Single signal reflexed regenerative receiver
A proof of concept receiver

This is one of my best receiver designs so far. Best in terms of simplicity in conjunction with it's features. Using common components, a superior receiver can be built. I wanted a shortwave receiver that could receive many HF bands without changing any components apart from the resonator. This would save cost and pain, for having one circuit made for each band. However, at the same time, this receiver should be easy to build and replicate, with few and mostly easy to find components.

Note that Initially started the project as a regenerative receiver and then I reflexed it. Please scroll down at the bottom of the page, for the latest schematic on this receiver.

Before describing the circuit, I have to warn you. Beware of the simple KIT receivers or schematics on the web. All simple receivers you will encounter on the web or in KIT form, are not single signal receivers. This means that you receive both sidebands at once. In practice, this means that:

1. When you receive a USB signal and there is no other LSB signal, you will hear the noise of the LSB as well. That is double the noise.
2. When you receive an SSB USB signal and there is another station at the LSB, you will hear both stations simultaneously.
3. Not only you will hear both stations, but in fact you may be unable to receive any of them. That is because, given the fact that HAMs transmit only in one of the two modes (USB/LSB) in each band, the audio from the opposite sideband appears inverted. So you will not only hear both stations, but one of them will be inverted, which will make the wanted station difficult to distinguish.
4. When you receive a USB signal and as you tune your dial, after a while you will receive the same signal again in the LSB. You would think you have heard two stations, but it is in fact just one.
5. In crowded areas of the band, it may be very difficult to distinguish single signals.
6. When operating CW, signal inversion of the opposite sideband (point 3 above) is not important and your brain may be able to distinguish the audio tones by comparing the pitches as you tune the dial and concentrating on one of them. However, all the rest disadvantages still apply.

It is ok to build such circuits for fun and experimentation, but if you are more serious hobbyist, you would need a single signal receiver. There are two ways and a half, a single signal receiver can be built:

1. The filter method. It is complex and it uses crystal filters difficult to build. Usually even more complex to be applied to multi-band receivers.
2. The phasing method. It is complex and it relies on very accurate phase difference in order to achieve good carrier suppression and sideband cancellation. Usually even more complex to be applied to multi-band receivers.
3. The binaural method. This is the "half" method I mentioned earlier. This is not actually a single signal receiver (hence the half) but a DSB one, that let's your brain distinguish between signals that appear in different "places" in the 3D listening space. Not very complex as the previous methods, but it still requires a phased local oscillator, which may be difficult to build. Not a real single signal method anyway for the complexity involved.

I thought these were the only methods for single signal reception, until I found another method that uses very simple circuits:

4. The regenerative method. This is only for USB reception, not LSB and not AM. In other words the oscillator is set below the signal you want to receive.

This is the principle behind which my receiver works. By careful design of a regenerative stage, the regeneration can be made to be so fine, that you can adjust it to be effective on one sideband and not effective on the other. To understand this better, suppose you receive a DSB signal (both sidebands tone modulated). By fine adjustment of the regeneration, you will receive a tone on the USB and just a hiss on the LSB. The hiss will be like listening to an SSB modulated signal on an AM receiver. Not only that, but the received hiss on the LSB will be much attenuated, as a result of the lack of regeneration effectiveness on that sideband.

100nF

+12V

100K

20K
MT

2.2K

+12V

2.2M

+12V

2.2K

32R-
600R

+12V

4.7K

100nF

+12V

4.7pF
NP0

Ceram.
Filter
Reson.

3.3K

2N2222

470uF

ANT

470K

BC549C

2N2222

BC549C

100pF
NP0

10K

100nF

47pF

5K

100nF

2N2907

+12V

100K

100pF
NP0

10K

22nF

10K

BC549C

2.2M

100nF

20K
MT

100nF

MV209

SSB
CW

The receiver in it's simplest form

My receiver was inspired from this and this video, but I have made a lot improvements on the original circuit. All the improvements, were done with simplicity in mind, after all, this is a receiver without any coils to wind!

First of all, I changed the transistors of the preamplifier and the regenerative stages to BC549C. The BC549 is still cheap but has lower noise, so I don't see the reason why not using it. The C version has also increased gain, which improves the sensitivity of the receiver much. I have tried to replace the preamplifier with a high dynamic range one, but it's gain was not enough so I decided to use the original circuit, as it was thought to be more important to have more sensitivity than trying to achieve high dynamic range. There is obviously some ground for improvement here.

The second change, was to use varactor tuning instead of the classic air variable capacitor. Not only the cost of such a capacitor today, can be more than the whole receiver, but also the size and the weight of the receiver was reduced with the varactor, there are no hand effects and most importantly no expensive mechanical gear reduction drives are needed. A single multi-turn trimmer is cheap, it has low parasitic inductance compared to the multi-turn potentiometers and it does the job perfectly well. I used multi-turn trimmers for both the frequency and the regeneration set. The latter one helped the pleasant tuning of the regeneration, so that single signal reception is easily possible. I used a 7.160MHz ceramic resonator and an MV209 varactor, controlled by a multi-turn trimmer, which tuned the receiver from 7.003MHz to 7.080MHz with great frequency stability (built on breadboard). The receiver works on other bands too by just changing the ceramic resonator.

The next improvement, was to add a little bit more audio filtering without getting too complex. I could use better low pass or CW band pass filters, but they would require LC circuits or active RC circuits which are more complex, almost as complex as the whole receiver. Not to mention, I wanted to keep the whole circuit made with discrete components. The 22nF capacitor is a simple SSB audio BPF, but the benefit of adding 100nF to it to limit it's high response for CW operation was obvious. Not only that, but this extra capacitor attenuates quite a lot the audio hiss and noise in strong QRM. There is obviously some ground for improvement here.

Next, I tweaked the values of the regeneration resistors/potentiometers, so that the regeneration could be set with greater precision (finer). The 100K trimmer, is preset, so that at the minimum frequency, you should be able to set the regeneration (using the regeneration potentiometer) so that the receiver can go out of puff (just receiving AM). At the same time, at maximum frequency the regeneration should be able to be set, so that the receiver can go into puff (just receiving SSB). When this setting is done correctly, you don't need to reset this trimmer anymore and you should be able to "use" the whole range of the regeneration potentiometer from end-to-end (maximizing regeneration precision), without any strange oscillation occurring.

I used two 20K 30-turn potentiometers I happened to have available, for the regeneration and the frequency control. To reduce the value of the regeneration potentiometer, I used a 2.2K in parallel to it. You can use a smaller potentiometer instead, if this is available to you. The potentiometers I used, were taken out of an old B&W TV set (you do save old electronics, don't you?), as shown in picture below. They were used there, to preset the analogue "memories" of the channels. These potentiometers are great value for QRP projects, as they combine 4 components into one, in a tiny space, a low parasitic inductance linear carbon wafer, a 30-turn reduction gear, a side knob and a display needle that moves along with the potentiometer. The feature of adding an analogue frequency and regeneration display to the receiver, without using any extra components, is great!

The final change I did, was to add a suitable audio amplifier to the receiver. The obvious solution was to use an LM386 and in fact I have tried different LM386 configurations, including this one. The gain was good but I couldn't get rid of the hiss from this IC no matter what I have tried. This is indeed a noisy IC and surprisingly this receiver has quite a low noise, to be ruined from a noisy audio amplifier. Thus, I used a discrete audio amplifier, which has much less noise than the LM386, it provides more than 30dB of gain and plenty of output power when using standard 32R or higher impedance earphones. Moreover, the components count is less than the LM386 circuit and they are more common to be found. To adjust the amplifier, set the volume potentiometer to minimum, so that there is no signal at the input and adjust the 5k trimmer so that the current that flows between the VCC and the 2n2222 transistor closest to the electrolytic capacitor, is 5mA.

Building and operating this first version of the receiver, has been proven to be much easier than thought and this is mainly due to the use of the ceramic resonator and the smooth regeneration control. First of all, there are no coils to wind and no mechanical variable capacitors, so there are absolutely no hand effects. The regeneration control, affects little the reception frequency, so there is not much frequency pulling, when you try to adjust the regeneration. When power is disconnected and connected back in, the receiver stays at the frequency where it was set previously. There is no audio oscillation at all (caused in other receivers by setting the regeneration too high), no matter how hard you try.

Another feature, is that you can tune across the whole tuning range without the need to reset the regeneration control. It handles like a direct conversion receiver then and not a two-handed tuned regenerative. However, when you do not reset the regeneration, you cannot receive single signal. So this one-handed tuning process assumes that you would receive DSB, not USB only. However most of the signals on air today are SSB, so the only bad thing in receiving DSB is the noise picked up from the opposite sideband, assuming no station is present on that sideband. It is only in crowded areas of the band, where you would need to reset the regeneration control, so that you receive the USB only and reject the LSB stations.

Another feature, is that my version, shows no audio oscillation at all, no matter how high the regeneration is set. This really saves your ears from abrupt loud sounds, when listenning with earphones.

Addition of a better audio BPF

Single signal reception is great, but in this receiver it leads to reduced sensitivity. As the regeneration is decreased, so that the LSB starts to be out of puff (starting from the high pitch frequencies and progressing to the low), the sensitivity or the regenerative detector in receiving the USB, is also decreased. Then you only depend on the gain of the RF preamplifier and the AF amplifier to compensate for this reduced sensitivity.

However, there is something else that can be done to improve the sensitivity of single signal reception, without receiving more of the unwanted sideband. Especially in CW or other narrowband modes, by proper AF bandpass filtering, one can increase the sensitivity of the USB, and still keep the LSB out of puff. To understand this better, let's assume a 700Hz CW BPF after the regenerative detector. This filter allows only the frequencies +/- 700Hz away from the oscillator, to pass through (USB/LSB). As the regeneration is decreased to receive only the USB, the LSB is progressively attenuated and set out of puff, starting from the high pitch frequencies and progressing to the low. If the CW BPF was not present, one would need to reduce the regeneration more, so that this out of puff point would be at the lower audio portion of the LSB, or ideally at the carrier frequency. But since there is a 700Hz BPF filter in place, we can set this out of puff point to be at higher pitch (higher LSB frequencies), increasing the regeneration, since the BPF will attenuate the low audio tones anyway. In other words, we increase the regeneration and increase the USB sensitivity, whereas at the same time the lower portion of the LSB (close to the carrier) despite being still received, is attenuated by the AF BPF. At the same time, the higher portion of the LSB is out of puff because we do not increase the regeneration that much. Simply put, we move the out of puff point more towards the high LSB frequencies and rely on the AF BPF to attenuate the lower frequencies of the now received low end portion of the LSB. Of course this BPF will attenuate the lower frequencies of the USB as well, but this is desirable especially in CW or other narrowband modes. From the above analysis, it seems that despite it's complexity, an AF BPF would be very beneficial to this receiver, especially in CW and other narrowband modes. In operation, you may find out that it is better to use a higher frequency BPF than 700Hz. Something like 1.5KHz to 3KHz could be better, as this will give you more sensitivity in single signal reception, but also more headroom for the setting that rejects the LSB. However the higher pitch tone may get your ears tired quickly.

100nF

+12V

100K

To PC
MIC

+12V

20K
Multi
Turn

100nF

2.2K

470uF

32R-
600R

2.2K

+12V

2.2M

+12V

10K

SSB

+12V

4.7K

3.3K

2N2222

470uF

4.7K

100nF

+12V

2N2222

4.7pF
NP0

Ceram.
Filter
Reson.

47nF

6.8K

47nF

10K

6.8K

47nF

10K

ANT

470K

BC549C

5K

100nF

2N2907

BC549C

100pF
NP0

10K

100nF

6.8K

47K

10K

BC549C

2.2M

47pF

BC549C

+12V

100K

100pF
NP0

10K

22nF

560R

47nF

560R

47nF

560R

47nF

100nF

20K
Multi
Turn

MV209

The receiver with the 700Hz audio BPF filter added

The filter that was added to the receiver, is called multiple feedback band pass filter. I have actually used three such filters cascaded. I have previously tested many different narrow filter topologies based on regeneration, but none proved to be stable enough. This filter, gives very good selectivity without being unstable or needing for any adjustments, so despite it's complexity, it is worth it. There is a switch, that selects between all three filter stages in series for narrow modes or just one stage, which when combined with the 22nF input capacitor, is useful in SSB operation. The SSB passband is not ultra flat of course, but in practice, the filter works just fine on SSB and there is no need for extra complexity.

Altering the value of the AF bandpass filter

I have decided to raise the frequency of the narrow band pass AF filter a bit to 978Hz. I just changed the 47nF capacitors in the filter to 33nF. This has been done, because I found that it was a little bit tricky to reject the unwanted sideband in CW, as the regeneration had to be set very low. This in turn, decreased the receiver sensitivity a lot in the wanted sideband. By separating the LSB and USB a bit more, it was much easier to reject the opposite sideband, without reducing the wanted sideband sensitivity a lot.

100nF

+12V

100K

To PC
MIC

+12V

20K
Multi
Turn

100nF

2.2K

470uF

32R-
600R

2.2K

+12V

2.2M

+12V

10K

SSB

+12V

4.7K

3.3K

2N2222

470uF

4.7K

100nF

+12V

2N2222

4.7pF
NP0

Ceram.
Filter
Reson.

33nF

6.8K

33nF

10K

6.8K

33nF

10K

ANT

470K

BC549C

5K

100nF

2N2907

BC549C

100pF
NP0

10K

100nF

6.8K

47K

10K

BC549C

2.2M

47pF

BC549C

+12V

100K

100pF
NP0

10K

22nF

560R

33nF

560R

33nF

560R

33nF

100nF

20K
Multi
Turn

MV209

The receiver with the 978Hz audio BPF filter added

Reflexing the receiver

A significant improvement in the sensitivity of this regenerative receiver, can be done if it is made reflexed. That is, if one or more transistors could be used to amplify not only RF but also AF. A few people have managed to make reflexed regenerative sets with one transistor, with good results. However, there is a major disadvantage in using a single transistor. That is, the leakage of the oscillating detector signal to the antenna. There is a technique that can be used to avoid this leakage, but the Wheatstone bridge it uses, needs readjusting for optimum results. Adding a second transistor costs almost nothing and solves the leakage and antenna isolation problems, once and for all.

The RF front end of my receiver, uses two transistors, one for the RF preamplifier and another one for the regenerative detector. With an ingenious trick, the RF preamplifier can be used for AF amplification as well. I have seen this technique applied in a very old circuit (pp 59 in this pdf) that uses vacuum tubes, that also used two stages in its RF front end. Applying to my circuit, is only a matter of using just 2 more components. Apart from that, anything stays the same, but I have redrawn the circuit below, to avoid long interconnect lines between components. The 3.3mH inductor is a molder choke (the one that looks like a resistor) and it's value is not critical. Anything in the low mH range (1-10mH or so) can be used. It's purpose is to block RF and let only audio pass through it.

To PC
MIC

+12V

2.2K

+12V

4.7K

100nF

2.2K

470uF

32R-
600R

100K

+12V

100nF

100K

20K
Multi
Turn

100nF

10K

SSB

+12V

2.2M

MV209

3.3K

2N2222

470uF

100nF

4.7K

+12V

2N2222

Ceram.
Filter
Reson

4.7K

100nF

6.8K

33nF

6.8K

33nF

10K

6.8K

33nF

10K

BC549C

5K

100nF

2N2907

100pF
NP0

4.7pF
NP0

470K

ANT

47K

10K

BC549C

2.2M

BC549C

10K

100pF
NP0

10K

47pF

3.3mH

560R

33nF

560R

33nF

560R

33nF

22nF

100nF

The regenerative receiver reflexed

Minimizing the power consumption of the receiver

The receiver, is already battery-friendly, since it is a reflexed regenerative design with discrete transistors. Most of the current is drawn by the audio power amplifier (about 5mA when idle and a max of about 8mA in operation). If you choose to connect it to a PC, you will normally disconnect the headphones from the audio power amplifier and connect them to the PC sound blaster output. Since the headphones, are connected in series with the power supply to the audio amplifier (when they are connected to the receiver), by removing them, power is automatically disconnected from the audio amplifier, which serves in reducing the overall receiver power drawn. However, if you choose to use a PC for digital signals decoding and also to use the receiver audio amplifier as a monitor, you do not save any power in that case.

With a simple additional trick, you can reduce the power consumption even more on demand, allowing for an even more flexible power distribution scheme. A switch is already used for selecting between SSB and CW. In SSB, two out of the three transistors of the audio filter, are not used. By changing the switch to a DPDT (marked as SW1a and SW1b in the schematic), you can power these two transistors off, when you do not use them, reducing the overall receiver power drawn in that case. Since the switch is DPDT, there is no additional knob in the front panel, just a more clever circuit wiring.

100nF

2.2K

4.7K

100nF

SW1a

To PC
MIC

470uF

2.2K

32R-
600R

270R

100K

4.7K

10K

100nF

SW1b

100nF

SSB/CW

3v
white
LED

20K
Multi
Turn

3.3K

2N2222

470uF

4.7K

100K

20K
Multi
Turn

2N2222

3v
white
LED

2.2M

Ceram.
Filter

MV209

100nF

6.8K

33nF

6.8K

33nF

10K

6.8K

33nF

10K

SW2

10K

BC549C

100nF

+12v

100nF

2N2907

3v
white
LED

100pF
NP0

4.7pF
NP0

470K

ANT

47K

BC549C

2.2M

BC549C

3v
white
LED

10K

100pF
NP0

10K

47pF

3.3mH

560R

33nF

560R

33nF

560R

33nF

22nF

SW3
momen-
tary

100nF

Minimizing the power consumption of the receiver

The circuit diagram above, has been redrawn to include the power modifications and also to show the complete VCC path in the different stages. Notice the volume potentiometer that is used, which is one that includes a switch as well. This way, the number of the front panel knobs is minimized. Another reason why such potentiometer is used, is to save your ears, when you power on the receiver. This potentiometer ensures that the audio volume will be minimum each time you switch on the circuit.

Also, notice the four white LEDs that have been added at the very right end of the circuit. These do not serve for any functional purpose and they are used as display illuminators only. I intended to use multi-turn potentiometers, like the ones shown above in this page and combine them with a printed paper scale with markings on it, for the different frequencies, band segments and regeneration settings. This way, I would create a useful analogue display for the receiver, at almost no extra cost. Well I thought, it would be nice if this display could be also illuminated, so I can see the different settings at total darkness and that is the purpose of these LEDs. The types of LEDs that were used, were the ones that have straw hat lens, which allows for wide light angle. Together with their proper spacing below the printed paper scale, they provide a good and even illumination. I have chosen to illuminate the display from below, because it seemed better than just placing a LED above it, in terms of illumination spread, but also because of more stable LED mounting and overall display size reduction, since the space below the scale was not used for anything else anyway. Note that SW3 is intentionally chosen to be a momentary switch, because I do not want the LEDs to draw precious power from the battery all the time, but only during the switch press. Four 3v LEDs are connected in series, because the voltage of the receiver is 12v. If you choose to use 9v for the receiver voltage, place just three of these LEDs in series. The 270 ohm series resistor, further limits the current drawn by the LEDs and it seems to provide a good balance between LED brightness (even at ambient light present) and current drawn. The total current drawn by the LEDs with this resistor in place, is about 5mA at 12v, when SW3 is closed. Initially, I have thought to put the paper scale above the potentiometers needles, so that I could project the shadow of the needles onto the paper scale. This was also a reason why the illumination was initially thought to be done from below the paper scale. However, this would require the LEDs to be switched ON all the time, which would draw power continuously. So I abandoned this idea and placed the potentiometer needles above the paper scale. This way, when there is ambient light, you do not need to switch ON the illumination.

Adapting the receiver for LSB (UNDER CONSTRUCTION)

The receiver works nicely up to this point. However, as already mentioned, there is a slight problem with it. On the lower HF bands, HAMs tend to use LSB, not USB and this receiver is only capable of receiving USB in single signal reception mode. Of course you can operate the receiver in DSB mode to receive LSB, but then the single signal reception advantage would be lost. In narrow band single tone modes, there is no difference if using USB to receive an LSB signal, but in voice SSB and multiple tone modes, the problem arises. If you use a USB receiver such as this one, to receive an LSB voice signal, by setting the local oscillator of the receiver below the LSB signal, the recovered audio would be inverted in frequency. The same applies to digital modes such as RTTY. The highest pitch tones would appear lower and vice versa.

However, there is a trick we can do to allow for LSB reception in single signal mode. Instead of trying to select the LSB in the RF part of the receiver, which would be difficult if not impossible in such a simple circuit, we can set the receiver to USB mode and set it's local oscillator to be below the LSB signal. The recovered audio, would then be inverted in frequency of course. If we re-invert the recovered audio spectrum, then the result would be the same such as if we were receiving LSB. I have not seen this audio spectrum inversion technique in any HAM receivers, so I thought to give it a try.

A possible advantage of this technique, is that the local oscillator, does not need to change necessarily in frequency when switching from voice USB to voice LSB. However, in narrowband modes and when the 978Hz BPF is enabled, it may need to be changed, since the 978Hz is not in the middle of the audio passband.

Receiver performance (UNDER CONSTRUCTION)

The performance of this receiver is very good for such a simple circuit. I have tested it, using a simple 12m wire slopper antenna, starting at about 4m and ending at 10m. The antenna has not got any ground. There is a coaxial connected to it and it's shield is left unconnected at the antenna side. The shield of the coaxial is only connected to the receiver ground side, which is not connected to any real or good ground, but just the negative end of the battery I am using to power up the circuit. With this antenna, I have only managed to talk throughout Europe with 100W on an Icom IC-728. Even with this poor antenna, the receiver performs good, although I have not made any side by side comparisons with other receivers. The sensitivity is more than enough and the audio sound quality is clear and crisp. In fact I enjoy listening to CW stations with this receiver much more than the IC-728, as there is no excessive noise that leads sooner or later to ear fatigue.

As said earlier, the tuning of the receiver is particularly easy for a regenerative receiver. Since HAMs transmit mostly SSB nowadays, most of the time, I set the regeneration so that the receiver just oscillates when tuned to the highest frequency and then decrease the frequency to scan the band down to the lowest frequency. This ensures maximum sensitivity at all times, which is usable when you want to dig out for weak signals, but also DSB only reception. If there is a point where lots of stations exist, then I reduce the regeneration control, so that I can reject the LSB signals. It is this flexibility that makes this a nice receiver, operate it as a direct conversion DSB receiver or as a USB regenerative receiver if you need to. As said, there is no weird oscillation to worry about, no matter how hard you try.

If there is a ham transmitting DSB CW with a cheap homebrew gear, you can always tune the receiver oscillator below his signal and reject the LSB if you wish, but not the USB. If he transmits USB voice, you can tune the receiver oscillator below his signal and receive it, either you decide to reject LSB or not. However if he transmits LSB voice, you will hear his voice inverted when you try to tune the receiver oscillator below his signal. The only thing you can do in this case, is to tune the receiver oscillator above his signal and set the regeneration to receive DSB. In that case, if a station appears on USB you will hear it as well.

For some people this procedure may be difficult to follow, but in practice it is very easy to cope with. More technically skilled HAMs usually find the flexibility that a regenerative receiver offers, very desirable. By appropriate adjusting of the regeneration control, the operator can do many "tricks" and control the behaviour of the receiver more precisely, unlike any other receiver. The simplicity of such receivers in contrast to their complexity, is mind blowing.

Back to main site