Saturday, October 30, 2010

Swing Set

Swing Set is a double-parametric, regenerative, 80m receiver. Despite the unusual circuity this is a classic regenerative receiver setup (Q-multiplier -> diode mixer -> AF amplifier). As shown above, a single RF signal source (I've been using my bench, RF signal generator) can be used to pump both the Q-multiplier and the AF amplifier. This tactic minimizes the overall parts-count at the cost of having to track a second RF resonator while tuning across the band.

A degenerate parametric oscillator produces the negative resistance required for proper Q-multiplier operation. A pair of varactors are parametrically pumped at twice the operating frequency with sufficient vigor to drive the input L-C tank circuit just over the threshold of instability. The result is much the same whether one employs a traditional three-terminal feedback amplifier, a tunnel diode or a parametric amplifier. In each case the total positive equivalent conductance appearing across a parallel resonant L-C tank circuit is partially canceled by the application of a negative conductance of nearly identical magnitude. The oscillating Q-multiplier produces high-gain at narrow-bandwidth while providing the necessary "beat-note" for proper CW or SSB reception.

The varactor-based audio frequency (AF) amplifier may be thought of as an RF modulator followed by a diode demodulator. Subsequent to the discovery of synthetic ferroelectric materials this circuit gained a brief popularity in research circles. Another burst of interest followed the introduction of semiconductor varactor diodes, such as I'm using in the above circuit. For best results the RF tank circuit in the AF amplifier is tuned to either side of peak resonance. In practice, one simply tunes this tank for maximum AF gain. So long as the varactors aren't driven into conduction the input impedance of this amplifier is extremely high; thus placing a negligible load upon the Q-multiplied resonator.

The power gain of a three-frequency, parametric upper-sideband up-converter (USBUC) is numerically limited to the output frequency divided by the signal frequency. The power gain of a three-frequency, parametric lower-sideband up-converter (LSBUC) is limited to the negative of this same ratio. Thus, the USBUC can produce modest, but stable gain, whereas the LSBUC can theoretically provide extremely high, but potentially unstable gain.

My varactor-based AF amplifier may also be thought of as a four-frequency parametric up-converter. These devices support both the upper and lower sideband outputs in order to combine the virtues of both the three-frequency USBUC and LSBUC, while avoiding their potential shortcomings. However, the output is an amplitude modulated RF signal which must be demodulated in order to reproduce an amplified version of the input waveform. Despite the loss associated with a demodulator the overall power gain can be substantial.

I've been using SMV1212's for all four varactors; mostly because that's what I have on-hand. The zero-bias capacitance per varactor is approximately 75pF. At a bias of 4v the capacitance drops to 9pF (with a Q of at least 150 @ 50MHz). I've also had luck in the past using certain common rectifiers (such as the 1N4002) in my HF parametric regenerative receiver experiments.

The AF amplifier resonator is simply tuned for maximum signal or noise. The Q-multiplier tuning takes a bit of practice. With the antenna connected, I initially adjust the Q-multiplier resonance (with the pump amplitude set to a fairly high value) until it pops into degenerate parametric oscillation. I gradually lower the pump amplitude while continually adjusting the tank resonance in order to maintain parametric oscillation. With the pump amplitude reduced to the lowest value that affords stable parametric oscillation, I further reduce the oscillation amplitude by slightly detuning the resonator to a lower frequency. As with any regenerative receiver, the aim is to set the oscillation to a value just above the onset of oscillation in order to maximize the gain and narrow the bandwidth.

If the AF amplifier squeggs, or is otherwise unstable, I lower the pump amplitude or raise the number of demodulator coupling turns on the RF resonator (or both). Stability is also increased (at the expense of gain) by raising the pump coupling capacitor value (I've used values from 5 to 15pF). It's important to use a low-loss RF resonator in the AF amplifier in order to achieve a high amplifier gain.

It's been fun listening to the 80m CW band with this receiver in the past two weeks. The Europeans are quite loud in the early and late evening hours. I've mentioned elsewhere how I enjoy hearing fellows operating vintage equipment. One evening I heard K8FN running an Eico 720 into an open-wire fed dipole through his Johnson Matchbox. On several nights I copied K4JJW putting out a lovely signal with his Johnson Ranger @ 40w into an inverted-L antenna.

The stronger stations provide armchair copy with the headphones resting on the bench. In fact, I had a perfect copy of W1AW one evening with no antenna connected. Speaking of which, it's necessary to attenuate the very strongest signals; especially as my varactors are unbiased.

1 November 2010
I popped in two pair of 1N4001 rectifiers in place of my SMV1212 varactors this evening. A bit of retuning rewarded me with 80m signals. The Q-multiplier appeared to be working much the same as before. However, the AF amplifier gain was considerably reduced. I suspect the dc/dv slope for the 1N4001 isn't as steep as with my SMV1212's. The Qu might be lower as well. Still, tooling around the 80m CW band with a 0-V-1 regenerative receiver made from four power supply rectifiers and my bench, RF signal generator is pretty cool!

Peter, DL3PB, suggested that I try LEDs for my varactors. He notes their higher forward-barrier voltage would be a virtue in non-biased receiver circuit. Secondly, they come equipped with a visual forward-bias indicator.

I briefly tried single, red, green and yellow LEDs last night, as well as paralleled pairs, as replacement varactors. While they didn't work as direct replacements for my SMV1212's (I couldn't pump the Q-multiplier over the threshold of instability), this was anything but a conclusive test. I'll need to make some measurements on the LEDs. Thanks again, Peter, for the excellent idea!


  1. Michael,

    Is the RF detector Ge diode leaky enough to not need a DC return path with the high-Z AF amp input?

  2. G'day Alan,

    That's exactly right. I got the idea from Chuck, N1TEV. The same leaky germanium diode also provides a DC path to ground for the varicaps; without which the resonance floats erratically.

    "Therefore, since brevity is the soul of wit,
    And tediousness the limbs and outward flourishes, I will be brief.
    Your noble son is mad..."

    Polonius; speaking about QRP minimalism in "Hamlet" :o)


  3. This comment has been removed by a blog administrator.