My FET regenerative radio project.  Notice the really neat little tuning dial.  Boy, they sure don't make nice ones like this anymore.  It just begged me to make radio using it.  Using a shortwave receiver, I listened for the regenerative signal when the unit was in oscillation and by matching my dial to the radio's dial, I was able to put calibration marks on the dial face (not shown in this photograph).

My Armstrong FET shortwave radio. This radio works well and it's something like this I originally wanted to build.

     What a huge difference between the two configurations.  This particular design is stable, it sounds great and has much more volume than the other design.  This radio really works well and I am extremely pleased with it.  I have it adjusted to cover from 3.5 mHz to 8 mHz (or megacycles, as we used to say) and the tuning and regeneration control is smooth.  The RF stage doesn't pull the detector and cause all kinds of unwanted regeneration effects like the NPN version did and even the volume control is nice and smooth.  

    Although I am using a really nice, very old dial with precision gear reduction, the tuning is so sharp that it is hard to get it right on frequency without a lot of trouble, especially on SSB.  To overcome the sharpness of the tuning, I installed a really excellent little fine tuning control which uses an ordinary 1N4007 rectifier diode as a variactor.  By the way, I built the fine tuning control on the potentiometer itself so that none of its components are mounted on the circuit board.  

 The circuit elements are as follows:
1)  The RF stage, which uses a grounded base amplifier.  The signal level goes smoothly from loud to soft as the RF Gain Control potentiometer's goes through its travel.  I had expected a lot less linearity, so this control works better than I would have believed.  
2) The Armstrong regenerative detector with its tickler coil and throttle is, of course, the heart of the project. The way it is configured, it is easy to tune, stable and fun to use.  
3) A LM386 0.25 watt audio amplifier that is configured for a voltage gain of 200 (which is its maximum).  As mentioned, it is arranged to be fed with a 100 K ohm potentiometer, which I've never seen done this way before, but seems to work just fine.  I have it connected to an internal 3 inch speaker and it really pumps out the sound.

     You may have noticed that there is a subtle difference between the way this FET is configured and the way a thermionic triode tube (like my 1H4) is configured.  After Armstrong figured out how the Audion tube actually worked (by thermionic emission of electrons and not by positive ions), he discovered that the grid would self bias and become negative due to picking up electrons from the stream that left the hot cathode on its way to the anode (plate).  All he had to do was select a "grid tickler" resistor of sufficient value (like 1 or 2 megohms) that would allow the tube to self bias without becoming too negative.  Too much resistance or an open circuit and the tube would self bias to cutoff, but too little resistance and the bias would be too small and the tube conduct too heavily and would be at the wrong part of its conduction curve for amplification.

     With the FET, the equivalent "gate tickler" resistor shown will not have the same self biasing effect since there is no way for the gate to pick up passing electrons to become self biased.  To bias the FET properly (make the gate "negative" with respect to the source (or cathode) and thus set the current through the FET to the right level), it is necessary to have a bias network between the source and ground.  Biasing is accomplished by simply using a 10K resistor and a bypass 0.001 MFD capacitor.  It just so happens that this is also a really excellent place to tap off the audio too.  (By the way, you might want to experiment with a resistor between 5K and 10K in this circuit and determine for yourself what value works best when listening to SSB)     

     When listening to AM broadcast, the throttle should set the regeneration just slightly below the threshold  of oscillation.  At this level the FET is working at the correct bias, current flow through the FET is proper and everything is good.   When the FET goes into hard oscillation, as happens as you tune to a higher frequency or when listening to Single Side Band (SSB) or Morse Code (CW), the FET conducts way too much and goes into a kind of run-away.  The current through it is all wrong and it starts acting squirrelly and if the current rises too much and the voltage produced at the Zener diode looses regulation, it gets really, really squirrelly.

     A self biased tube does not experience run-away under these conditions because the more it conducts, the more electrons the grid picks up and the more negative it biases itself.  Because the gate of an FET can't pick up extra electrons and become increasingly negative like that, putting a high speed diode (the 1N914 shown in the schematic) in the gate circuit insures that extra negative bias will be automatically produced to keep the FET under control.  With the diode arranged thus in the circuit, it allows negative bias to rise above the set value as oscillations gain in amplitude and thus the FET's current and operating parameters are kept where they should be.  This diode works great to  " de-squirrel  "  the operation of the radio when listening to SSB or CW signals and I recommend putting one in there.  Of course, there are lots of diodes out there, but be sure you use a fast switching type in this circuit.

     At least, that's what I think the 1N914 diode is doing to the gate circuit.  I'm sure the smart guys who really know their theory will soon be setting me right about all this, so stay tuned for a better explanation why this diode in the gate circuit works so well.

     To protect all the circuits, I'm using a 1N914 diode to block reverse voltages in case somebody tries to put the battery  in backwards while the set is turned on.  Yes, the connections on top are polarized and normally you wouldn't have the set turned on when changing batteries, but I like to have my stuff fool-proof whenever possible.  Why a 1N914 diode?     I have a lot of 1N914s and they will easily handle the current, but any diode will work just as well.  You have a lot of 1N4007 or 1N4001s?  Yeah, they will work fine too.

     Finally, I left the high impedance audio control potentiometer alone since it works smoothly throughout its range and the audio level in this configuration sounds just as loud as when using a more conventional 10 K potentiometer.  I discovered that there was some kind of audio oscillation that started when the potentiometer was set all the way open, but that a small value resistor or choke coil easily eliminated it.  I guess I should have added some RF filtering here, but a simple resistor in line with the potentiometer's high side works just fine.

      As a future addition,  I want to experiment with a band switch to short out some coil windings so I can extend the range of the set up to 15 mHz or so.  I think it would be nice to have two bands, 3.5 to 8 and 8 to 15 that would be selected by a simple toggle switch.



    Well, in words and pictures, that is the story of  my latest regenerative radio project.  If you are interested in building a similar radio, I suggest you follow Mr. Kitchen's advanced regenerative radio design but strictly avoid his "Beginner's Scout" design.  It will help a lot if you can get the right circuit board for the project, but if you are careful and use those stick-on strips and pads, a perforated circuit board will work wonderfully for you.  I have recently built several BFO's and HF buffer amplifiers using perf boards and stick-on pads and they work just great and look good too.  If you have to, you can use the "dead bug" method where you use globs of solder to connect everything and it will work well at these frequencies too, only just don't show anybody the insides of your radio.  It works, but god is it ugly!

     I'd like to thank the long deceased Edwin Armstrong for the many invaluable contributions he made to understanding electronic devices and circuits and for formulating the ideas behind most of the neat stuff I've built over the years.  I'm sorry you lost your temper, hit your wife and jumped out of a window to your death, but I've been screwed by evil villains like David Sarnoff myself and I think I know a little of how you must have felt.  If it means anything now, I too acknowledge the fact the U.S. Supreme Court was wrong, you were right, Lee Deforest was a fraud and he received way too much credit for things he never even understood.  I'm sorry, but RCA was right about the need for FM to move up to the VHF high band.  The truth is, some of those young engineers actually understood your invention better than you did.  It happens all the time in high technology where us old guys get beat out by the whippersnappers and have to take an early retirement.  If I could, I would like to tell you about what I have discovered regarding money, recognition and happiness. What I've found is that there is no hateful person and certainly there is no amount of money or fame that is worth sacrificing your happiness, the happiness of those you love and especially your life for.  Greedy, ruthless, scheming, manipulative bastards like Sarnoff may cheat you out of your money, the recognition you deserve and your plans for a better world, but they can take your happiness only if you give them the power to take it.  We should never give those people that kind of power.  Screw the money, the fame and the whole sorry lot; it won't buy you anything if you end up killing yourself or if, in your frustration, you hurt the ones you love.  It doesn't take all that much to live comfortably and happily and, in the end, we all finish up dead and forgotten with no way to take it with us anyway.  Naked you came forth from nothingness when you were born and naked into nothingness did you return when you died, not withstanding your accomplishments in life -- as shall we all in the end.  

     Finally, I'd like to apologize to Mr. Charles Kitchen for the mean (but true) things I said about his Beginner's Scout Radio.  However, I'm not taking anything back.  I do want to thank him for all the ideas that I shamelessly ripped off from him.  I also want to acknowledge that most of the schematics I'm presenting here were originally drawn by him and that I only modified them.