The ring modulator is one of those effects which have been in the synthesists toolbox since the early bleeps of R2D2. It’s a fun little thing to have even if you don’t really know what it does exactly. It makes things go whrrrrrrbrrrrfliieeep, and that’s cool enough. So, why not build one. I thought to name it ‘The one Ring’ but that’s a bit of a weird and stupid name and I already have more than enough modules with stupid names. So I ended up calling it, very unimaginative, ‘Ring modulator’, if you have a better idea, please, send me a postcard with your proposal.
What’s a ring modulator?
For frequencies A & B (fA and fB) it will yield the following at it’s output. fA+fB and fA-fB. So, if A and B are perfect sines at 400Hz and 300Hz you would see a signal on the output which consists of a 700Hz and a 100Hz sine wave without the original frequencies of the input present. You could call it frequency mixing and throw a lot of fancier math towards it, but that wouldn’t help in understanding the effect. It’s that simple, you can see it easily on a frequency analyser. With more complex signals for A & B, the output will become a lot more interesting and a lot more difficult to predict. A ring modulator is often used to create inharmonic signals since the resulting frequencies are generally not harmonically related to the input signal. Simply put, it turns your incoming signal in an inharmonic mess. It’s a useful tool in synthesis for all kind of inharmonic type sounds like bells, chimes, metal hits, percussion and, of course the all time favourite, weirdness.
The easiest way to build a ring modulator is by having 2 transformers and a ring of diodes in between them ( google “ring modulator schematics” and you find it easily ). It looks a bit like a bridge rectifier, but it’s not, the diodes are oriented differently and form a nice ring. The original design is a completely passive circuit and pretty easy to build. The downside however is those pesky audio transformers which are bulky and can be tough to source.
So, I wanted to make one without any transformers. I found some information to go on in some documentation on the subject while googling around. It comes as no real surprise, it’s also based around the typical diode ring, but instead of transformers, opamps are used. Now it was a bit of tricky thing to mimic the whole transformer behaviour but if you look long enough at it, it starts to make sense. Or it doesn’t, depends. It’s a bit of mixing waveforms together to insert the right thing into the diode ring and get the right thing out.
But that’s the core of the whole thing. 4 opamps, a bunch of resistors and 4 diodes. So, now for the added fluff.
A ring modulator on its own is just 2 inputs and a single output, which is a bit sad really. You’d also need 2 oscillators to get something out of it, which isn’t always an option. Luckily I do have a simple oscillator circuit design on hand, as used in the ‘can I kick it‘ module which can provide a carrier wave if nothing is present at the carrier input. Add a waveshaper to it, also ‘as seen in’ the ‘can I kick it‘ module. Add CV control to that oscillator and you already have the start of a more interesting module.
That leaves me with one half of an LM13700 with nothing to do. And what does one do with a spare OTA? The most useful thing I could think of was adding a VCA to the carrier level so it can be controlled through CV. How else would we manage to get it to do wobwobwob.
Speaking of wobwobwob. It would be nice to add an LFO mode to the carrier VCO. And turns out that’s easy to do. All oscillator designs are based around the cyclic charging and discharging of a capacitor (*) in some form of another. That capacitor is basically the core of the oscillator. By controlling the amount of current going to/from the capacitor we can control the speed at which is charges/discharges, hence, change the frequency. An OTA, like in this oscillator design is ‘a way’ to achieve this through CV control. However, there are limits to the whole idea which limit the min/max frequencies you could reliably get out of this oscillator design. To get lower frequencies you would need to decrease the current going to/from the capacitor. At some point the current involved is getting so small you’re starting to run into issues. Another way to change the charging/discharge speed of the oscillator would be by varying the capacitance of the capacitor. So we can add a switch with to a bigger capacitor and the oscillator frequency will drop in LFO range. Much easier then designing an oscillator which can sweep through the whole LFO-VCO range in one smooth motion. If you wonder. Yes, you could also use a smaller capacitor and make an oscillator that’s well beyond human hearing capabilities, but apart from entertaining bats or dogs I see little use.
*I’m sure some smart-ass will correct me on this as there probably is some esoteric design out there which uses a totally different approach.
So, we now have a voltage controller carrier wave ranging from LFO to VCO range, with a variable waveform tri-sin-sqr and CV controlled level control. Seems to cover the basics pretty well. Since the ring modulators output doesn’t contain the input – well, in theory it shouldn’t, in reality there will be some leakage. Although the opamp design seems to leak far less then the design I made using transformers. It were cheap-ass transformers though. Nevertheless, I thought it would be a fun idea to actually add a blend pot by which one can mix the input/output.
Any ringmodulator can be used as a VCA if you want. Applying an envelop to the modulator input will result in a behaviour that’s similar to a VCA for the signal present at the carrier input. In this case that’s either the internal oscillator or the signal at the carrier input. It will go through zero and come back out the other end, so it can be used creatively.