with Germanium clipping/limiting diodes in the RF feedback path, to chop off the modulation and to give 'exalted carrier' /homodyne/synchrodyne -like operation to the Regeneration characteristic. This enables "phase-lock" onto medium-to-strong carriers and allows "Hi-Fi" audio reception!
This radio is LOUD, on strong stations with an external antenna and tuned ground!
[Preliminary notes at bottom of this page]
Rough Draft Notes:
This radio is LOUD, on strong stations with an external antenna and tuned ground! I use a Radio Shack Amplified Speaker, into a quality auxiliary speaker borrowed from a stereo system, with a "Ewe" antenna and earth ground. Adding the extra L-C circuit for "ground tuning" made quite a difference in signal strength..
1) The 50 ohm antenna input goes to a 10 turn winding on the FIRST of 2 ferrite antenna cores. The other end of the 10 t winding goes, not straight to ground, but through a series cap to gnd. The cap value is .0047uF, which with an inductor value of about 10 uH, resonates in the AM broadcast band.
2) The secondary on the first antenna core is about 60 t. The top end of it connects to the 365 pF "GND Tuning" variable cap, whose other end connects to circuit ground. The bottom end of the 60 t winding goes to EARTH ground. So the first core and first variable cap are a series L-C tuner that minimizes the impedance between receiver circuit ground, and earth ground. NOTE: The braid of the coax connecting between the One-R-Flex receiver and antenna, is connected to circuit ground/chassis ground at the receiver, but is NOT connected to any ground at the 9:1 transformer box at the base of the Ewe antenna - it is left 'floating' at the antenna end. To be very clear, chassis grnd = circuit gnd; EARTH gnd is separate, and the "gnd tuner" series L-C circuit tunes to resonance between them.
3) The 1st ferrite rod core is positioned vertically [using candle wax to cement it] approximately 2" apart from the 2nd ferrite rod [also mounted vertically], which with the "MAIN Tuning" cap [also 365 pF], forms the L-C tank connected to the transistor. So the only way antenna signals get into the Reflex/Regen stage is via magnetic link coupling. I find that this improves selectivity and helps with out of band rejection, after having tried several versions of coupling the antenna signal to the 2nd [main] ferrite core or to the LC through a small cap. Signal strength is not impaired by leaving the antenna connection at the 1st LC circuit, and having no direct connection to the 2nd LC. The 2" spacing between ferrite rods was determined by trial and error. Too close, and the tuning changes, along with introduction of overload and too many overlapping stations, especially at night and at the upper end of the band.
4) Giving the receiver a tunable earth ground enables maximum sensitivity and signal strength. With the Reflex circuit in combination with Regeneration, strong stations come in so loud that the volume control on the Amplified Speaker is no more than 1/4 way up from the bottom of the pot.
NOTE! [To anyone who may be wondering about the somewhat unconventional connection of the 2nd ferrite rod coil to the emitter/47 ohm resistor: The circuit will not oscillate at all if you don't do this! I reckoned that needing RF feedback to the emitter, an emitter resistor of 47 ohms just about matches the antenna impedance coming in. 'Almost' ground, but still feeding the low-impedance point at the transistor's emitter. This is somewhat like the 7 pF "feedback" capacitor seen on my VHF Aircraft receiver, except the whole L-C tank is put across the collector-emitter path. "C" [the Tuning cap] is connected to ground, "L" is 'almost' grounded but still feeding RF back to the emitter, and they appear as a parallel LC tank to the transistor. Works very well, and avoids the need to wind a low-impedance secondary on the coil, for the transistor base, which is seen in other Reflex designs.]
Here is a diagram of a previous iteration of the circuit, with blue lines and arrows showing the RF feedback path, and red lines and arrows showing the Audio Reflex feedback path. Can you see why the 47 ohm emitter resistor does not degrade the circuit's operation, but instead enables it?
5) Regen is abrupt but fairly smooth-- i.e, it appears suddenly, but does not "pop" into oscillation -- there is a small but smooth transition into oscillation.
6) When peaking a signal with the GND Tuning cap, back the Regen feedback off until it's below the point of oscillation, but still weakly audible. Now peak the signal with the GND Tuner. Now turn the Regen back up, and touch up the Fine Tuning pot to get the station locked and centered. If you attempt to "peak" an already-oscillating detector, you may find that when the GND Tuner hits resonance, the oscillating detector is STOPPED and you're below the point of oscillation. It's a little harder to find the best 'peak' point on the GND Tuning cap unless you start out below the oscillation threshold and peak your GND Tuning there. You will get a 'feel' for this as you play with the receiver.
7) Once just into oscillation, the receiver will 'sync' or lock onto the AM carrier, and once adjusted, will not drift noticeably for hours at a time, if the station is strong enough. Weak stations may require touch-up with the Fine Tuning pot once in a while.
8) The audio quality advances from "mellow/sea-shell" filtering quality, to bright, high-fidelity audio once into oscillation. Advancing beyond that point increases the "tinniness", and the bass [which can be heard through a decent stereo system speaker] begins to disappear. Adjacent-channel interference also increases once the audio expands into "hi-fi".
9) Next step, adding bandswitching. At least 160 meters, if not 80 also, would be nice.