6 Valve Regenerative Receiver with Automatic Regeneration Control

This receiver is a reincarnation of a set I built in 1984, which was actually my first homemade valve radio. The design is based on Radio & Hobbies' "Three Band Two" of May 1957, later reprinted in October 1966. This project was a broadcast and shortwave receiver with switched coils, using a 6BL8 or 6U8 as a regenerative detector (pentode) and output stage (triode). Following issues added a 6AQ5 output valve for improved performance, with the 6BL8 triode converted to a voltage amplifier. Rectifier was a 6X4.
In September 1983, Electronics Australia featured the design again, as to explore the possibilities of building a valve project in the 1980's. That article was rather unenthusiastic and anyone reading it today will notice the author having a general dislike of valves.
When I was attempting to build a 27MHz CB transceiver back in 1984 , I chose this circuit for the receiving section as coil data was available for up to 30Mc/s. Incidentally, a regenerative receiver can be used for SSB and CW reception, so is ideal for shortwave use.
It certainly did receive CB transmissions, and in view of its performance, I wound a coil for the broadcast band and never looked back.
Instead of using a 6AQ5, I used a 6BM8 pentode as I had many of them. The 6BM8 was also to function as the transmitter modulator.
However, the transmitter part never eventuated and the project became a broadcast receiver only. Seeing as the triode of the 6BM8 was doing nothing, I made it into a further voltage amplification stage.
My set had a lot of use, and in fact was my main receiver for a number of years. I added an automatic regeneration control, which worked very well, a vibrator power supply for battery operation (it once spent a day out in a boat), and an RF amplifier to allow use with a car radio aerial (the spin off from this was a separate car radio project in 1985). Later additions were a tape recording output, negative feedback, and in its later days, provided the test bed for FM receiver additions, initially with a Fremodyne tuner, and then simpler straight super regenerative circuits.
Finally, I experimented with a hybrid audio amplifier, using a valve to drive a solid state output stage.
Eventually around 1990 I pulled the whole thing apart as the chassis was now full of holes and the effects odd numerous experiments. I had lost interest in AM radio, and my super regenerative FM receivers were not performing properly.

The circuit of the reincarnation is largely the same as my original.

As the years went by my interest in AM came back, and so did my sentimental attachment to the receiver I learnt so much from. I decided to rebuild the receiver on a new chassis, as close as possible to my original design. I decided against the FM add on, although my present super regenerative tuners work well. As I already have this receiver in 12V form when I built it for car use, I decided not to include the vibrator supply either. The rebuild was to incorporate automatic regeneration control, but this time using valves. An RF amplifier would also be included to eliminate aerial loading effects.

The Regenerative Detector
Everything is completely conventional. The beauty of the design is that it uses some of the most common Australian TV valves. The 6BL8/ECF80 was initially designed for VHF TV tuner use, with the triode for the local oscillator, and the pentode as the mixer. The 6BM8/ECL82 has been discussed elsewhere, and was common in field deflection and audio use.
The pentode in this circuit functions as a grid leak detector, with regeneration control by screen voltage. The aerial coil is tapped for feedback.
Potentiometer control of regeneration was one of the factors that got my interest in the original design. As it turned out, this method of control is the smoothest with no backlash or detuning, as happens with variable condenser feedback. My original receiver had one aerial terminal which worked satisfactorily with a 10m long wire aerial in Sydney at the time. Now that I have a much longer aerial with peculiar resonance and loading effects, I decided the new receiver would be a bit more flexible. So, a simple untuned RF stage was added.
Not only does this mean the regeneration control is not dependent on the aerial, some extra gain is also obtained.
The triode of the 6BL8 is a straightforward audio amplifier stage. The cathode is not bypassed as there is sufficient gain over the various stages, and it does create some local negative feedback which improves sound quality. The unbypassed cathode is the source of low impedance line level signals, and I did in fact use it to feed a tape recorder in the original set. If you wish to do this, connect the cathode via a .47uF condenser to the audio output connection. Other valves which have the same pin connections and work with no changes are 6U8/ECF82, 6EA8, 6CQ8, 6JW8/ECF802. However, that is not to say they are interchangeable in other circuits. Needless to say, for European constructors, the series heater equivalents might be easier to get. 9A8/PCF80, 9U8/PCF82, 9JW8/PCF802 for the RF and 16A8/PCL82, 50BM8/UCL82 for the audio. Of course, separate similar triodes and pentodes can be used if more convenient.

       550Kc/s-2.2Mc/s                          2.2Mc/s-12Mc/s                         8Mc/s-30Mc/s
Coil data for 550Kc/s-30Mc/s. Frequency coverage is with a 10-415pF variable condenser. All coils 3/4" (20mm) diameter. The Radio & Hobbies design used a rotary switch to select coils, but pointed out that absorbtion effects could occur depending how close the coils were to each other. It was suggested that the two unused coils be shorted out with extra switch poles to prevent this. My coil former is a piece of 32mm conduit attached to an octal plug. Coil winding data was adjusted to suit the wider diameter.

Audio Amplifier
The audio amplifier is discussed here so won't go into elaborate detail. Suffice to say, this was the first project I'd used an M1100 line transformer as sold by Dick Smith, and was delighted with its performance. I have made some slight changes though. As some 6BM8's suffer from grid emission, I have reduced the pentode grid resistor to 220K. It is true that this would reduce gain compared to the usual 470K, but the reduction is small, and there is plenty of audio gain over the whole circuit. As M1100's are no longer available, I have used a new old stock Rola E14 speaker transformer. Although its primary impedance is 7K instead of 5K the difference is unimportant in this application. As the E14 had only a 15R secondary, I unwound it and brought out an 8R tapping. The negative feedback circuit is the same.
With the new receiver, I also decided to elaborate on the headphone socket. Previously, the headphones would connect straight to the speaker transformer secondary and switch out the speaker in the usual way. However, as Jaycar now have a headphone socket with DPDT switching, I decided to use this to improve the matching when headphones are in use. As most headphones are 32R per channel these days, a 10R resistor loads the secondary when the phones are plugged in. Additionally, a 33R is in series with the phones to reduce the possibility of overloading them.

Power Supply
The power supply is different in my new set. For the original, I used a transformer from an Astor radiogram. 1N4007 diodes were used to rectify the high tension. My new transformer has a 5V 2A winding so couldn't resist using a valve rectifier. I used a 5Y3.
It also provides 225V aside at 60mA and has a 6.3V 3A heater winding. (Yes, I could have used a 6.3V heater rectifier in the original but never did).
Like most projects, I used resistive filtering of the B+. This works and is more convenient than a choke. The 800R 5W resistor, and the two 16uF filter condensers are from the original set (as is the tuning gang). The original Astor power transformer, filter capacitors, fuse and mains lead ended up in my first pulse counting FM  receiver. Note that I've put the 800R resistor in the negative line. Electrically, it works the same as when put in the positive line as is conventionally done, but doing it this way creates a source of negative voltage which may have been necessary for various forms of automatic regeneration control I was experimenting with.
As the DC output from the transformer and valve rectifier are lower than the original set, the two feed resistors for the audio and RF stages have been reduced in value. They were previously 3.9K 5W and 47K 1W respectively.

RF Amplifier
This is simpy a 6AK5 pentode with the primary of the aerial coil in the plate circuit. Input is untuned, with the 100pF and 47K performing as a simple high pass filter, as well as the usual grid leak and DC isolation functions. The high pass filter is desirable to prevent the 6AK5 being overdriven by 50 cycle mains picked up on a long wire aerial. This could cause modulation hum. The low plate voltage might seem strange, but increasing it beyond around 30V did not increase gain. In fact, many valves work perfectly well at low voltages and the automatic assumption one needs to start out with 250V is often a needless waste of power and valve life. Gain of the receiver is higher with this RF amplifier, and selectivity is also improved because the aerial is no longer loading the tuned circuit. However, with this configuration gain is not high enough to use something as short as a a car radio aerial.

Automatic Regeneration Control
For an understanding of how this works, see this article first. Basically, the idea is that when regeneration is excessive, the detector puts out an abnormally large signal. The level of this is then used to back off the regeneration to the point where it is below the point of oscillation. In this particular receiver, regeneration is controlled by the screen grid voltage of the 6BL8 pentode. For manual adjustment, the 250K pot controls the voltage in the usual way.
The audio signal from the 6BL8 triode is further amplified by half a 12AX7 triode which drives a 6AL5 voltage doubler. The output of this is the DC control voltage and becomes more negative with increasing signal strength. The other 12AX7 triode then controls the B+ level fed to the 6BL8 screen grid.
Let us now look at the operation of this circuit in detail. Audio is picked off from the volume control as here the voltage is at its highest and not affected by the volume control setting.
The first half of the 12AX7 triode operates as a further audio amplifier. Only for convenience did I use a cathode rheostat to control gain. A conventional volume control of 500K to 2M could have been placed in parallel with the normal volume control instead. The triode uses contact bias for simplicity, but of course cathode bias could also have been used.
From the plate, the signal is now strong enough to drive the voltage doubling rectifier. The second 12AX7 triode is not used as a cathode follower, as it may seem at first glance, but the control voltage is used to control its conduction. With no control voltage, the triode has no bias and passes full current. In this situation the 6BL8 screen grid receives full voltage.
With an excessive control voltage, the 12AX7 is cut off and the 6BL8 receives no screen grid voltage. The control voltage is normally between these two extremes.
Note that the voltage doubler input cannot be referenced directly to earth because it has to float at the 12AX7 cathode, and thus 6BL8 screen grid, voltage. Nevertheless, it still receives full input signal because the 330K, and when the ARC is selected, the  .22uF, are in the return path. The time constant is formed by the 10M and 5uF. With the coil I used in this set, oscillation occurs around 10V on the 6BL8 screen grid.
When the receiver is not tuned to a station, it weakly oscillates as a result of the ARC negative feedback. Thus, the detector is operating at full gain. When a station is tuned in, the strong audio signal takes over, developing a higher control voltage which then backs off the regeneration. The receiver now operates below the oscillation point, thus providing a clear signal. The ARC will also compensate for fading to a useful degree. As signal strength increases, so does the audio signal, thus reducing gain. And, vice versa, although once weak enough, the set will start weakly oscillating until signal strength picks up again.
To set the automatic regeneration, the set is tuned to a weak station with the manual control. Then, switched to ARC, the ARC set control is slowly advanced from minimum until oscillation just stops. A voltmeter; preferably of reasonably high impedance, and analog, can be connected across the 330K to function as a crude tuning meter.