To see this working, head to your live site.

Edited: Jul 06, 2022

Solid state superregenerative RX with squelch

Solid state superregenerative RX with squelch....

the superregenerative detector is designed around TR1, a BF191 VHF transistor. The base has a tight fixed bias by the diode string, forcing a stable operating point over a wide tuning range, aiding the squelch to maintain good performance. The audio is low pass filtered to remove the 30kC quench frequency and amplified in TR2. IC1a is the main gain block, its low output impedance preventing any unwanted interaction between the squelch and volume controls. The audio then is amplified in the LM386 gated audio power amplifier. The signal from IC1a is also passed to the multiple feedback bandpass filter around IC1b, the second half of the LM358. The filter has zero gain at the Fc of 12kC, dropping rapidly from each side. This filters out any residual quench, and any audio modulation. The output is rectified and the recovered vo

ltage used to gate the LM386 off, muting the radio. Set correctly, the squelch will "spit" off signal, and fully open on even the weakest transmission. This design is the classic "noise" squelch method, relying on the "quieting" effect a transmission has on the severe background hiss of any high gain RF amplifying system.

Purists will note the absence of any isolation between detector and aerial. My experiments have found little to no attenuation of SR hash with a buffer stage. Measured on my old spectrum analyser, the 6 microwatts of RF hash is spread over a roughly 200kC spectrum at the quench frequency , reducing any radiated power at any one point. My Yupiteru 7200 is struggling to detect any interference at a 3 meter distance, and that is one sensitive receiver.

It is also interesting to note, the performance of this version is very similar to the three valve version of the same thing, the former using a high active component count, the valved set using just 6 active stages.

Edit.... I forgot the pin numbering for the LM386... the transistor collector connects to pin 7, when switched on, it deprives the preamp stages within the IC of operating current. The capacitor bypass on pin 8 slows the voltage change to remove switching clicks, making the squelch "softer" with the attack and decay.

6 comments

6 Comments

qrp-gaijin

Jul 02, 2022

@dayleedwards wrote:

Here we have a 10 meg wide sweep centered on 120 megs.

Thanks for the interesting photos. To check my understanding, I annotated the photograph with the frequencies at the bottom.

@dayleedwards wrote:

You can clearly see the awesome sidebands caused by the self quenching. These are at Frx + Fq xN where N is the harmonic. The dropoff is the bandwidth of the RF tank. The actual RF tuned frequency is the first peak on the left.

OK, so in the photo above, this means the RF tuned frequency is 117 MHz. This means the reception frequency Frx = 117 MHz. The quench frequency Fq = 30 kHz according to the first post in this thread. So, we expect to see the "sidebands caused by the self quenching" at:

Frx + Fq x N
which equals 117 MHz + 0.030 MHz x N, giving for example:
117.030 MHz
117.060 MHz
117.090 MHz
117.120 MHz
etc...

But the "sidebands" (the peaks above 117 MHz, right?) appear to be spaced much farther apart, perhaps 500 kHz between each peak.

Does this mean the quench frequency is actually 500 kHz? If not, what determines the 500 kHz spacing between the peaks above 117 MHz?

dayleedwards

Jul 02, 2022

Replying to

This is a measurement anomaly I think, Either quench freq is nowhere near 500kC. IF I could synchronise the SA sweep to the quench frequency it would make more sense. Im thinking the sweep is behind the decay, similar to a sampling error, the screen updates effectively dividing the input by x amount.... but this does not explain the lack of similar sidebands in the valve version, as this is also unsynchronised, so one would expect a similar error. SRs are difficult to measure confidently, any disturbance by attaching measurement equipment upsets the operating conditions, thereby largely negating the measurement. .... at least with the equipment I have . My knowledge too is lacking in this regard. Using a SA signal bandwidth of even 30 kC requires a very slow scan, or sweep rate, nowhere near the quench frequency. A modern SA would be nice, but pricey.

qrp-gaijin

Jul 01, 2022

@dayleedwards wrote:

My experiments have found little to no attenuation of SR hash with a buffer stage. Measured on my old spectrum analyser, the 6 microwatts of RF hash is spread over a roughly 200kC spectrum at the quench frequency , reducing any radiated power at any one point. My Yupiteru 7200 is struggling to detect any interference at a 3 meter distance, and that is one sensitive receiver.

I would be interested in seeing any pictures or videos of these radiation experiments. I did similar tests a long time ago trying to detect my HF superregen's radiation on my FT-817 and (as I recall) the radiation was so low that I couldn't detect it.

dayleedwards

Jul 02, 2022

Replying to

OK, despite my poor photography skills......

Here we have a 10 meg wide sweep centered on 120 megs. You can clearly see the awesome sidebands caused by the self quenching. These are at Frx + Fq xN where N is the harmonic. The dropoff is the bandwidth of the RF tank. The actual RF tuned frequency is the first peak on the left A more informed display would be had if the sweep time could be syncronised to the quench frequency.

I will take the unprecedented step of reading the SA manual, it will be possible Im sure.

Here is a scarey wideband scan from 100kC to 1,8 gigs of the SR, the SA input simply connected into the aerial socket of the SR. Peak power of the fundamental is 6 microwatts, tapering down on the log scale to a few pW

Finally, the original measurement on my Anritsu analog SA, confirming the original HP trace.

dayleedwards

Jul 02, 2022

Replying to

AND.... lets confuse the issue even more. Superregens, are like women, and Chinese puzzles.

Below is the IDENTICAL measurement taken on the valved SR. Where are the sidebands??? I will leave that question open..... again, a 10 meg wide sweep.

Another quick observation is the peak power out of the SR aerial socket is now 30o nW, remembering this one has a grounded grid isolation stage,, the reverse attenuation being quite good. Here lays the difference between vacuum state and solid state. There is a PHYSICAL barrier between anode and cathode with the valved RF buffer in the form of a grounded mesh. The coupling is by a unidirectional elcctron stream only.

This is not the case with a transistor or FET input stage, the reverse isolation is almost non existent in simple circuits, hence my previous comments on their effectiveness, or lack thereof. They do however semi isolate circuit loading by creating a constant load on the detector, otherwise it can be an act to get a good transfer of energy., In the SSSR, I have connected it via a capacitive tap on the ground lead of the tuning capacitor, giving a very low impedance connection point negating the need for any isolation stage.

So, all very interesting, a picture is worth a thousand words.

dayleedwards

Jul 01, 2022

schematic