The same LO (or BFO if you like) signal generated within SA612 is also routed to MC1496 as its carrier signal. I used here that IC because it has oscillator already built in which is very convenient and simplify circuit to some extent. Such signal is then entering product detector based on SA612. After that, signal is limited in three stage differential limiter with 85dB gain and passed again through another identical crystal filter to remove all spurious product generated in limiting stages.

On its output, MC1496 produce DSB signal that is then passed through 2,7KHz wide 8-pole crystal filter with 501,650KHz central frequency. First mixer is old but excellent double balanced mixer MC1496 chosen before ubiquitous SA612 because of better balance and carrier suppression and variable gain capability. Both stages are done with one dual OPAMP LM358. Input is gain variable microphone preamplification followed by soft limiting stage to slightly per-shape audio prior to first mixing. My design is not new and is based on well known overall concept. I am an “analog chap” and my design is pretty much the only way I am capable to solve the problem. RF speech processing can be done in wide variety of configurations from quite simple and cheep with one IC and small ceramic filter up to much more sophisticated and complex designs in a mixture of analog and DSP stages.

Thanks to a generous donation from one friend, I recently got as a gift two identical SSB Crystal filters from an older military radio and finally got a chance to try my version of RF Speech Processor brewing for years in this half a pint of gray matter I still have (I hope?). DIY is not very common activity this days as it was up to decade or two ago, but I am still deep in those waters.

Ham radio speech processor