Recently, a user on the Electro-tech-online forum was trying to get a ring modulator working for a DIY music synthesiser. They mentioned that parts, either suitable audio transformers or analog multiplier ICs are often either expensive or difficult to obtain - especially in some countries where shipping costs can be high.
That got me thinking about alternate ways to produce the same results, without using expensive or rare components.
What I came up with was a PWM based concept; switching between the true and inverted versions of the audio input at a very high frequency that could easily be filtered out without affecting the wanted audio.
The PWM duty cycle centred at 50% and linearly modulated by the audio "carrier" input.
The result after filtering is the product of the two input signals - a four-quadrant multiplication, just what is needed for a ring modulator!
This is a block diagram of one channel, to help clarify the operating principle:
The first rough paper design included a built-up PWM generator, however looking at that I realised that in principle I'd created a Class D audio amplifier type circuit - and there are lots of very cheap integrated Class D amp ICs which are readily available!
Looking at various common and cheap types, I settled on the PAM8403, which has several sources. The PAM 8406 is also suitable, just one pin needing a different level. They are surface mount ICs, but small PCB modules with the IC and a few other components are readily available.
I found I need to add attenuation and level control for the carrier input, as the level settings were very touchy and I was getting some distortion. The mic audio level was also too high at times.
The output filter is rather inadequate in this build as I did not have enough capacitors in a suitable range of values to get the correct cut-off frequency - it really needs something like 470 or 680pF and a fourth order filter; this has a 2nd order using 220pF caps.
The intended fourth order filter version is shown in the schematics below, which also has the carrier input adjustment facility etc.
This is the audio from the first test of the prototype, "warts and all" - it's overloading & distorting in places due to excess levels, but I believe it does prove the circuit works; it's compressed to 320K MP3 to reduce the file size:
Audio from initial prototype testing
And the second test session, after adding an adjustment for the carrier input and better setting the output level:
Audio from second prototype tests, after adding a carrier attenuation control
A video showing details of the prototype on my YouTube channel:
This is my prototype, without the extra level adjustments etc. It has a 5V regulator included, which would not be needed in such as a modular synth that has +/-12V and +5V available. Most of the signal connections are made with wire-wrap wire on the underside of the board.
ps. It's also stereo! Or two fully independent channels.
The initial test setup is shown below.
The +/-15V power is from a battery pack (the large grey box) and carrier from the little function generator, running from the USB power pack underneath it and the input to the Focusrite interface is via the small transformer hanging off the end of the board, to isolate everything from the PC electrical noise.
The small second board is a lash-up preamp for the SM58 mic I was using for testing.
Provisional schematics, shrunk to fit the page; component values may be varies as needed as testing continues.
There is also a full resolution .PDF of these at the end.
The .pdf is the definitive current version and will incorporate any minor changes not in the screengrab images
Schematics as a two page PDF:
RJ Ring Modulator Schematic PDF