Audio spectrum inversion using transistors?

In my page http://qrp.gr/regenrxcer/index.htm I have described a regenerative receiver that can under some circumstances receive USB (not LSB).

I need an audio circuit that can do audio spectrum inversion for 200Hz or so to 3KHz or so. This means, the higher audio frequencies must appear lower and the lower, higher.
Is is similar to the analogue audio scramblers I think. This will allow me to distinguish LSB signals as well, by inverting the audio received.

Can this be done some way using simple discrete circuits?

Any info or schematics are appreciated.

Will something like this (replacing the oscillator with a discrete one) be useful?
http://www.discovercircuits.com/Andy/Voice-Scrambler_Descrambler.htm

Yes it can be done but will cost you a lot of time and tests.
Basic idea of analog scrambler is to mix the audio spectrum with a higher carrier, then mix again but with another carrier and obtain a reversed spectrum but in original BW so will be UN-intelligible.

Take a llok on CML website, have circuits that do that already in a small package
**broken link removed**

Otherwise, you should mix your self which isn't a big deal, but filtering will be a huge effort
https://circuitsstream.blogspot.de/2013/06/build-voice-scrambler-or-descrambler.html


For example

300....3000 mixed (use MC1496 or NE612) with fixed 10 KHz will produce two side bands

10300... 13000 (the normal audio, shifterd) <-- filter this out
7000... 9700 (the reversed spectrum) <-- keep this one

Mix again with 6700 Hz and will obtain

300... 3000 <-- scrambled, but in the same audio band

Now the problem is filtering on side band. If you are clever, will mix with something to use a professional crystal filter used by hamradio amateurs in SSB communication (search for SSB crystal filter). For this particular filter (in KHz or MHz range) you will need two carriers that can be obtained by using either proper crystals (hard to find) or some DDS chips.

And finally, all this crap can be easily obtained inside a DSP circuit like dsPIC or something.

Regards,

- - - Updated - - -

PS.

Best result for mixing in KHz range will be achieved with switched CMOS mixers such as CD4066. Search on internet for MPX stereo encoders like this one to see the mixer stage. This will already cancel out some unwanted products of mixing
https://www.qsl.net/zl1wtt/stereo.htm

Also see the OA and other stuff nearby.

Now for SSB processing, look for some old russian electromechanically filters like EMF500 or something similar. This is in KHz range area so operational amplifiers will work nice, no complex schematic. You will need a 2400...3000 Hz bandwidth filter.

For using a crystal filter, usually you will find them in 9 MHz...10.7 MHz range or, you can build yourself one by buying a 100 pieces of inexpensive 1 MHz crystals and select them for a ladder filter.

zuzurelu said:

Yes it can be done but will cost you a lot of time and tests.
Basic idea of analog scrambler is to mix the audio spectrum with a higher carrier, then mix again but with another carrier and obtain a reversed spectrum but in original BW so will be UN-intelligible.

Take a llok on CML website, have circuits that do that already in a small package
**broken link removed**

Otherwise, you should mix your self which isn't a big deal, but filtering will be a huge effort
http://circuitsstream.blogspot.de/2013/06/build-voice-scrambler-or-descrambler.html


For example

300....3000 mixed (use MC1496 or NE612) with fixed 10 KHz will produce two side bands

10300... 13000 (the normal audio, shifterd) <-- filter this out
7000... 9700 (the reversed spectrum) <-- keep this one

Mix again with 6700 Hz and will obtain

300... 3000 <-- scrambled, but in the same audio band

Now the problem is filtering on side band. If you are clever, will mix with something to use a professional crystal filter used by hamradio amateurs in SSB communication (search for SSB crystal filter). For this particular filter (in KHz or MHz range) you will need two carriers that can be obtained by using either proper crystals (hard to find) or some DDS chips.

And finally, all this crap can be easily obtained inside a DSP circuit like dsPIC or something.

Regards,

Click to expand...

Thanks for the idea of how it works and the schematics.
Will something like this work? (replacing the oscillator with a discrete one)
http://www.discovercircuits.com/Andy/Voice-Scrambler_Descrambler.htm

Also I was wondering why voice shifting is needed? Can't you mix 300....3000 with a 3500 carrier so as to get 300....3000 inverted and 3800....6500 shifted? Then you can filter the shifted one with a simple LPF.

Yes, you can. This is the idea inside of DSP where is much simpler to implement a LPF with 80dB slope or something.

In real life, you will not be able to cancel out the 3500 Hz carrier which is in the same audio band, so basically will hear a continuous tone along scrambled and de-scrambled signal. A crystal filter will have easily this attenuation so that double mixing with higher carrier will obtain far better performances.

- - - Updated - - -

But I saw that you want this for SSB receiving.
Regenerative circuits are quite fan but poor in selectivity.

Why don't you build yourself a direct conversion one (easily too) and use quadrature demodulator such this one
http://lea.hamradio.si/~s53mv/zifssb/block.html

This is called zero-if receiver and has advantage that only changing RF block and OSC will get all bands you are interested with proper USB/LSB and CW reconstruction.

Why can't the orginal receiver resolve LSB?

Brian.

betwixt said:

Why can't the orginal receiver resolve LSB?

Brian.

Click to expand...

Well it can receive LSB (in DSB reception "mode"), but it cannot set the USB "out of puff" and attenuate it, at the same time.
It can however do the inverse, Receive USB and attenuate and set the LSB "out of puff" with careful setting of the regeneration (tested). Thus you do not hear any tones from the LSB just hissing, which sounds like listenning to an SSB transmission on an AM RX.
If I can easily do audio spectrum inversion, I could set the RX LO to be below the input signal (like receiving USB) and perform audio spectral inversion to recover the original LSB conversation.

- - - Updated - - -

zuzurelu said:

Yes, you can. This is the idea inside of DSP where is much simpler to implement a LPF with 80dB slope or something.

In real life, you will not be able to cancel out the 3500 Hz carrier which is in the same audio band, so basically will hear a continuous tone along scrambled and de-scrambled signal.

Click to expand...

Ok I got it, the problem is because you can't filter effectively the audio carrier and the upper audio sideband.

I wonder if this circuit is capable of such a single inversion, which you mentioned is bad, but I just wonder.

As you probably figure out already, you can receive USB only due to nature of natural shifting in TX's mixer like we discussed above.
SSB is just ordinarily AM modulation, but with carrier suppressed (no more wasted energy for carrier itself) and one side band suppressed too because information is just redundant.

Basically a re-generative RX will be capable of "demodulating" all kind of standard modulations (not quadrature for eg. ) ) by positioning itself (the oscillator) nearby carrier for FM, on carrier for AM and so on, basically the principle is that modulated information will interfere with your oscillator that will act in the same time as mixer too (more or less).

Since LSB is reversed you can try demodulate by positioning yourself under the carrier, so will appear like USB for you. But in the same time will receive also LSB from bellow frequencies so will end up in listening a nice mix of 2 stations.

zuzurelu said:

Since LSB is reversed you can try demodulate by positioning yourself under the carrier, so will appear like USB for you. But in the same time will receive also LSB from bellow frequencies so will end up in listening a nice mix of 2 stations.

Click to expand...

If I position the LO below an LSB signal (assuming my RX is set for USB), I will receive the audio inverted. If I re-invert the audio I can understand the message.
This is needed because this special RX, can ONLY receive USB (kind of single signal) or DSB, not LSB alone.

If I tune the LO below an LSB signal, and set the RX in USB, I need to invert the detected audio to understand the message.

*I still haven't got a clue if the circuit in post #6 will work at all.

I made an audio scrambler circuit with an MC1496 balanced modulator/demodulator IC and a couple of switched capacitor lowpass filter ICs, about 31 years ago. The demodulator was almost the same and its output sounded perfect. The transmitted signal had a suppressed scrambler carrier and it sounded like high audio frequencies were low and low audio frequencies were high, completely scrambled.

Audioguru said:

I made an audio scrambler circuit with an MC1496 balanced modulator/demodulator IC and a couple of switched capacitor lowpass filter ICs, about 31 years ago. The demodulator was almost the same and its output sounded perfect. The transmitted signal had a suppressed scrambler carrier and it sounded like high audio frequencies were low and low audio frequencies were high, completely scrambled.

Click to expand...

I do not see why the carrier should be present in high levels using a balanced modulator like mentioned in post #4.
Will the circuit in post #6 be able to do it? I cannot see any balanced thing on it..

The circuit in post #6 does not invert audio frequencies. Instead it simply adds buzzing modulation then a voice sounds like an old robot.
If the scramble frequency is a little higher than the audio frequency band then the audio might have an upper and a lower sideband, one which could be filtered to make inverted frequencies.

Audioguru said:

The circuit in post #6 does not invert audio frequencies. Instead it simply adds buzzing modulation then a voice sounds like an old robot.
If the scramble frequency is a little higher than the audio frequency band then the audio might have an upper and a lower sideband, one which could be filtered to make inverted frequencies.

Click to expand...

Ok it is clear that a balanced modulator is needed for this inversion.
The one attached, is a discrete one made for HF but I think it should work on AF too as it is, maybe by increasing the input and output capacitors a bit.

Or maybe a single balanded mixer can be ised **broken link removed** which requires less components. This should reduce the audio carrier and keep only the audio sidebands. It is ok for the two AF signals not to be too isolated I think.

- - - Updated - - -

Now that I am thinking of it, the single balanced mixer won't work nicely, because either the carrier or the input AF will be present at the output.
A double balanced mixer like the gilbert cell I attached should work ok I guess.

You can use the single balanced mixer but you need two of them, a filter and two oscillators to make it invert the spectrum. All it does at the moment is shift the frequency but it does not provide inversion. It also has no supression of the input components so it would be difficult to keep only the wanted part of the spectrum. If you do use a circuit like that, I would suggest the LO should be quite high, possibly in the MHz region, you filter it to remove the LO and unwanted sideband then use the same circuit to down-mix back to back to AF using a slightly higher LO frequency.

Brian.