This page describes a simple regenerative receiver design intended for receiving amateur-radio radiotelegraph (CW) signals. (It also works well for receiving SSB signals.) Although I originally built it for 7-MHz reception, it can be scaled to work on other frequency bands. Figure 1 and Figure 2 show two variations on the design.

Figure 1—Regenerative receiver using 12AU7/ECC82 dual triodes. The detector is a synthetic tetrode; the calculated inductance of L1 is 1.42 μH; and the 105-V supply is VR-tube regulated. The resistive attenuation at the receiver input (approximately 18 dB, provided by a terminated 10-dB pi-attenuator pad in combination with mismatch loss) brings the noise floor ("band noise") of my antenna system down to a level just above the receiver noise floor at just above critical regeneration. (Other antenna systems may require more or less attenuation.) Although the regeneration control (REG) arrangement shown here uses a 10-kilohm control bracketed by fixed resistors to spread the circuit's critical-regeneration point across the control's rotation, you may want to begin with the regeneration-control arrangement of Figure 2 (50-kilohm control in series with 100-kilohm resistor) when setting up your tuned circuit and its regeneration tap.

Figure 2—The Figure 1 circuit modified for use with 6AK6 pentode tubes. (6AK6s are generally less expensive than 12AU7/ECC82s or 6C4s [the equivalent of one 12AU7/ECC82 section] on Ebay.) Although a triode-connected 6AK6 works in the triode AF amplifier portion of Figure 1, I wanted more gain than the amplification factor of a triode-connected 6AK6 can provide (μ of 9.3 as opposed to 20 for a 6C4 or 12AU7/ECC82 section) and so used the fourth 6AK6 as a pentode voltage amplifier as shown here. (A 6BH6 also works well in this stage with attention paid to the fact that the tubes' cathode and grid 3 connections are flipped relative to each other.) So smooth is the transition across critical regeneration with the 6AK6s that the simple regeneration control arrangement shown here is sufficient for operational use without further modification. This site's Homemade Equipment Pictures 3 page shows a practical implementation of this circuit.

If the detector cathode tap is placed too far up from common on the tuned-circuit coil, oscillation may not cease with the REGENERATION control turned all the way down (placing the detector "screen" at common). If you encounter this condition, adding a few hundred (100 to 470) ohms, unbypassed, between the cathode tap and the detector cathode may restore more-normal control of regeneration. This circuit change, which adds some cathode bias (at dc) and degeneration (at ac, which includes RF), may also help to smooth the transition through critical regeneration in any Hartley regenerative detector that goes into oscillation "too abruptly" as the REGENERATION control is turned up. (Note, however, that overabrupt transition across critical regeneration commonly indicates that the detector is loaded too heavily by—is coupled to closely to—the antenna or the stage that otherwise precedes the detector.

Muting the Figure 2 circuit (for silence during transmitting periods) with relatively thump-free make and break required some experimentation. The gain of the 6AK6 pentode AF voltage amplifier is high enough that grounding and ungrounding its grid—a muting method I've successfully used for years with triodes in this first-AF stage—results in unacceptable bangs. Using a high-voltage power MOSFET to ground and unground the AF amplifier screen (with 1 kΩ of series resistance to slow discharge of the 1-μF screen-bypass capacitor) was pleasantly soft on make but too thumpy on break, as was grounding and ungrounding the pentode plate for dc. (Similar brute-force ground-the-B+ methods were sometimes used to obtain thump-free Morse-code-transmitter keying in the 1930s.) In the end, merely grounding and ungrounding the circuit's AF output terminal with a 12-V-solenoid reed relay proved to be an acceptable solution.

This circuit can be easily scaled to other frequencies. So far, I've used it with good results in spot-tuned form at intermediate frequencies (455, 1700, 1750, 1843 and 1850 kHz), in tunable form at 80 meters (3480–3600 kHz), 40 meters (7100–7125 kHz), 30 meters (10100–10150 kHz), and 20 meters (14000–14100 kHz), and driven it (after suitable additional attenuation) with the 3180-kHz second-IF strip of a Yaesu FR-101S MF/HF receiver. The FR-101's heavy automatic gain control usefully augmented the already good frequency-pulling dynamic range of this high-capacitance detector design.