The R-392 receiver is hot, stable, and provides full coverage between 0.5 and 30 MHz. However, the compromises involved in modifying it for 28 Vde on filaments and plates need shaping up. Here are some suggestions.
First bring out the plate and filament leads separately. The receiver is happier with somewhat more than 24 volts on the plates, somewhere between 30 and 35 volts. Build a supply that will provide over 30V from a bridge rectifier. Filter the supply with 3000 uF or more (the receiver tolerates 2 to 4 V of hum). Power the entire receiver this way first. The tubes in my set withstood 37 V for a few hours. Next add a dropping resistor (2 ohms, 20 watts) to the filament lead. Leave the plate lead on the higher voltage.
The audio output tube has to go. This tube pulls a total of 1.3 watts, plate and filament. The designers must have known it had to go: the manual shows a plug-in transistor substitute. However, their substitute uses 15 components including four obsolete transistors and two custom transformers. A better substitute is shown in fig. 4.
An improved circuit is shown in fig. 5.
Next we operate on the detectors, V602, V603 are 12AU7s with filaments in series. They are used as four diodes for detectors and agc. Pull out V603 and insert two germanium diodes into the socket holes, one with anode to pin 1 and cathode to pin 3 (see fig. 6), and the other with anode to pin 6 and cathode to pin 8.
You can now increase the power supply dropping resistor to about 8 ohms, and the filament string will have decreased from 3 to just over 2 amps. The radio has now become practical, and you haven't had to take the chassis apart. The next step should involve fet substitutes, a very practical possibility with only 30 volts on the plate bus.
N. J. Thompson, KH6FOX
from Ham Radio Magazine, April 1976, page 65