The "One Tube FM Tuner" described in Popular Electronics, August 1960 is about the simplest valve receiver for reception of the 88-108Mc/s FM broadcast band.
However after constructing it, I was less than impressed. While it "received" stations, the sensitivity was poor, the aerial loading caused dropouts or changes in oscillation level, and of course the stereo/SCA subcarrier beat problem was quite evident. The quench filtering was virtually non existant. With all these problems the receiver was not really useable. It was vastly inferior to either the Fremodyne or 12AT7 receivers.
As I mentioned in the original article, my plan was to modify the design while keeping with the single 6C4, along the lines of my successful 12AT7 receiver, so that it would be a practical receiver to use. If you have not come from the "One Tube FM Tuner" page, I suggest you familiarise yourself with its design before proceeding here.

New design

  A couple of options were considered in rebuilding this super regenerative receiver. Given that there was already a three gang tuning condenser, the thought of rebuilding as a Fremodyne was enticing. It would also mean not having to mount a pot on the front panel for use as a regeneration control. There was enough room for a nine pin valve socket to accomodate the 12AT7 required, and the power supply suited.
 However, this came unstuck considering the mechanical layout. The original 6C4 receiver had no groundplane as such which would make it awkward with all the VHF earth returns required in the Fremodyne, and nowhere to mount the IF coil. The additional components would be too much of a squeeze to fit in. Of course a new chassis could be made up allowing for proper earthing, additonal tagstrips etc, but it would be too much fiddling getting everything to line up with the existing dial.
  It was thus decided to retain the single 6C4 straight super regenerative concept but modify to something along the lines of the 12AT7 circuit. This would be keeping in with the theme of "simplest possible valve FM receiver", and hardware wise it would still look much the same as the original receiver.
  As I only was going to need one of the 15pF tuning gang sections, I toyed around with the idea of replacing the existing three gang unit with another single gang capacitor, but it would be too difficult to mount. I had a feeling the plywood chassis would end up looking like an unsightly sieve by the end of it, and I didn't want to ruin this receiver's good looks.

The new version of the receiver looks much the same as its predecessor. The preset regeneration control can be seen just behind the front panel.

The New 6C4 receiver

First thing was to pull all the superflous parts out of the old circuit and rebuild according the the 12AT7 design. Of course, with only one triode, there wasn't going to be an RF amplifier stage. The aerial would be coupled as per the original 6C4 circuit; ie. the aerial coupled by a one turn coil. I knew this would not be optimum as the aerial loading varies across the band. With a super regenerative receiver, this causes problems in that while close coupling improves sensitivity, it can cause the receiver to stop oscillating over part of the band. I would simply have to put up with less coupling and the resulting inferior sensitivity. The other problem with this method of coupling is that changes in the aerial will cause the tuning to drift. With a fixed aerial at home this isn't too bad, but it could be awkward for a portable receiver where the aerial is constantly being moved around. Hopefully the other attributes would make up for this.

The same component values were used as per the 12AT7 design, but as I already had the RF choke I decided to use it. Simple quench filtering was by a 120K and 470pF in the audio output lead. Quite a bit of quench is still evident, but to further reduce it starts to noticeably cut off the treble response. This is why an active filter is far better. It has a much sharper cut off characteristic. My mains powered 12AT7 receiver uses such a filter. For regeneration control, I used a pot in the cathode circuit, as per later developments with the 12AT7 receiver. The RF choke is described in the Popular Electronics article for those who want to wind it.

The circuit of the new design is about as simple as you can get for an FM receiver.

Results were good on the initial power up. At this point I hadn't had any aerial connected; this type of super regenerative receiver will work without on the stronger stations well enough. The optimum cathode bias appeared to be about 5.6V. This is more than in the same circuit with the 12AT7, but I am using a different RFC, and the 12AT7 is a higher gain valve.

I did have to raise the quench frequency. Best results were around 30Kc/s. This meant changing the 330K grid resistor to 270K. Only for physical convenience did I connect it across the 33pF grid capacitor rather than from grid to earth.

Regarding the regeneration control, I found that 1uF was quite sufficient bypass for the audio and quench frequencies. This eliminates the time delay factor evident when larger capacitance values are used and the control is adjusted. I used a 10K pot because it was to hand, which was fed from the existing 100K power supply bleeder. A 50K pot can be used equally as well if it is desired to eliminate the need for bleed current. Again, there was no problem with time lag in adjusting the control.

At this point I still had no aerial and this little receiver was doing a very good job. Stations like C91.3 (Campbelltown) and Wave FM (Wollongong) were coming in at entertainment quality. Of course all the mainstream Sydney stations were loud and clear; even 2RES, a low power public station could be heard as could 2ST from Bowral. What I suddenly realised is that in tuning across the band, I hadn't been adjusting the regeneration control. It always seemed to be in the right position. Very interesting, I thought. Well, looks like I won't have to drill the front panel and install a pot!

Looking into this "automatic" regeneration control, my theory is that now cathode bias is used, it is providing DC feedback, constant plate current and therefore constant oscillator amplitude. Indeed, the measured plate current did stay at a constant 560uA across the band. Further thought for future receivers. As before, the regeneration is adjusted for the loudest and most sensitive point of receiver operation (which occurs just before oscillation drops out). A slight adjustment is usually required to then miminise the stereo/SCA subcarrier beat.

The output waveform on a strong station. Each division on the CRO is 50mV, so it can be seen there's sufficient level for a two stage audio amplifier. The thickness of the waveform is what remains of the quench frequency.

Final part of construction was to make up a bracket for the pot and optimise the aerial coupling. As expected, ideal aerial coupling did make the reciever drop out at the high end of the band and there was a dip in the low end where it momentarily dropped out. This was clearly visible on a spectrum analyser. So, the coupling has to be very loose to reduce this effect. Thus the receiver is desensitised as far as aerial signal goes, but as there's considerable direct pick up from the oscillator coil, it isn't as bad as might be thought. Most testing and design of this receiver has been without an aerial, and realistically, if you only want the stronger stations it will work without.

Setting up

Without the aerial primary coil in place, ie. just the oscillator coil in circuit, turn on the receiver and see that it oscillates. Keep the cathode voltage at minimum for now. A hissing sound should be heard and stations should be receivable. Now, increase the cathode voltage to the point where oscillation (and general best performance) is maintaned across the band. A slight touch up might be needed to reduce the stereo/SCA interference.
Then it's time to install the aerial primary coil and connect the aerial. Space the primary coil away (start with about 1/4") from the secondary (oscillator coil) and try the receiver. You'll probably find oscillation drops over part of the band. Reduce the cathode voltage slightly which should improve this. If the drop outs are still severe, then space the two coils apart further.
Eventually you'll have to compromise between aerial coupling and the amount of regeneration. As regeneration is increased to compensate for close aerial coupling, the sensitivity and audio output will be reduced.
If you have the regeneration control user adjustable, it will allow a closer aerial coupling. Remember, maxiumum sensitivity and audio output are obtained just before the receiver stops oscillating.

How does it perform compared to the original Popular Electronics/RCA design? The performance is much more consistent across the band, the quench filtering is better, and it is much more sensitive. And of course you only need a single gang tuning condenser. Certainly as one's first super regenerative receiver I highly recommend the new design.
Having said that, if you need maximum sensitivity, the original 12AT7 receiver is preferable.

Further Ideas
If it is going to be used by a non technical person, then leave the regeneration preset control inside the enclosure. That way it will tune like any "normal" radio receiver.
However, for those who are more adventurous, I would recommend having the regeneration control on the front panel. It will allow the receiver to accomodate the effects of aerial loading, and you'll be able to have a much closer aerial coupling, resutling in better sensitivity.
On the subject of aerial coupling, other ways to connect an unbalanced aerial into the circuit are through a low value condenser (a few pF) to either the cathode of the 6C4, the top of the oscillator coil, the grid of the 6C4 or to a tapping on the oscillator coil.
These will give a much tighter coupling, but again loading effects will be evident. You will need to experiment for what is best in your situation. I would start with about 2pF of capacitance. I have ideas of using this method for a mobile receiver where the aerial and transmission line are fixed items.

May 2006 update:
I've now modified the receiver for capacitive aerial coupling as it's easier to adjust and I suspect less critical with the aerial used.

As can be seen, it's simply a loop of insulated wire near the hot end of the aerial coil. I've also changed the aerial input for unbalanced transmission line. Sensitivity is good and there's no dropouts tuning through the band.
The 6C4 triode is the same as that used in a 12AU7, so if you don't have a 6C4, you can use a 12AU7 with the second half for an audio or RF stage if you wish, or even as the rectifier. I have actually tested a 12AU7 in this set by means of a crude 7 to 9 pin adapter, and it does work the same as the 6C4.  Incidentally, if you're only going to use one of the 12AU7 triodes, don't connect the heater section that isn't going to be used...it will poison the cathode as there'll be no plate current.
In case someone is wondering, will a 12AU7 work in the 12AT7 circuits? I'm working on that now as having drawn so heavily on my 12AT7 stocks, I'm looking for substitutes. So far it seems that if the 12AU7 can be made to oscillate over the band it works as well as the 12AT7. I'll provide more info on that once I've got it to work...or not. The RF amp side of things works perfectly with the 12AU7 at least.
An important point which has come to light is the RFC. The 15uH choke used in the 12AT7 receivers will not work with the 6C4 or 12AU7. You must wind the choke according to the instructions. I used a 6.2mm piece of plastic rod with 75cm of 25 B&S (.55mm) enamelled copper wire. The 75cm is a quarter wavelength. Presumably other wire gauges could be used provided the length is the same.

home
 

email me: cablehack at yahoo dot com