, pp. 10251031. 8 F.X. Moncunill, P. P-Schonwalder and O.M-Casals, A Generic Approach to the Theory of Super-regenerative Reception, IEEE Transactions on

Circuit and Systems-I, Vol. 52, January 2005, pp. 54-70.

9 R. Feick and O. Rojas, Modeling and Simulation of the Super-regenerative Receiver, IEEE Transactions on Consumer Electronics, May 1997, pp. 92102. 10 D.M.W. Leenaerts, Chaotic Behavior in Superregenerative Detectors,

IEEE Transactions on Circuits and Systems-I: Fundamental Theory and Applications, Vol. 43, March 1996, pp. 169-176.
11 A. Jamet, A 10 GHz Super-regenerative Receiver, VHF Communications, Vol. 29, K.M. Publications, UK, 1997, pp. 2-12. 12 J. Whitehead, Super-regenerative Receivers, Cambridge Press, 1950. 13 B.P. Otis, Ultra-low Power Wireless Technologies for Sensor Networks, Thesis, University of California, Berkeley, February 2005.