2014年2月16日日曜日

Low-complexity and low-power regenerative receiver with remote control of tuning and regeneration for use in electromagnetically noisy urban enviroments

Abstract

A novel bipolar transistor regenerative receiver design is presented that covers medium-wave to short-wave frequencies. The main novel feature is the remote control of tuning and regeneration achieved through DC control voltages. This arrangement is particularly useful for use in electromagnetically noisy urban environments because it allows physical separation of and optimal separate placements of the RF and the DC portions of the receiver: the RF portion of the receiver responsible for intercepting incoming radio frequency energy can be placed in an electromagnetically non-noisy area, whereas the DC portion of the receiver responsible for specifying the receiver control voltages can be placed in a convenient, humanly-accessible operating location which however may be electromagnetically noisy. While other existing solutions can achieve the same result by using other means, for example by using remotely-located active antennas or passive tuned antennas, the novel arrangement proposed here has the benefits of preserving the low circuit complexity, the single non-ganged tuning arrangement, and the low power consumption of the regenerative receiver architecture. Other receiver features include low fractional coupling between the resonant tank and the amplifier, use of a simply-constructed two-terminal inductor, ability to operate off of only 1.2 volts, a biasing arrangement that does not require the use of a radio frequency choke, and a wide tuning range of approximately 3 octaves.

Introduction

Regenerative receivers remain popular among hobby and amateur radio constructors. Though the performance of the regenerative receiver architecture has been surpassed by later architectures such as the superheterodyne or direct conversion, the ease of understanding the principles of the basic regenerative receiver makes it appropriate both as a beginners' construction project as well as a platform for exploring in depth the detailed behavior of an almost- or only marginally-oscillating system. This paper focuses on the former application area, namely, a regenerative receiver designed for practical use that can be simply built by a hobbyist and used for reception of medium-wave and short-wave broadcasts.

Simplicity: The appeal of the regenerative receiver


One reason for the popularity, both past and present, of regenerative receivers is that a comparatively simple circuit can achieve acceptably high levels of sensitivity and selectivity thanks to the repeated selective amplification within a single active device. Additionally, an active device used as a regenerative detector can typically run off of very low power because the active device requires only enough power to sustain an oscillator loop gain of 1.

Another attractively simple aspect of the regenerative receiver is the minimal antenna requirement. Simple regenerative receiver circuits often directly connect a short, high-impedance whip antenna directly to top of the oscillator's LC tank. The high impedance of the short whip antenna matches the high impedance of the Q-multipled LC tank without any need for additonal impedance transformation mechanisms such as a transformer or buffer stages. The impedance match allows efficient energy transfer from the antenna to the LC tank, with the result that good reception - in an electromagnetically non-noisy environment - is possible with a short antenna. Some designs even dispense with an antenna altogether and make the tank inductor a single-turn or multi-turn solenoidal coil with a physical diameter large enough to allow the tank inductor to directly intercept incoming radio energy, acting as a tuned and Q-multiplied magnetic loop antenna - again achieving good reception in an electrically- and physically-simple circuit topology that requires little power.

Problem: Electromagnetically noisy urban environments

The problem that this paper attempts to address is that of designing a simple regenerative receiver for use in an electromagnetically noisy urban environment. Although there are a number of simple regenerative receiver designs available in print and on the WWW, attempting to use these simple regenerative receivers with a short whip antenna in an electromagnetically noisy urban environment is likely to yield poor results. Urban environments, such as the environment inside an office or apartment building, degrade performance of receivers through two major mechanisms: the attenuation of desired signals due to the electromagnetic shielding of the building structure, and interference to the desired signals by electromagnetic noise sources inside the building. The short whip antenna, often recommended for simple regenerative receiver designs, doesn't work inside a noisy building.

Existing solutions

Let us now consider existing solutions to improving reception, and why the existing solutions are inappropriate for the problem of good reception with a simple circuit in an electromagnetically noisy urban environment.

Simple but inadequate solutions


Resonant outdoor antenna


One simple solution is to put up a larger resonant antenna, such as a half-wave dipole. The antenna should be located outdoors and away from noise sources, with a shielded coaxial cable being used to bring the signal from the antenna to the receiver without noise pickup. This solution is inadequate because an urban environment typically will not offer enough space to put up a full-size resonant antenna outdoors and away from noise sources.

Random-wire outdoor antenna


Another simple solution is a short, outdoor random wire antenna. Even in an urban environment, such an outdoor antenna might be feasible - for instance, by throwing a length of wire out a window or stringing it across a balcony. But such an antenna will be susceptible to noise pickup where it enters the building and is strung across the electromagnetically noisy room to reach the receiver. Reducing noise pickup on random wire antennas might be achieved through use of coaxial cable and good RF grounding practices, but this is again impossible in a typical office or apartment building that will have no good RF ground available.

Passive high-impedance remote untuned whip with matching transformer


A passive antenna was tried consisting of 2 meters of wire formed into a short doublet, with the center of the doublet wound 25 times through an FT50-43 core as the primary of an impedance transformer. The secondary consisted of 5 turns, connected to 50-ohm coaxial cable, intended to transform the high impedance of the short antenna to something closer to 50 ohms. The hope was that this might allow a remotely-located and untuned short whip antenna to be simply and efficiently coupled to the regenerative detector. The results were disappointing: the antenna delivered very low signal levels regardless of its connection to the regenerative detector (through a grounded-base amplifier, or connected directly to the tank).

More complex yet still inadequate solution


Active high-impedance untuned whip antenna


A remotely-located, active, untuned whip antenna such as the PA0RDT mini-whip antenna can deliver high signal levels to the receiver, but suffers from common mode noise. Electromagnetic noise from the noisy indoor environment is conducted up feedline and is coupled into the very-high-impedance amplifier, where it then is audible at the receiver. An RF choke by itself is ineffective because the choke impedance can realistically never be greater than the very high impedance of the whip antenna amplifier, meaning the amplifier still sees a significant portion of the noise even with a choke present. Properly addressing the common-mode noise problem requires providing a low-impedance RF ground that is located outside of the noisy indoor environment. Such a grounding arrangement is typically impossible in an urban environment such as an apartment building.

Adequate but complex solutions


Remotely-tuned, passive, small loop antenna


A remotely-tuned small loop antenna, also called a magnetic loop antenna, is a solution that can work well. The small loop can be located in an electrically non-noisy location. A smaller coupling loop serves as an impedance transformer that matches the antenna's impedance to the 50-ohm coaxial cable. The coupling loop has an important benefit: it galvanically isolates the loop antenna from the feedline and has a high rejection of common-mode noise conducted on the outside of the coaxial cable shield, meaning that noise picked up on the indoor coaxial cable run is effectively blocked from reaching the antenna and the receiver. Some care needs to be taken in the remote tuning arrangement; for example, if a varactor is used, then the DC control lines that supply the varactor reverse bias must be heavily isolated from the loop's resonant circuit (using high-value chokes or resistors), or else any RF or AF noise picked up by the control cables will modulate the varactor control voltage which can lead to distortion of incoming signals. The disadvantage of the remotely-tuned small loop antenna is that it must be constantly retuned to the reception frequency. With a regenerative receiver, this means tuning requires operation of at least three controls: antenna tuning, main tuning, and regeneration. Another disadvantage is that the tuned loop antenna typically requires an isolating RF amplifier stage ahead of the regenerative detector to prevent unwanted interaction between the loop antenna's high-Q LC resonant tank and the regenerative detector's high-Q LC tank. Furthermore, although the regenerative detector itself can run off of low power and low voltage, it is difficult to design a well-performing isolating RF amplifier (exhibiting both gain and low distortion) that runs off of low power and low voltage, thus requiring higher voltages and higher power consumption if an RF amplifier is desired.

Active low-impedance untuned small loop antenna


A remotely-located, active, untuned loop antenna, such as the M0AYF design, can work very well as it requires no separate RF ground (being a balanced design), and uses a low-impedance amplifier, making it possible to choke off common-mode noise. The only disadvantage of this solution is that it requires significant power to achieve acceptable sensitivity and IMD performance, which detracts from the minimalist appeal of low-power operation possible with regenerative detectors.

Remote control: A simple and adequate solution


In order to maintain the low circuit complexity and low power consumption of the regenerative receiver design, while simultaneously allowing remote placement of the antenna in a non-noisy location, the following design decisions were made:

  1. to use a high-impedance whip antenna connected directly to the regenerative detector's LC tank,
  2. to locate the regenerative detector in a non-noisy location, and
  3. to use a remote-control cable to operate regenerative detector from a comfortable operating position.
This design requires no impedance matching between the antenna and the detector, and also requires no additional power as would be required by an active antenna.

As mentioned earlier, an alternative but equally simple solution is to eliminate the whip antenna and make the LC tank's inductor a large-diameter air-core solenoidal inductor, allowing the tank's inductor to act directly as a tuned loop antenna. The remote control of the regenerative detector is essential to allowing this simple solution to be practical in a noisy urban environment.

Remote controlled design


The easiest way to achieve remote control of a regenerative receiver is by using DC control voltages. The circuit must allow voltage control of tuning and regeneration. Then, the DC control voltages can be sent over a long control cable, with the potentiometers controlling the voltages being located at a convenient operating position. When designing the control cable wiring, it is essential to avoid ground loops and to have adequate RF and AF bypassing on the control lines to prevent any noise on the control cable from reaching the regenerative detector.

For remote control of tuning, the obvious choice is a varactor. For remote control of regeneration, it was decided to alter the regenerative detector's supply voltage. Both of these remote control schemes allow heavy RF and AF bypassing of the DC control lines for high noise immunity.

Other design parameters


Other parameters desired for the receiver design included:

  • Use of a DC-grounded inductor. Grounding the inductor is important to prevent low-frequency noise coupled into the inductor from reaching the rest of the circuit. The easiest way to achieve this is to have the inductor grounded at DC. (Some oscillator topologies such as the Vackar have the inductor floating, which requires use of a choke to keep the inductor floating at RF but grounded for low-frequency noise.) 
  • Use of a two-terminal inductor without a tap or a tickler winding. This allows ease of inductor construction, particularly if a large-diameter solenoidal coil is used as a loop antenna.
  • Fractional coupling of the active device to the LC tank. The active device should not be connected directly to the top of the LC tank, but instead should be connected to a lower-impedance point to avoid loading of or frequency-shifting of the tank by the active device's parasitic capacitances. Since the previous design parameter precludes use of a tapped coil or a link winding, the only way to achieve fractional coupling is to use a capacitive tap on the LC tank.
  • No use of radio frequency chokes. Radio frequency chokes can have low Q and care needs to be taken during circuit design with regard to the choke's self-resonant frequency. To avoid these issues completely, the design should avoid the use of chokes entirely.
  • Low-voltage 1.2-volt operation. The receiver (with the exception of the varactor bias voltage) should run off of a single 1.2-volt rechargeable cell. Combined with the fractional tank coupling and the choice to avoid using chokes, the low voltage requirement makes it challenging to design the circuit with enough gain to oscillate.
  • Wide tuning range of three octaves (3x). Some oscillator designs like the Colpitts or the Seiler have a limited tuning range due to the large feedback capacitors in parallel with the tuning capacitor, inflating the minimum tank capacitance and thus limiting the upper tuning range. The total series capacitance in parallel with the tuning capacitance should be kept as small as possible: large enough to ensure oscillation, but small enough not to unnecessarily inflate the minimum tank capacitance.

The circuit

A circuit was designed fitting the above design parameters. The circuit uses a 2N3904 bipolar junction transistor in a common-collector oscillator configuration. C2 and C3 provide an approximately 1:100 impedance transformation between the transistor base and the top of the LC tank. Because of the 1:100 impedance transformation, the base bias resistor R2 does not significantly load down the tank. The total series capacitance of C2, C3, and C4 is less than that of the lowest value capacitor, which in this case is 20 pF. Therefore, the total capacitance in parallel with the tuning capacitance (provided by D1) is less than 20 pF, which is not excessively large and does not unnecessarily restrict the minimum tuning capacitance. This allows a wide tuning range.

Varactor D1 provides an approximately 35-500 pF variable capacitance and is not grounded directly, but instead is grounded through C4, which raises the variable capacitance above ground. This is equivalent to a Vackar oscillator (see reference 3); the difference is that the Vackar grounds the common point of the variable capacitance and C4, whereas this "topologically rotated" Vackar oscillator grounds the inductor L1 instead.

C5, C6, C7, and C8 provide RF and AF bypass for the DC control lines. The DC control lines run from the main receiver board through a CAT5 network cable to the control panel. The control panel contains the batteries V1 and V2, the remote control potentiometers VR1-VR3, and the AF amplifier. VR1 controls the regenerative detector supply voltage. VR2 and VR3 provide coarse and fine control of the varactor tuning voltage. Audio is taken off of the Q1 emitter and fed over the control cable into the AF amplifier, consisting of a low-impedance common-base stage followed by three common-emitter stages. A crystal earphone is used for output.

To avoid excessive losses in the resonant circuit, capacitors C1, C2, C3, and C4 should be high-quality, low-loss capacitors such as NP0 types or equivalent. Failure to use high-quality capacitors here may lead to difficulty in bringing the set into oscillation. For operating convenience, VR1, VR2, and VR3 should be high-quality ten-turn potentiometers, to allow fine adjustment of tuning and regeneration. Normal one-turn potentiometers can be used instead, but adjustments will be much more critical.

Performance


General


In the indoor test environment, no reception was possible with the receiver and whip antenna located indoors. Remotely locating the receiver and whip antenna on a balcony enabled adequate-volume reception of shortwave broadcasts and low-volume reception of local amateur transmissions.

Cable noise


Routing the control cable through an electromagnetically-noisy area revealed a slight tendency for frequency-modulation of the tuning control voltage. This could be reduced by increasing C5 to 10 uF or greater, but this carries with it the disadvantage that the tuning control is less responsive, since the larger capacitor requires more time to charge or discharge.

An additional source of noise is likely the AF output line which is connected directly to the emitter, which is in turn directly connected to the bottom of the varactor. Since the AF output line supplies AF output voltage it cannot be AF bypassed, making this line susceptible to noise pickup; furthermore, any noise pickup will directly affect the varactor's tuning voltage. This could be avoided by adding a separate detector or buffer stage and taking the AF output from that stage, so that the noise on the AF output line is isolated from the sensitive varactor tuning circuitry.

Possible signal distortion through AF modulation of varactor voltage


Another potential issue with the varactor configuration is signal distortion due to the floating configuration of the low-impedance side of the varactor: it is not directly grounded, but is instead grounded through C4 (which is necessary for the Vackar-style feedback and allows a wider tuning range and a lower fractional tank coupling than a directly-grounded varactor). The disadvantage of this configuration is that the bottom of the varactor is then also connected to the emitter, which develops the detected AF output across the resistor R3. This means that any AF voltage developed at the emitter raises or lowers the DC voltage at the bottom of the varactor, which theoretically leads to varactor detuning and signal distortion.

In practice, however, listening to amateur SSB transmissions on 7 MHz with a 50 cm whip antenna revealed minimal distortion and clear intelligibility. The audio circuit is currently not high-fidelity, using a crystal earphone for output, so it is difficult to determine by ear the amount of distortion present. Most likely, the low signal levels due to the short whip antenna, the high detector linearity, and the low R3/C4 RC time constant yield acceptably small AF signal levels at the emitter such that the varactor detuning is minimized.

Using an attenuator to minimize AF modulation of varactor voltage


An RF attenuator might be added to reduce signal levels in those cases when high detected audio levels at the emitter cause disruption of the tuning voltage.

Using a DC-grounded varactor to avoid AF modulation of varactor voltage


An alternative design might avoid this issue completely by using a traditional Vackar topology where the tuning capacitance is grounded and the inductor is floating. Grounding the varactor would isolate it from any AF voltage appearing at the emitter.

Unfortunately, such a topology requires use of a radio-frequency choke as the transistor's collector load. This contradicts one of the original design constraints to avoid the use of chokes. A collector load choke is necessary in this configuration due to the other design constraints of low fractional tank coupling, restriction to a two-terminal inductor, and only 1.2 volts of supply voltage; use of a collector load resistor instead would reduce gain excessively. The collector load choke could be replaced with a link winding that is magnetically coupled to the tank coil, if the original design restriction of "two terminal inductor with no link windings" is relaxed.

However, at least one other choke is still required in this configuration, to ground the inductor at AF while keeping it floating at RF. As mentioned earlier, grounding the inductor at AF is necessary to reduce susceptibility of the receiver to low-frequency noise (such as mains hum) induced into the coil.

Using an RF-floating and AF-grounded varactor to minimize AF modulation of varactor voltage


Another way to use a choke to reduce the potential for tuning voltage AF modulation would be to keep the existing Vackar-style circuit with the floating varactor, but to replace R3 with a choke, which would have the effect of almost grounding the varactor at AF while keeping it floating at RF. This also would remove the emitter RC network previously consisting of R3 and C4, meaning that there will be minimal filtering of the RF signal at the emitter and hence minimal AF voltage fluctuations at the emitter.

Circuit simulations showed that a 176 uH choke, which could be hand-wound with 20 turns on a FT50-43 ferrite toroidal core, allowed circuit oscillation over a wide frequency range from medium-wave to short-wave frequencies, using tank inductors from 200 uH to 0.5 uH.

If a choke is used in place of R3, some care needs to be taken to avoid unwanted resonances between the choke and C4. For instance, a value of 176 uH gives a resonance with C4 (3280 pF) of 209 kHz, which, depending on one's geographical location, may or may not have any strong broadcast signals that might disturb the detector and cause ghost signals or other distortion to appear. The chosen resonant frequency between the choke and C4 should be free of any strong broadcast signals.

Of course, if emitter resistor R3 is replaced with a choke as just described, the audio can no longer be taken off of the emitter and must instead be taken off of the collector or a separate detector stage.

Miscellaneous observations on performance


There is some tendency for the AF stage to motorboat, particularly with a longer antenna.

An increase in volume can be had by shorting out C9 and removing R4. This has the effect of directly coupling the regenerative detector and the common-base AF amplifier. This however may increase unwanted interactions between the AF stage and the regenerative detector, evidenced for example by motorboating.

Super-regeneration can occur if the regeneration is turned too high, particularly at the high end of the tuning range.

The set does not have quite enough gain to oscillate at the extreme low end of its tuning range, when the varactor reverse bias is zero volts and the tuning capacitance is at its maximum. A varactor reverse bias of zero volts is typically not recommended because it can lead to increased varactor losses. The variable bias network for the varactor could be adjusted so that the minimum reverse bias is always at least one volt.

Depending on the time of day and the antenna location, sometimes strong AM stations will be audible over wide portions of the tuning range (a phenomenon known as "blanketing"), caused by the strong AM signal breaking through the LC tank selectivity and being directly detected by the Q1 base-emitter junction. Again, as with the potential varactor distortion issue mentioned above, an RF attenuator could be used to address this problem. For this remotely-controlled design, a remotely-controllable RF attenuator would be necessary for on-the-fly adjustment of RF attenuation. Though this would increase the circuit complexity, this might be accomplished with a JFET configured as a variable resistor, with the JFET bias being controlled by a remotely located potentiometer. A separate RF stage might also be needed to implement RF attenuation.

Conclusion and future work

A novel regenerative receiver design was presented with remote control of tuning and regeneration. This arrangement preserves the low circuit complexity and low power consumption of the regenerative receiver while allowing operation in noisy urban environments through separate siting of the receiver-antenna assembly and the control panel. Future improvements could focus on improving the noise immunity of the control cable, improving detector efficiency by using a separate and separately-biased transistor stage, and better AF amplifier design to reduce motorboating and increase gain. The AF amplifier might also be co-located with the regenerative detector, to increase the amplitude and noise immunity of the AF signals passed along the control cable and to isolate the potentially noisy audio output line from the sensitive varactor tuning circuitry. A balanced audio arrangement might also offer higher noise immunity. A remote-controlled RF attenuator could also be useful for reducing signal levels in cases of varactor-induced distortion or AM blanketing.

References

Revision history

  • 2014/02/16: Initial revision
  • 2014/02/17: Corrected errors on schematic: C4 should be 3280 pF (not 3680 pF). VR1 should be 10k (not 1k). Added some more notes on performance.
  • 2014/02/20: Added requirement of 3x tuning range. Rearranged active antenna/passive antenna discussion. Added comments about varactor voltage modulation by AF and a remote-controlled RF attenuator.
  • 2014/02/21: Expanded discussion of possible problems with AF modulation of varactor tuning, and suggested solutions.