A regenerative receiver satisfying both the least components count and the maximum performance is the 'Twinplex'. A regenerative detector stage plus an audio amplifier, which makes a '0V1' configuration. To decouple the antenna from the tuning circuit, a grounded grid amplifier has been added, thus turning the '0V1' into a '1V1'. The frontend is a 6AS6/5725 (but any triode or pentode is OK), and the '0V1' section is made by a '5814', or a special 12AU7/ECC82 (any double triode with comparable voltage amplification would do fine equally).
One of the crucial aspects with the regeneratives to achieve the best performance, is a relatively high amplification level, which in the original concept of the 20's shall be partially achieved by an inter-stage audio transformer, between the detector and the audio amplifier. This component is almost vital to get a strong audio with normal amplification level achieveable by only one 'low mu' triode (second half of the 5814. In my receiver it is connected as voltage elevator providing an audio gain of 4 (i.e. about +12 dB).
The second crucial aspect is a 'high Q' coil in the detector: the use of 'Litz' wire is the only way to get the lowest losses at RF, which raises the intrinsic selectivity of the tuned circuit at the onset of oscillation in the detector.
In my prototype (... nowadays, for me any homemade device is a 'prototype' due to lack of time which forbids to buld something showing proper look, stability and performance ...) an audio filter has been added, ahead to the audio amplifier section: featured by a low loss inductor, offers a high selectivity around 800 Hz, which allows at some extents to pick up signals even in 'pile-ups' on the radio amateurs bands.
The knobs on the front panel are: main and fine tuning, regeneration control and volume control.
On the back panel, RF attenuator and the filter switch + power supply, antenna and headset sockets.
Both top and bottom layout pictures are self explanatory. Everything has been taken from the 'radio-depot', the metal pieces for the chassis included !
The RF attenuator is very useful when full size antennas are used, but the receiver shows an impressive sensitivity even with a piece of wire thrown on the bench. Besides, the grounded-grid front-end offers such a separation from the antenna, that the audio pitch when listening to CW stations is little affected by the usual oscillations of the same antenna in the wind. The front-end offers only a limited voltage gain, but is essential to preserve the tuned circuit 'Q', that would otherwise be damped by a straight connection to the antenna, even with very light capacitive direct coupling, as is usually seen in the schematics.
As it is, the listening to CW stations on low HF bands is very comfortable, even though the overall selectivity against strong near signals is poor. Experiments made on another regenerative have shown that the use of a small tuned 'magnetic' loop improves dramatically the lack of strong interfering signals rejection. Others have tried the use of tuned passive loops with regeneratives, reporting very interesting results.
High impedance headsets are used with this receiver and normally I have to keep the volume at 1/4 .. 1/3 due to the very loud sound achieveable by this configuration.
A side to side comparison with a professional, surplus, receiver, shows no relevant differences in sensitivity. The AGC absence in the Twinplex makes sometimes the reception more comfortable in presence of natural atmosferic noise.
What else ? The next step will be to couple the Twinplex to the Hartley transmitter: the circle will be closed. The 'evolution' will be complete: a series of progressive steps back in the radio technology, to arrive at the point where it started. Because nothing more than a handful of triodes is what's needed ... to surf the Shortwaves ...
Radioman, Oct 2020