Dave's Homemade Radios Banner

Regenerative Tube Radio Building Notes

by David Schmarder

Font Decrease Font Increase



Regenerative Radio

Hello fellow radio builders or to be builders. After building a couple dozen regenerative tube radios, I have learned some lessons. Some of them have been difficult. If you like designing your own sets, this page is for you.

The information here specifically addresses tube radio building. Solid state regenerators use the same principles but sometimes in different ways. I will defer to others that have about building solid state radios.

This is a subject where entire books have been dedicated to this subject. I intend on doing it all in one web page! No big theory discussions or math here. You are here to build a radio, not to earn your college degree.


Antenna Coupling

Here is the beginning, where the antenna and ground meet the radio. I will use the example of an outdoor antenna, around 50 to 100 feet long (15-30m). Regen sets can be also made using loop antennas and ferrites. But lets talk about the old outdoor long wire antenna.

It is important to be able to adjust the signal level to the receiver. Regen sets overload easy, especially when listening to SSB and CW signals in the oscillating mode. Also, general overloads happen with MW reception. If too much signal comes in, it may be heard over half the dial. There are methods to limit the signals coming in, including mechanical and electrical adjustments.

The amount of coupling will depend on your local radio conditions as well as how good your antenna is. A short wire will need more coupling than the long wire. If you are not completely familiar with your antenna, you may have a little more radio tweaking to do.

On the dual coil radios, mechanical variable coupling is quite useful for adjusting the signal levels. This involves moving a coil (best it is the antenna coil) closer and further from the other coils.

Input Trimmer Capacitor

This is the tried and true single coil coupling solution. A small variable capacitor, around 50pF for a MW set, and just a few pF for shortwave connects the antenna to the top of the tank coil. It isn't my favorite way of antenna coupling but it is the way to go with a single coil. I have used it in radios with good success. It is most useful with plug in coil radios. You don't need to wind the extra coil. The frame of the capacitor should go to the antenna as that will lessen hand capacitance problems. The capacitor should be isolated from the knob.

Untuned Coil

I used this on a recent radio with great success. But I had to make the coupling adjustable. This meant a linkage to the front panel to move the coil closer and further from the main tuning coil. In this radio, I wound 25 turns on a 1 inch (24mm) hub spider form. At first, I had more turns, but it had overload that I couldn't control. The amount of turns needed depends on several factors, including antenna length, local signal strengths, and the amount of distance the coil can travel. I recommend at least 3 inches (75 mm) distance of coil travel. More is even better. Coils can also be rotated on an axis to control the coupling. This, by the way, is the basic variocoupler.

Fixed Antenna Coil With Taps

Here we have a situation where you place the antenna coil in a fixed position, perhaps an inch or two from the main coil. The coil could be a maximum of 25 turns as in the above example, with two taps, at 15 and 5 turns. The higher in the band you go, the fewer turns you will need. I used this in another receiver with good results.

Series Capacitor To The Antenna Coil

In this configuration you use a capacitor (typically a 365pF) in series with a fairly small coil (like the 25 turn for MW, less for SW). The capacitor will reduce the signal that reaches the coil. This is a good non-mechanical way to adjust the coupling.

Series Resonant Tuned

A series resonant circuit is a time tested way to build an input circuit. For the MW bands, the amount of capacitance and inductance is fairly large. In my opinion, this is a good circuit for shortwave regen sets.

Series & Parallel Resonant Tuned

I think this is the best antenna tuner. Some will recognize this circuit from some of my dx crystal sets. Both this and the Series Resonant Tuned circuit need a lot of coupling distance from the main coil. I usually build these on two bases so the spacing can be easily adjusted. If your regen is going to be used as a 1AD radio, then use this circuit.

Radio input coupling methods


RF Amplifiers

RF amplification can be handy, especially in shortwave receivers. With MW radios, rf amplifiers don't help very much as the band noise is already limiting your reception sensitivity. Shortwave radios are a different story. I use rf amplification to not only improve the sensitivity but to isolate the antenna from the detector. There are two reasons for this. One is that antenna movements can shift the frequency of the regenerative detector slightly. This is annoying when receiving CW or SSB signals. The other reason is that with the tube between the detector and antenna, there is no signal radiation from your oscillating detector to the antenna.

The signals are usually but not always coupled to the detector stage by a capacitor. On shortwave, this capacitor must be very small, only a few pF. On medium wave, they can be 10-50 pF.

Grounded Grid

Grounded grid amplifiers are usually use triodes. They are simple and very stable. The grid shields the tube's input from it's output. The gain is low. There is great isolation of the antenna from the detector and isolation from detector oscillations reaching the antenna. There is a fairly close match between the antenna and cathode. An rf gain pot is useful here.

Untuned Grid

The untuned rf amplifier has higher gain than the grounded grid amplifier. Stability is pretty good when a tetrode or pentode tube is used. This circuit still isolates the detector from the antenna.

Tuned Grid

This is the best of the best. A tuned input is a good way to keep those out of band signals that might cause unwanted rf products from appearing at the detector. A triode will work but a pentode style tube is recommended due to inter electrode capacitance problems. Also, this is where a metal chassis or base might pay off. Other problems such as tracking differences can hurt the performance. But if you are willing to go through the effort, you will be rewarded.

Tube RF Amplifier Input Options


Detection

Detection is what turns your RF into AF, but you probably already knew that. Here are a couple of ways that detection can be accomplished with a tube. Both take advantage of non-linear operation of the tube.

Keep in mind that detection and regeneration are two different topics. Regeneration only has to do with a feedback system that improves sensitivity, while detection involves signal rectification. You can have detection without regeneration and you can have regeneration without detection.

Grid Leak Detectiion

Grid leak detection is by far the most popular type of detection in regen radios. This system involves the use of grid rectification to retrieve the audio from an AM signal. There is a capacitor that charges to a peak level during the positive half of the modulated rf wave. This can happen because the grid conducts when driven positive with respect to the cathode. During the negative half of the wave, the capacitor then starts to discharge through the grid leak resistor that is across the capacitor. All this causes the plate current to change at an audio rate, but not symmetrically. The result is audio recovery at the plate. I am mentioning all this, to show the difference with another type of detection mentioned next.

Since the bias is at zero, if a triode is used, the plate voltage should be below 30 volts. Otherwise excessive current could flow. Pentodes can operate at higher plate voltages, as the screen voltage can be used to adjust the current flow.

Plate or Anode Bend Detection

A much less popular type of detection is called Plate Detection or as our friends in other countries would call it, Anode Bend Detection. While the grid leak detection method mentioned above operates at grid saturation, the plate detector operates at near grid cutoff. Both of these bias points are at the edge of tube non-linearity, and thus good detection.

With this detector, the tuned circuit is connected directly to the grid. There is a resistor connected between the cathode and B- or ground that is set to a value where very little plate current flows. That is, the cathode resistor provides the needed grid bias to operate this tube at near cutoff. I typically shoot for a 20 or so microamp plate current. By using a cathode resistor for the grid to cathode bias, a near constant plate current can be maintained as the tube ages.

When the modulated carrier wave comes in, half of the wave is in the conduction range of the tube and half is operating the tube at beyond cutoff. There is a small rf bypass capacitor across the cathode resistor that helps with the detection, by bypassing the rf to ground and allowing audio rate changes to pass. All this causes plate rectification, with the resultant audio pulses.

Until recently, I have never seen a regenerative plate detector, but it does work and works well. Normally plate detection worked best at large signals, with distortion at the low level signals. But maybe something is different due to the added regeneration as the quality is good with weaker signals too.

Another advantage of the plate detector is in SSB and CW reception in the oscillation mode. Since there is no grid rectification, there is no frequency pulling of the tuned circuit. This is the now popular "product detector" circuit that became popular in the late fifties.

Types of detectors used in regenerative radios


Feedback Control

Every regenerative receiver going back to Edwin Armstrong's discovery nearly 100 years ago depends on a portion of the output rf signal being coupled back into the input circuit.

All forms of feedback discussed here make use of inductive coupling for providing this feedback. I am not aware of any tube regen radios that don't involve coils in the feedback. Basically there are two feedback systems. One type is the feedback coil in the plate circuit and the other in the cathode. The cathode feedback can be a separate coil or a tapped main tuning coil. I usually prefer the separate coil because it allows me to exactly set the efficiency of the coupling to the main coil, either by distance or number of turns. This feedback must be regulated. Here are four popular methods.

Plate Voltage Control

This is the old time way of regulating the gain of a triode tube. A voltage divider potentiometer determines the voltage that is applied to the plate of the detector. The higher the voltage, the higher the feedback. A capacitor, typically .1uF to 1uF is placed from the arm of the pot to ground as a way to keep the electrical noise from the pot out of the detector. By reducing the tickler feedback as much as possible, the voltage can be raised and this will give more audio output.

Screen Grid Control

If you have a tetrode or pentode tube for a regenerative detector, then maybe screen grid control is the way for you to go. By adjusting the screen voltage, this will control the gain and the feedback. A capacitor from the arm to ground is still valid for noise reduction. This type of control should not be used for plate detection as you will be changing the static plate current at the same time as the gain. This will cause problems in how this detector works.

Shunt Tickler Resistance

This method is especially nice if you want to use a SPST switch to cut off all current flow in a battery tube radio. Since you can build a radio with no resistance going to ground, shutting off the filament cuts off all B+ current flow. Shunt tickler resistance control can be used with triodes to pentodes. With the latter, the screen voltage is tied to the B+ usually through a dropping resistor.

A variation of this that has worked out very well is to break the connection between the arm and end terminal of the pot Then move the top of the throttle capacitor to the arm of the pot. This makes it work better. The value of the pot will need experimentation, but generally has to be somewhere from 1k to 10k ohms. There also can be some problems with RF sneaking around the pot and causing instability. All in all, while this is not the perfect way to control regeneration, it is often used in battery sets.

Variable Throttle Capacitor

This is real old time radio operation! Pots were not too common in the early 20's but variable capacitors were in good supply. This works well for every kind of tube. A variable costs more than a pot usually, but it is noiseless and efficient. It was a long time before I used a variable throttle capacitor. It didn't look that wonderful to me. Was I wrong! I first tried it on my breadboard to see if there was merit to this. This has the possibility, with battery tubes to require only a SPST switch in the filament line to shut off the whole radio. A 365 pF works well with this circuit, MW or SW. There are cases that you may need an additional fixed capacitor. This is usually because the tickler is a little less than it should be. But using a fixed capacitor across the variable is entirely ok.

Variable Resistance, Fixed Throttle Capacitor

This is a circuit that I personally like. The regeneration control is very smooth. There are no extra parts used, compared to the variable throttle capacitor circuit described above. As with all the feedback circuits some experimentation will yield great results. For a MW radio, I would start with a 50k pot and a .001uF throttle capacitor. Adjusting the tickler coil may help too. There is less loading on the tickler coil because the resistance across it is the full value of the pot. As the pot is adjusted away from the plate towards the RFC, the throttle capacitor becomes more of an rf bypass, thus promoting rf feedback.

Regenerative control methods


Interstage Coupling

There are several good ways to couple your detector to the next stage, whether it will drive a speaker or headphones. Here are some of them.

RC Coupled

This is best used with a tetrode or pentode tube as you may want to operate with higher voltages. This will give you enough audio voltage to directly drive your output tube to speaker levels. The plate load resistor is somewhere around 100k - 330k ohm.

Choke Coupled

This is used only with triodes. The choke that is used measures somewhere in the 100 - 500 henry region. The wire is very thin with many turns on it. Sometimes a capacitor is placed across it to tune the choke for an audio peak for CW reception. Many times a high value resistor has to be placed across it to prevent "fringe howl".

Bogen Autoformer Coupled

Hey, it's our old friend, the Bogen T725. This time it is used as an interstage coupling device. You can experiment with which tap goes to the plate. This will give an audio voltage boost to the next stage. This is used with battery tubes mostly but will work with any kind. Triodes are used more often than pentodes.

Transformer Coupling

This was used quite often in battery radios with triode tubes. A transformer with a 1:3 windings ratio (1:9 impedance ratio) gives an audio voltage boost to the next stage. The transformer might be expensive.

Interstage Coupling Methods For Regenerative Receivers


Output

Headphones

This is where the radio meets the ears. There are different ways to handle the output. The simple way is to place a pair of high impedance headphones in the plate circuit of the detector stage. Make sure the voltages at the detector are low enough so that if an insulation problem with the headphones happened, that you would not be harmed. During the heat of dx, sweat can happen.

Speakers

Speakers: Maybe you would like to drive a speaker? A one or two stage amplifier is just what you need. If you are using battery tubes, sometimes a 3:1 ratio windings interstage transformer will give you the gain you need. It is best to look at some of the already published projects for ideas.


Tube Selection

Most any triode to pentode tube will work in a regenerative receiver. The tube gain will determine the tickler coil composition and how much feedback you will need. There is a lot of leeway here. Just start with a larger tickler and peel off turns.

Triodes

Triodes are the original detectors. A low mu tube works best, but medium mu tubes are ok too. B+ Voltages need to be kept low if used as a grid leak detector. These work well as grid leak or plate detection detectors. Not too much to go wrong here.

Tetrodes & Pentodes

Tetrodes and pentodes of about any kind make good grid leak detector regen tubes. You can use any plate voltage that the tube will handle, but the screen is generally lower than 20 volts. These tubes can be used in the plate detector, but the regeneration should not be controlled by varying the screen voltage. The voltages need to be kept steady.

In the case of a grid leak style detector, it doesn't matter if the tube is a sharp or remote cutoff type. The grid always operates at around zero voltage anyway, so plug in what you have. But if you use the tube as a plate detector, it should be a sharp cutoff style.

Tubes With Cathodes

Tubes with cathodes (indirectly heated) are versatile. The cathode lead can be grounded as a filament tube would be, or components can be connected to that lead. A cathode tube can always be substituted for a battery tube, so long as you take care of the cathode lead.

Directly Heated Tubes

Filament tubes are the original regen tubes, but they sometimes need special handling. They generally will not substitute for cathode type tubes without special things being done. If the cathode is not directly grounded, then the components have to be connected to one side of the filament. This means that rf chokes of low dc resistance have to be added. It really isn't worth it.

DC must be used to light the directly heated filament tubes. Also, the polarity of the filament must be observed. In detector circuits the positive grid pin (+) has to be grounded. Filament tubes can be run from ac to dc power supplies. A lot of filtering is needed.


Coils

There are up to three coils in any regen radio. The coils can be wound on spider forms, such as I usually do, or cylinder coils will also work very well.

Main Tuning Coil

This is the coil that selects the station you are listening to. I like to use litz such as 40/44 size. Some may want to use something bigger, such as 165/46 wire. This may help with the performance some, but you are probably as well off with the lower count wire. 100/45 might be a good compromise.

Antenna Coil

I like to wind this coil also with litz, but for a different reason. Many times this coil is moveable and having a multi strand wire keeps the wire very flexible and less prone to breaking. There are no set rules for winding because the type of circuit and how good your antenna determines what this will be. Radios with very short antennas need larger coils.

Tickler Coil

Nothing funny about this. It is essential for a regenerative radio to work. This also requires experimentation. But a good way to judge the set up is to first tune your radio to the lowest frequency you want to hear. Then increase the antenna coupling to maximum. You can then take turns off the tickler coil so that the radio goes into the oscillation mode at about 3/4 of full rotation of the regen pot, or throttle capacitor value. This is where you need the most tickler action. Once you have that set, check the high end of the band to make sure you have smooth operation there. Too much tickler coil can cause a weak output and instability.


Mechanical Considerations

Stability

The best construction of a regen is one that is solidly built. If the wires are flopping around, then your radio will warble. Keeping the leads short is a good idea in MW sets and essential in SW radios.

Shielding

Metal chassis are desirable but not always a requirement. If the radio is one or two stages, not much shielding is needed. If you have grid driven rf amplifiers, then some shielding should be used. Keep the coils away from each other.


Other Thoughts

Series Variable Resistors vs. Voltage Dividers

There are some designs that use a series variable resistor going to the triode plate or screen grid, rather then a voltage divider (with the bottom pot lug going to ground). This way of regen control has not been very satisfactory. Try to avoid this if you can.

Decoupling and Bypassing and Isolation

Good bypassing is the watch word in regen sets. I recently built a set and tried to leave out the rf plate choke. I spent days trying to get the radio to work right. I showed the circuit to someone else and he spotted it right away. The plate rf chokes keeps the rf where it should be, not wandering around the chassis. Keep the bypass elements near the tube and all will be ok.

Well, there it is folks, a complete guide to regenerative radio design all on one page. I didn't cover every aspect of designing a regen radio, just a enough to keep you slightly confused. You can print the page and read it while you are sitting in the little room or out at your workbench. Good luck with your design and build of your next regen radio!