Back to RF

Regenerative Receiver

"What we have heah" is my latest version of the classic Regenerative Receiver. It is an outgrowth of the Tapped Cap Loop Antenna, though it does not use a loop antenna for stability reasons. Regens are VERY sensitive to conditions in and around the regenerative front end. I did use the tapped cap approach to level sensitivity and regeneration through the tuning range.

The Regen is an interesting old time (80 yrs) receiver that was developed to enhance the basic sensitivity and selectivity of the basic crystal set (direct rectifying receiver). Back then, tubes were new and Expensive. It was unrealistic to use lots of them for gain. Though later, the TRF (Tuned RF) receiver did become a reality.

By applying an oscillator circuit to the resonant circuit of the crystal set, adjustable regeneration, less than oscillation, is used to increase the natural Q of the resonance. By having an increase in Q, the BandWidth / selectivity and sensitivity of the resonant circuit could be increased. One could start out with a natural Q of 200 and easily boost it to 2K or more. This would yield a BW of 5KHz at a 10MHz input frequency, not too bad for a ShortWave receiver. At the same time, sensitivity has been increased by a factor of ten, making direct detection much more viable. The only drawback is a funky user interface. One must re adjust the regeneration every time the tuning is changed.

The following drawing is of the whole circuit. Ive split it into two smaller sections for the description.

The following section details the regenerative front end.

Refer to the top part of the above schematic. Qrf and Qrg form the regenerative oscillator with the resonant tank Ltap, Ltnk, Dtl, Dtc, and Ctap. Dtl,c are the tuning varactors. They are each a parallel five stack of VHF tuning diodes of the type used in TV tuners. They are of high Q. The usual AM tuning diodes are too low in Q to work up at SW.

Qrg is an emitter follower, and Qrf is a common base amplifier. Together, they form a differential amplifier with the emitters coupled by Crg. Qrf is the RF amplifier. It does two things. First, it takes the low impedance antenna current from Rgr, the RF attenuator, and translates it to the inductor tap Tapl. The main advantage of common base amps is that they have good isolation from input to output. Next, it does the same for regeneration current from Qrg. Regeneration occurs because the two trannies are non inverting, as well as for signals that propagate through the resonant tank. Regeneration is adjusted by Reg, by sets the supply current fed to Qrg. The conduction of Qrg is directly proportional to its supply current. Thus, regeneration can be set from below oscillation to above. Signals that propagate through the tank are band pass filtered by the action of resonance. They are applied to the base of Qrg to complete the loop.

Qrg is also used to perform envelope demodulation or detection. The base emitter junction does the rectification. This is coupled out of the emitter of Qrg by Cc1 to the audio attenuator pot Rga.

Below oscillation, regeneration is used to increase Q, as above. Above the oscillation threshold / into oscillation, it generates its own signal. In this mode, this oscillation signal is mixed / heterodyned with the filtered RF signal through the tank. Demodulation takes place in the same way.

Refer to the top part of the above schematic. This next section is the high gain audio amplifier. It is a form of pre amplifier that ive come up with. It dispenses with the input cap as the first stage is an upside down emitter follower. It buffers the incoming audio from Rga, the audio attenuator.

Qa2-4 for a three stage common emitter inverting amplifier. Open loop gain is very high, >5K. Rf4, 3, 2, and Cb3 are the negative feedback components. They attenuate the feedback to a more civil form. Otherwise, thered be too much of it for the buffer to work with. They, in combination with Rgs, set the closed loop gain of the amp to ~1K5. This is still pretty high as it is needed to amplify the feeble demod signals from the regen stage.

Refer to the bottom part of the above schematic, the HZR. This is the tuning voltage booster / supply for the varactors. These need at least 15V of voltage range to perform their function, but the supply is only 2.4V. Qas and its associated circuitry form a power oscillator. It is one the new oscillator forms, the HZR High Impedance Resonator type. Qas is an emitter follower. Its emitter, node drv, drives the series resonator Lrez, Crez. Now, remember that series resonators have the ability to boost voltage. The series resonant impedance is low compared to Xo, the resonant reactance. Thus, as the series resonant impedance determines current, the current works on Xo to produce a higher voltage. In this case it boosts whats happening at drv by a factor of about 10. At resonance, the current is in phase with the drive voltage because at resonance, only resistance is left. Node rez lags 90 degrees behind drv because of the integration by Crez. Cfb couples this to the base of Qos by way of current flow. This base current leads the voltage at rez by 90 degrees. Thus, the base current is now back in phase with drv, or the emitter voltage. Hmmm . . . everythings in phase. Sounds like regeneration doesnt it. It oscillates at 20KHz.

Now, there are ICs out there that are made to boost low supplies for tuning voltages, but they are PWM square wave devices, with lots of harmonics up to SW. Not good. Theyre OK for cell phones and the like that run in the UHF. The oscillator here is sinusoidal with low harmonic content. Its frequency is kept low at 20KHz to help in this regard. It is ow enough not to be an EMI problem, but high enough not to hear, and is easy to filter later on.

Cbl couples the boosted sinusoid to Dbt,b which is a peak to peak rectifier with a diode ground return. It rectifies this sinusoid and feeds node bst. Cbr smooths the DC produced. The tuning pots are connected to bst. Rfb and Rst divide this voltage and feed it back to Qes, the error sensor. Qal is the active load for the oscillator, and is biased on by Rlb. Qes is a common emitter inverter. As bst rises, the divided voltage matches the base emitter turn on voltage of Qes. This turns Qes on, shunting bias away from Qal, tending to shut it down. This occurs when the divided voltage reaches 1/20th of bst. This is negative feedback to regulate the strength of the oscillation, to regulate the voltage at bst.

The tuning pots Rset and Rspr are used to provide band set and band spread functions, respectively. Rset is a pot to variably divide the full voltage. Rspr is a small tweak of the voltage across Rset. It acts as a vernier. Now, a mechanically variable capacitor can be used to tune the receiver, but I thought it would be novel to try varactors. Also, the mechanical method make the spread function difficult. A separate spread cap would be needed and they don&rsquot act proportionately across the tuning range. The spreading method above does, as the spread ratio is also divided by the Rset pot.

Power for the circuit is from a two stack of NiCd or NiMH cells. Supply current is on the order of 30mA. Tuning coils are pugged into the circuit by euro style block barrier strip connectors. I made the coils using the spider method. That is, you take a disk of insulating material, like plastic or PCB material, and cut an odd number of radial slots around the periphery. The winding weaves in and out of the slots and never overlaps turns. The result is a coil with the ideal one wire spacing between turns. This produces the best Q for the wire used. All pots are chitty chack units. Construction is done on perf board. The layout should flow from the regen stages back through the audio amp. Use one buss wire for ground. The booster is done one the other side of the ground wire with the oscillator stuff being the farthest from the regen stage. Cost is low overall. Performance is on par with most regens. Its fun to use.

Back to RF