1928 Regen Detector + One Step Audio

1928 Regen Detector + One Step Audio


Many shacks of the 20s and 30s listed the receiver as a regen with one step of audio. Pictured is the one I built based on common constuction/design practices of that period. This particular receiver is closest to one described by Ross Hull in the November 1928 issue of QST. James Lamb wrote about a similar design in the April 1928 issue of QST.


This basic receiver consists of a triode regenerative detector transformer coupled to one stage of audio.

The character of this radio is a set by the Air King Short Wave Coils. These three coils with a 140mmfd capacitor cover 80, 40 and 20 meters. This coil set has a movable antenna coil mounted on the base/socket but each coil includes it's own tickler winding.

C1 is the main tuning control with C2 acting as bandspread. Three methods of bandspread were typical in the 20s. One was two capacitors in series, the second was a capacitor connected to a tap on the coil and, finally, parallel capacitors. The first two, with the right coils, allows each ham band to be spread across the entire dial. The third works for any coil but bandspread will not be optimum for each band. Given my "store bought" coils that I did not want to modify, I opted to use the parallel capacitor method of bandspread. 80 meters covers the entire dial. Other bands are squeezed down a little.

R3 controls regeneration by controlling the plate voltage of the detector. Controlled regeneration requires a way to change the feedback and/or gain of the regenerative detector. Most receivers of this period used either variable coupling between the tickler/plate coil and the grid/tuning coil or a "trottle" capacitor between the tickler coil and ground or variable plate voltage. James Lamb and Ross Hull, both respected receiver experts, recommended varying the plate voltage on short wave regenerative detectors because this method had the least impact on tuning. C4 eliminates any tendency of scratching noises as R3 is changed.

The audio stage is fairly standard. R5 is shunted across the secondary of the audio transformer to control fringe (or regeneration) howl. R4 contols volume by varying the filament voltage on the audio stage tube. By putting R4 in the ground side of the filament, it also provides a little C bias to improve audio quality. The headphones are shown in the plate circuit. This reflects the vintage design point but it is NOT the way to use this receiver today. 100 plus volts through headphones clamped to one's head is not recommended. See my operation notes below for a safer way to lisen to this set.


Table 1 - Parts List
R1 2M Grid Leak
R2 3 ohm 1 Watt
R3 1K Regeneration Rheostat
R4 20ohm Filament Rheostat
R5 1Meg
C1 140mmfd Bandset
C2 25mmfd Bandspread
C3 100 mmfd
C4 1mfd
C5 .002mfd
T1 3:1 Interstage Xfmr
V1 '01 Triode
V2 '01 Triode
L1 Short Wave Coil, see text

My receiver is 12" wide and 10" deep including the 2" gap behind the front panel. The chassis and front panel are both of 1/4" oak plywood. The runners are each 3/4" x 1" x 10" pieces. All were heavily stained and varnished. The front panel itself is 6" high. R3 is mounted high in the middle of the front panel while R4 is mounted low. In order to minimize the effects of hand capacity, the front panel is backed by a sheet of copper. Note that many older pots and rheostats like what might be used for R3 and R4 were built with the arm connected to a metal case or shaft. This may not matter for hard rubber or bakelite panels but, if mounted incorrectly, R3 could short B+ to ground. To avoid a short, cut an extra large hole in the copper shield where R3 will be mounted and use an insulted knob. The case of R3 should not contact the copper hand capacity shield. Keep this in mind for R4 also. When laying out the front panal remember that many 20's capacitors are not mechanically symetric. The C1 or C2 tuning shaft may not be at the center of the capacitor frame and, at minimum capacity, the rotor may extend quite a bit out to one side. R3, R4, C1 and C2 must be mounted so that nothing interferes behind the front panal and all controls look symetric from the front.

Lay out the receiver chassis trying to keep leads, especially those around L1, C1 and C2, as short and as direct as possible. Much will depend on the particular coil set and socket that is used. I ended up with both tube sockets oriented with their filament connections away from the coil. C4, R2 and most of the power wiring is under the chassis.

Coils may be either vintage manufactured such as my Air King coils or homebrew. Homebrew coils can be wound on a variety of forms. One popular option was four pin tube bases. The information in Table 2 can be used as a starting point when using tube bases as coil forms.

Table 2 - Coil Information
Band Grid/Tuned Coil Turns Tickler/Plate Coil Turns
160 49 8
80 33 7
40 15 6
20 6 5

Both windings are closewound of 24 gauge enamel insulated wire. Wind the tickler winding at the bottom of the form and slightly spaced away from the grounded/bottom end of the grid/tuned coil. Both windings must be wound in the same direction. The number of grid/tuned coil turns should be adjusted so that the band of interest falls near maximum capacity. With only a four pin coil form, no pins are available for the antenna link to be part of the plugin coil. Instead, the antenna link may a few turns of wire wound in a coil that loosely slips over the tube base coil form. Another option is to eliminate the antenna coil all together and couple the antenna to the top of the grid/tuned coil (ie. C1/C2/C3/R1 junction) via a very small capacitor (around 5 mmfd).


For safety, headphones should not be connectected directly into the plate circuit of V2. Instead, transformer couple headphones or another amplifier into the circuit at this point. This set will not drive a speaker directly.

V1 and V2 filaments require 5VDC @ 250 mA each. R2 and R4 allow a little flexibility in the voltage applied to the set but do not exceed 5VDC at the tube sockets for any length of time.

The detector B+ voltage should be adjusted so that the set goes into regeneration when R3 is set to about half scale. For my set this was 45VDC. This varies according to the gain of V1 and the size of the L1 tickler winding.The audio stage B+ may be any between 45 and 90 volts.

After hooking up antenna, ground, power and "headphones", operation is typical of most regenerative receivers. Use a known signal to set the bandset capacitor so that the ham band of interest is covered by the bandspread capacitor. Advance the regeneration control until the set just hisses into regeneration. This is best setting for listening to CW. Strong SSB signals can be copied best if the regeneration is advanced a bit further. Once you are familar with the set, build calibration charts overing any bands of interest.


Plenty of 80 meter and SW broadcast signals were copied with this set. It worked, but with a few limitations. The regeneration setting changed with frequency. As I tuned across the band, I needed to tweak the regeneration control. I also found that varying the regeneration setting varied the frequency slightly. This may have not been noticable in 1928 but it did effect handling on modern SSB signals. I found that I used the regen setting as a final "clarifier" to tune in a QSO. This simple set does overload easily. One possible improvement might be a small variable capacitor in the antenna lead. This would allow stong signals to be attenuated before getting to the detector. Tunable hum is a major problem with this set. Above about 6MHz all that can be heard is 60 cycle hum. Better shielding including a metal chassis might have helped this problem.

This is a neat set but my use of it on today's bands will be pretty limited. Touchy regeneration control and hum wiping out 40 meters are the biggest limitations. I'll probably use it for general SW listening and a little on 80 CW with my TNT transmitter or 1928 Ross Hull Hartley.

Click here to view high resolution pictures of this receiver.

Back to W0VLZ's Ham Radio Page