homebrew single transistor regen (with a stupid resonance question)

Discussion in 'Homebrew and Kit Projects' started by KX6MWS, Sep 12, 2011.

Thread Status:
Not open for further replies.
ad: L-HROutlet
ad: l-assoc
ad: L-rfparts
ad: l-rl
ad: l-innov
ad: l-Waters
ad: l-gcopper
  1. KX6MWS

    KX6MWS Ham Member QRZ Page

    After looking at several regenerative radio circuits, I decided to build one. The goal is to end up with something that can listen at shortwave and 80/40/20 or maybe even 10 meters. Before venturing into HF, I decided to make an AM broadcast receiver, because its lower frequency and a much stronger signal: it's easier to tell when the thing is working.

    My question revolves around the tank circuit. It's an inductor and capacitor in parallel. Often, I see the output of the tank circuit connects to the base of a transistor, and the base has to be biased. To keep the base bias voltage from shorting through the inductor, a capacitor is placed between the tank and the base. It looks like this:

    (some one else's design) http://cool386.tripod.com/6tr/srrx.html

    I'm made a single transistor version, so I just have the 'core' circuit centered around Q2. L1, C3 and C4 make up the tank circuit, and choosing appropriate values for these for different bands is simple math. In fact, I have it working just fine in the AM broadcast band, just by using a much larger inductor and a larger capacitor. My question, is does the value of C5 contribute to the tank circuit?

    I originally thought not, it's just to keep the bias DC from shorting through the coil. I tried different large values, 1 uF, 10 uF, makes no difference. With values like 1uF, the DC is blocked, but any AC goes through just fine. But when I use smaller values (100 , 50, 20 pf etc) it seems to change the station I'm tuned to by a little bit My question is... why? It's not part of the tank, right? And why does smaller capacitors change it; if this cap is affecting tank resonance, then putting in something huge like 10uF should completely throw me off the band, but it seems the larger this value is, the less it matters. What's going on?

    My other question, is when I push this into oscillation (which is necessary to receive SSB), is the value of C5 going to affect the oscillation frequency? Is there an 'intelligent' way to choose what C5 is?
  2. WA6TLP

    WA6TLP Ham Member QRZ Page

    That's a really neat design for a super-regenerative receiver. I would think you would do well to use the values of capacitors called out.

    I am flatly amazed that you used 1 to 10 microfarad capacitors for coupling. I have never heard of anybody using anything that big to couple HF signals and I wouldn't have believed that the circuit would work even at 200 meters.

    OK, here's my take on it. I'm sure there will be others who know a lot more about interelectrode capacitance who will come in and correct me.

    With a coupling capacitor that is way too high, it's like it just isn't there and you are getting the full effect of the interelectrode capacitance of the transistor. With the more reasonable sized coupling capacitors (in the PF range), the overall capacitance of the tank circuit would be much lower and the interelectrode capacitance of the transistor would only insignificantly contribute to the tank circuit since it would be in series with this small value of capacitance. Thus, as you increase the capacitance from 1 microfarad to (say) 1000 MF, yes you will see a change in the resonant frequency as more and more interelectrode capacitance is coupled over, but after you get much past 1K, I very much suspect that you won't see much change because all the interelectrode capacitance is already coupled and very little is lost by being in series with such a huge coupling capacitor.

    Does that make sense to you?
  3. WA6TLP

    WA6TLP Ham Member QRZ Page

    Just reading over the article referenced, I'm wondering if a super-regen can be made to demodulate SSB? My whole life, I've never worked with a super-regen, so right now I'm thinking that it can't as it doesn't maintain a continuous oscillation to provide the beat frequency. If this radio you are thinking of building doesn't detect SSB or CW, it might be of only limited use to you as a ham receiver -- no?

    Just for laughs, you might read the saga of how I came to build a little regenerative radio that tunes 80 up to 40 meters. It detects AM, SSB and CW quite well and is cheap and easy to build with nothing critical or specialized in its parts list. The story is on my website at http://www.geojohn.org/Radios/MyRadios/Latest/LatestRadio.html
  4. AB9LZ

    AB9LZ Ham Member QRZ Page

    Sure it can, just turn the regen control up until the thing oscillates, it's like having a BFO and detector in one.

    Actually using one to make a contact is another matter ;)

    73 m/4
  5. KX6MWS

    KX6MWS Ham Member QRZ Page

    Interesting, I didn't even consider the parasitic capacitance that comes with the transistor. Putting a small capacitor in series with the base would both perform the DC decoupling I want AND help cancel out some of the transistor's parasitic capacitance. Awesome! Thanks!

    SO back to the 'gigantic' coupling capacitor I used: I mostly just plugged in the large capacitor to a) see what happens and b) prove it was there for what I thought it was there for. Why does it work? Yeah at 1 MFD, it's like its not there at all; any AC in the tank circuit goes right into the base of the transistor. I was using the capacitor for only one purpose: To allow me to bias the base of the transistor to the required .6 'turn on' voltage, and to keep that DC bias from shorting through the tank inductor to ground. I'm at work at the moment, but when I get home I'll put up a schematic of what I have so far. Keep in mind I was using this to receive AM band at about 1Mhz, very different from the original circuit at VHF frequencies.

    I noticed that, for this ~1 Mhz receiver, the coupling capacitor from the tank to the base has to be above 50 pf or so. When I made it smaller, the volume dropped off significantly. 200pf worked well. The original circuit runs at 100+ mhz and used a 4.3pf coupling cap; but I don't think 1 Mhz travels through such a tiny capacitor very well. According to http://www.electronics2000.co.uk/calc/reactance-calculator.php

    100 Mhz, 4.3 pf (original in the circuit) : 408 ohms (strong signal gets through to base!)
    1 Mhz, 4.3pf: 40.8 k Ohms (too weak)
    1 Mhz, 200pf: 795 Ohms (decent signal, sound good)

    It appears that as I lower the frequency of the original circuit, I need to lower the value of the coupling capacitor to let in the lower frequency. But.. what happens is the value of the coupling capacitor is actually affecting the station I'm tuned to. If I tune to a station, then replace the 200pf coupling capacitor, with say 100 pf or 150 pf, the station isn't tuned in anymore. I CAN tune it back in by adjusting the tank, but I'm not happy with just covering up the issue. I'm wondering what the math says the resonance should be. I have the formula for LC resonance in parallel... How does it change when you have: L and C1 in a parallel tank, with a C2 in series ?

    I suppose for the ultimate in tweak-ability, I should just make the coupling capacitor a varcap as well, and let the user play with it. But having main tune, coupling 'tune' and regen (plus volume control) seems like a bit many controls for such a single circuit!
  6. WA6TLP

    WA6TLP Ham Member QRZ Page

    Using a "quenching" circuit is supposed to prevent the SUPER-regenerative radio from going into oscillation, isn't it? Never experimented with a SUPER-regenerative radio, so I really don't know. Certainly, a straight regenerative radio may be "throttled" such that it easily and stably goes into oscillation and thus provides a "direct conversion" where the beat frequency is at an audio frequency and if it is right where the carrier should be, SSB will be demodulated.
  7. WA6TLP

    WA6TLP Ham Member QRZ Page

    If your amplifier is a high impedance device (like an FET) then the impedance of the coupling capacitor may be quite large (small capacitance) and you will still get good coupling. A low impedance amplifier will require more capacitance. This is why I've always liked to work with tubes and FETs. Still, the magnitude of the signal you actually demodulate is not a function of how much signal you couple to the detector, but by how many times the signal is amplified as it is 'regenerated' in the circuit. This is the principle that Armstrong developed in 1914 using early Audion tubes. My suggestion is to pick a capacitor between 50 and 100 PF and go with it. The gain of the radio can then be adjusted with the regeneration control and the audio gain potentiometer. I would think it would be a completely unnecessary complication to try to adapt a variactor diode to this part of the circuit and it very well might make smooth regeneration control difficult or impossible as they would tend to interact.
  8. KB7NRN

    KB7NRN Ham Member QRZ Page

    First, to try a take a VHF super regenerative circuit and convert it to a MF regerative circuit is not the way to go about things.
    There are many simple solid state regenerative circuits out there, why you attempted this is beyound me. I mean no offence, it's just not the way to go about it.

    Hmm... where do I start to expalin that what you have created more then likely is not a regen or super regen at all? From what you described you have a variable L/C hartley circuit connected to a transistor, it's no more then a gorified crystal radio. I question if it's really even oscillating.

    If it does oscillate, you have not described how you control the feed back (regeneration). Q6 was what controlled osciilation in the orginal circuit. And yes, super regens oscillate but the oscillation is quenched at sub-audible freqs as John describes in his article.

    "The emitter of the UJT provides an approximate sawtooth waveform, which as it also provides the bias supply for Q2 takes the detector in and out of oscillation at about 50KHz.
    It is necessary to be able to set the optimum quenching voltage and this is done by adjusting Q6's supply by pot VR2. This effectively functions as the regeneration control."

    So without a regeneration control either your circuit is not oscillating or is oscillating out ouf control. I would suspect the former but I could be wrong.

    To learn more about regenerative and super regerative receivers you might want to join the Yahoo regen group

    They have many schematics in the files section and an overflow files section for vacuum tube and solid state regens. You can also discuss regen theory with some of the most knowlegble people on the subject. Charles Kitchin, N1TEV, (Mr. Regen himself), contributes to the group.

    Hope this helps,
  9. WA6TLP

    WA6TLP Ham Member QRZ Page

    I agree with John (KB7NRN), trying to convert a VHF super regen to use on the HF bands is going about it the hard way and the experiments you have performed to date are highly suspect and (in my mind) of limited worth.

    Before you spend any more time trying to adapt the design you have referenced, you might consider building a radio based on an existing design known to work on HF. Once you have it working, you might try tweaking resistor and capacitor values to try to improve its performance and test your hypothesis' regarding what each component is actually doing in its circuit. I think you will have a lot of fun and will learn a lot doing that. The only thing I would suggest (as I state at length in my article), stay away from Mr. Kitchen's so-called "Scout Radio."

    If you want something simple and cheap to build, uses common, easy to get parts, tunes with smooth regeneration control and that works quite well, I've already linked to my little radio. John's suggestion regarding the chat group is really excellent too if you want to get knowledgeable regarding this subject before actually building something. Of course, I believe that the best way to learn is to just jump in and build something.

    By the way, in my earlier post I said "...as you increase the capacitance from 1 microfarad to (say) 1000 MF..." Boy was that wrong. I meant "...as you increase the capacitance from 1 picofarad to (say) 1000 picofarad..."

    I was wrong when I stated "magnitude of the signal you actually demodulate is not a function of how much signal you couple to the detector." Obviously, the stronger the RF signal, the louder will be the demodulated signal. Coupling can obviously be taken too far and there is an optimum amount. A compromise must be found so that the radio has good 'Q' and the regeneration is smooth and reliable. Most designs I've seen call for something between 5 and 100 PF at HF.

    Anyway, good luck.
    Last edited: Sep 14, 2011
  10. KX6MWS

    KX6MWS Ham Member QRZ Page

    Thanks for everyone's comments and advice!

    Whoops, I'm sorry, I forgot to mention I had a tickler coil coming off of the collector. WHen I wrote my initial post I edited it a bit, and I guess I accidentally deleted that part!

    This is a schematic of what I have right now:


    -When L2 is close to L1, this circuit makes a horrible noise. (oscillation)
    -If L2 is turned around the other way the noise stops (oscillation is suppressed)
    -Adjusting R2 can stop/start the oscillations.
    -If tank circuit is tuned to a particular station, right before R2 oscillates, the station is very loud
    -If R2 is adjusted to that the receiver is oscillating, tuning slightly up/down where the station is seems to cause the station to become high or lower pitched.

    I think it's actually regen-ing properly. Using the same tank circuit I built a crystal set, and I don't get near the amount of sensitivity this is getting.

    When I want a good performing radio I will likely end up choosing a design that's known to work and copy it. I tend to learn most things through trial-and-error. (Though there's a big difference between hardware and software: when a computer crashes, you press reset. When a circuit 'crashes', the magic smoke comes out! )
  11. WA6TLP

    WA6TLP Ham Member QRZ Page

    OK, yes I think you have a Armstrong regenerative radio there, but it certainly isn't designed for optimum performance, IMO. Still, what you are doing is providing yourself with some wonderful and interesting experiments and it's teaching you many things.

    Yes, I suggest you copy an existing HF design when you go to make up a "real" radio and go from there. However, you do have many of the elements already.

    If you want to continue to experiment, here are some suggestions:
    The way you "couple" (or don't couple) the antenna to the tank circuit won't work at HF unless your antenna is really short because there will be too much capacitance in the circuit. You really need to couple the tank circuit to the antenna using a small value capacitor or (better yet) an antenna coil that is inductively coupled to the tank circuit.

    The way you are controlling the regeneration and tapping off the audio is not really very good.

    In the old days, the regeneration control was called the "throttle" and "throttling" is still best done with a variable capacitor. I would suggest removing R1 and R2 from their present locations, put a 1-2 MH choke coil in where R2 is now, discard R2, put a variable "throttle" cap between the top of the tickler coil and ground for smooth regeneration control and then relocate R1 so it's between the choke coil (opposite the tickler coil) and the battery. Tap the audio off at R1 using the blocking cap C3 as shown.

    Have fun.
    Last edited: Sep 14, 2011


    The "Scout Regen" actually is a variant of Kitchin's "High Performance" JFET design using a tickler and throttle cap. Its manual can be seen here: http://www.qrpkits.com/files/Scout_Regen_v1.2.1.pdf . The other design, that you described in your web page as having poor performance, appears to be called the "beginner's receiver" or, in the case of your PC board, the "Scout Beginner's Receiver".

    The high performance regen is also viewable here: http://www.electronics-tutorials.com/receivers/regen-radio-receiver.htm . That article also has much good advice about proper construction of a regenerative receiver. I followed that article's advice closely and have a good working regen as a result (works even up to 30 MHz).

    Now if you want to go for even higher performance, do visit the regenrx group mentioned already. There was a recent discussion there about using a separate detector stage (placed after the Q multiplier) for increased dynamic range. You can also use varactors for tuning and throttle control, which allows all RF leads to be kept very short (very important in a regenerative receiver).
  13. KX6MWS

    KX6MWS Ham Member QRZ Page

    Thanks!, I'll try that. I'm really glad I decided to start with AM broadcast before trying any HF; the signal is lower frequency and much stronger, so it had a better chance of working. I just found this http://www.circuit-projects.com/radio-frequency/regenerative-receiver-with-dual-polarity.html which is close to what you're talking about. I guess need to get another variable capacitor so I can control regen! (make it 2, since the variable capacitor I have now for the tank is 'junk')

    After reading a lot of the comments, I'm seeing a lot of your guy's suggestions in other circuits, such as coupling antenna to tank circuit through a capacitor, etc.

    I have a million-and-one questions, but I think I'll settle for just a couple:

    - I've noticed, is that I see many AM and shortwave designs with ticklers, but none for FM broadcast with ticklers. Is it just impractical? FM receivers have small coils, so I don't see a 5 turn air core coil next to a tickler really working as well as a 40 turn coil does with AM.

    - How does this design oscillate? http://www.vk2zay.net/article/file/360 I've seen some where the coil is tapped, and some feedback is introduced that way, but I don't quite get how it works here. http://www.ke3ij.com/am-univ.htm is also similar, but the tank is off the emitter instead of the base. (I may end up using KE3IJ's design, since it seems it can scale from AM broadcast up to aircraft, which means maybe I can use the same circuit and just swap out parts of it to make a simple multi-band radio.)

    Again, thanks everyone!
  14. WA6TLP

    WA6TLP Ham Member QRZ Page

    There is enough unneutralized interelectrode capacitance to allow the amplifier to go into oscillation.

    I looked over the links to the "high performance regens" and they are essentially identical to what I built with just some variation in part values and the addition of a biasing diode in my design. These designs are proven to work and if you can get the circuit board instead of having to 'gin one up like I did with a diamond burr, it will be so much easier and may even work better (although mine works just great). If you can get it as a kit - even if it's a little pricy, you will save a lot of time, disappointment and frustration AND money by getting the kit.

    Remember this though, if you do build one of these high performance regenerative radios, you will learn a lot, you will feel a very special kind of well-earned pride, you will gain a lot of satisfaction listening to broadcasts on your own radio and, if you are hard-core enough, you may even use the radio in conjunction with a homebrewed transmitter to make a few QRP contacts on CW now and again. One thing you will never have is a really good ham band radio that works even as well as a 1940 Echophone EC-1.

    Good luck and have fun.


    If you do go ahead and build a high performance regen, I highly recommend that you not solder the transistors directly into the circuit, but instead place them in plug-in sockets (IC sockets or similar) so you can swap them easily. Regens are rather critical to build and there are many possible sources of poor performance (overcoupling the antenna, too much feedback, parasitic oscillations, stray capacitance, wire lengths, ground loops, mechanical stability, thermal stability, voltage regulation, device bias, etc). Your transistors can also be a source of trouble if you happen to get a noisy or otherwise poorly-performing one. By being able to swap out your transistors easily, you can quickly determine if poor circuit performance is caused by your transistors or not. I spent too long on one regen trying to track down poor perfomance, that turned out to be due to two noisy transistors (RF amp and detector). Since my transistors were soldered into the circuit, it was troublesome to swap them so I instead wasted much time investigating all sorts of other possibilities before narrowing down the problem to the transistors.

    Also, don't underestimate the importance of layout, wire lengths, and physical stability (of tuning mechanism and coil). These become particularly important as you approach VHF.

    Your idea to build a wide-range multi-band radio based on KE3IJ's "universal regen" is something I've also pondered. Do let us know how it goes. You might find it easier to use the alternate design of KE3IJ's "40 meter tweeter" which does away with the emitter choke and basically replaces it with a resistor.

    Here's another possibly interesting design that covers 4.5 through 50 MHz.
    http://www.sparktron.com/pbox.html (see item number 28-110).
  16. WA6TLP

    WA6TLP Ham Member QRZ Page

    I've been building transistorized projects since the 2N107 days of the late 50's and I've never had a project fail to work because of a bad transistor nor have I ever wished I had put them in sockets.

    Of course, I've always bought premium quality devices where there was no doubt as to to their actual type.

    The advice on the importance of layout is good. If you lay things out well, the probability of having all the problems mentioned in QRP-GAIJN's post should be minimal when building something as simple as one of these regenerative radios. When I had all my problems with the "Beginner's Scout" radio, perhaps they were do to some of this, but when I changed design to the straight forward Armstrong type regenerative radio, using a tickler coil, but keeping the component layout essentially the same within the same box, everything then worked smoothly and well.

    I guess the moral is: stick with a robust design that is straight forward and does not require hand-picking parts in order for it to work properly. Avoid designs or abandon designs that are "squirreley" where regeneration control is not smooth and you get beeps and hetrodynes where they shouldn't be. If you go with something like Kitchen's "High Performance" radio (my radio is essentially that radio), you can't go wrong. Just be sure you avoid anything like the "Beginner's Scout" radio. :)
  17. KX6MWS

    KX6MWS Ham Member QRZ Page

    Thanks! Yeah, I'll stay away from the scout regen.

    Soon it'll be time to stop tinkering with this on a breadboard and start solding it together. I have seen Youtube videos of FM receivers working on a breadboard, so I may want to try it there first, but I know I won't have solid performance until I move it onto a proper circuit board.

    So, I have another question for you (WA6TLP). You suggested I put a choke between the tickler and the battery. Why? I see the same thing on Harley oscillators as well: http://www.electronics-tutorials.ws/oscillator/hartley.html (Scroll down to shunt-fed hartley)

    The version of the Hartley that makes more sense to me is 'Basic Hartley Circuit' which is higher up on the same page: The base is biased the same way I'm currently doing it (resistor network between battery and ground), but what confuses me is the resistor and capacitor on the emitter. Do I even need that? The resistor on the collector should limit the current safely. What performance improvement is made by doing that instead of just grounding the emitter?

    In many designs I see a choke put in between RF and audio stages; this makes sense to me. The choke on KE3IJ's radio is right before the output to the 'radio shack' speaker. (Which, BTW is the amplifier I'm using in all my experiments so far! That little speaker is pretty handy. My dad has one I remember from when I was a kid, and I was surprised to see RS still sells them, so I picked one up real quick!)

    This Saturday in the garage is going to be pretty fun!

  18. VK2HHS

    VK2HHS Ham Member QRZ Page


    I can certainly suggest Kitchin's high performance regen as a good build candidate, even I managed to get one going some years ago with VERY little electronics experience. I dig it out from time to time, it's a hot performer!

    For FM broadcast band, I can recommend the "Radioshack Special". Google it and don't be put of by the name and inferred quality.

    de VK2HHS
  19. WA6TLP

    WA6TLP Ham Member QRZ Page

    "You suggested I put a choke between the tickler and the battery. Why?"

    Hummm, how can I explain this simply and without a diagram?


    OK start by marking up your schematic as I have suggested or reference my schematic above. Now looking at the schematic you will see that the collector of your NPN (or the drain of my FET) is connected to the bottom of the tickler coil. With me?

    This is so that currents flowing in the collector (or drain) must flow through the tickler and thus affect its magnetic coupling to the tank circuit. OK?

    Now consider this, in order to couple the signal back down to the tank circuit so the circuit can amplify through regeneration or (with enough feedback) go into self-oscillation (for CW and SSB), RF current must be able to flow through the tickler coil (in addition to the normal DC collector current) so that it can be inductively coupled to the tank circuit. OK?

    The amount of RF current flowing through the tickler must be "throttle-able" which means that you must be able to carefully control it and we do this throttling by increasing or decreasing the bypassing that is performed by the throttle capacitor. Think of it this way, without bypassing to ground at the top of the tickler coil, no RF current flows through the coil to ground (DC flows, but no RF). The more capacitance, the more bypassing, the heavier the RF current will be flowing in the tickler, the more will be coupled down to the tank circuit and the more the circuit will "want" to oscillate. Does that make sense?

    Now, without a choke coil in the DC path to isolate the top of the tickler coil and prevent bypassing by the battery (which is an excellent bypass in itself) or by components in the audio circuit, lots of RF current will flow in the tickler and that get coupled to the tank circuit and you won't be able to stop the circuit from oscillating. In effect, the circuit will run wide open and your throttle capacitor just won't do anything, you will have just too much RF bypassing.

    Here's the magic of the choke. The choke blocks all the bypassing by the other circuit elements and the only thing that the top of the tickler coil sees is the throttle's bypassing to ground -- and you can control that. You absolutely need a choke coil right there to block unwanted bypassing to ground by the follow-on circuits. Does any of that make sense?
    Last edited: Sep 16, 2011
  20. WA6TLP

    WA6TLP Ham Member QRZ Page

    "what confuses me is the resistor and capacitor on the emitter. Do I even need that?"
    Probably not unless the audio is tapped off there, which it is in many designs. You will see that in my diagram, the equivalent source circuit is where I tap off my audio. Assuming your audio is not tapped there and you are using an NPN bipolar transistor in this circuit, bipolars are notorious for going into "thermal runaway" and thus a conservative design calls for additional biasing circuits in the emitter circuit. You are probably running extremely tiny currents through this transistor with minimal chance for thermal runaway, so my guess is that it probably isn't needed. That's my take on it anyway.

    I hate working with bipolars and love working with FETs. Part of that is because they are so much more like the wonderful vacuum tubes of my youth. Still, in many ways, FETs are easier to work with.
    Last edited: Sep 16, 2011
Thread Status:
Not open for further replies.

Share This Page

ad: dxeng