This workshop is about adapting the architecture of some traditional analog synthesizer oscillator sections for use on the Nord Modular. Some analog synthesizers are taken as example and some patches are named like these synths. However the sounds on the Modular will have their own character and are not meant to be exact copies of original analog synths. Instead they are only meant to be inspirational for the Modular user.

The oscillator section of a synthesizer contains the waveform generators that provide us with the basic timbres at a controllable pitch. The simplest monosynths, like e.g. the Roland TB-303, MC-202, SH-101, the Yamaha CS10, the Korg MS10 and MS50, etc., feature only a single oscillator in the oscillator section. However, these synths are able to generate at least a sawtooth waveform and a pulse waveform (and some noise ;-). In general it will be possible to make a mix of these waveforms, effectively giving control over the balance of even and odd harmonics in the resulting waveform. As the two waveforms are generated by the same oscillator, there will be a tight relationship between the timing of the two waveforms.
In general the pulse waveform will be derived from the sawtooth waveform by means of a waveform processor, build into the analog oscillator.

The Nord Modular oscillators do not have multiple waveform outputs. Instead slave oscillators can be used to generate multiple waveforms. To get the tight waveform relationship between the waveforms hardsync can be used.

Single oscillator sections

Instead of using slave oscillators, which use up relatively much DSP-resources, an interesting single oscillator section with multiple waveforms can be patched by adding some waveform processors to a single oscillator module. The purpose of this section is to get a full, analog sound with the least amount of DSP-resources possible. OSC B is the best choise for the oscillator, as a hardsync input is not needed. This oscillator offers a choise of the four basic waveforms and PWidth modulation.

A cheap trick to 'beef up' the sound of a single oscillator analog synth was to add a suboctave divider. You can find this feature on a.o. the Korg Polysix and the Roland SH-101. Electronically the suboctave was generated by feeding the output of the oscillator into a digital frequency divider chip. A commonly used chip was e.g. the CMOS 4520, costing less than a dollar and featuring division rates of 1:2, 1:4, 1:8 and 1:16. The outputs are squarewaves 1, 2, 3, or 4 octaves below the oscillator frequency.
On the Modular we can patch a basic suboctave divider by using a S&H module and a logic inverter module. The S&H is clocked by the inverted oscillator output to avoid the siren-like vibrato when the suboctave squarewave would follow PWM on the oscillator. To do this an Inverter/Levelshift module from the audio tab is inserted between the oscillator output and the S&H clock input.
The output of the S&H is routed to the input of a Logic Inverter module from the Logic tab and the output of the Inverter is routed back to the input of the S&H. Every time the oscillators wave crosses zero in the positive direction the output of the S&H is flipped between either 0 units or 64 units. This effectively creates the suboctave squarewave. Before mixing the output of the S&H it is a good idea to subtract a constant value of 32 units by an Adjustable Offset module from the Mixer tab set to -32 to avoid chances of signal overflow later.

The level of the suboctave is half the level of the oscillator output. This is no problem at all, as now the suboctave level is more in balance with the levels of the sawtooth, the triangle and the sinewave.

As you can hear the mixed waveform gains a lot of 'beef' in the low registers.

If the waveform that is used to drive the suboctave generator is a pulsewidth modulated pulsewave the PWM effect can be transferred to the suboctave by a Gain Controller module. The suboctave signal is actually used to suppress every one of two cycles of the waveform. The cycle which is not suppressed still features the PWM effect, so that's how it's transferred to the suboctave waveform.

Using an envelope to control the pulsewidth and 12 dB filter creates a fuzzy bass. A clipping distorter and an overdrive are added to the patch to quickly change between a duller or brighter sound.

Using two suboctave dividers together with a PWidth modulated squarewave oscillator and adding a Chorus module can create a vintage "String" sound, like the Arp Omni or the Solina String Ensemble.

The Omni sound is also very useful for input to a Vocoder module.

Another trick to beef up the sound of waveforms is to use clipping. The Clipper module from the Modular should be used in a 'feedback' fashion to make up for the signal loss with high clipper settings. This is done in a way as can be seen in the next patch.

The Clipper module should be set to Sym to avoid signal overflow on the outputs of the Modular. Why this feedback works is because the Clipper module still clips at a level of plus and minus 64 units when the Clip controlknob is closed. By tweaking the feedback control on the feedback mixer instead the clipping level can be controlled while keeping a constant signal level on the output of the Clipper module. The effect on a sawtooth and triangle wave is that the level of the fundamental frequency is slightly increased, so the sound gains so more 'beef' in the low.
On the sinewave harmonics are introduces, making the waveform more suitable for filtering later on.

Another interesting addition to this circuit is to add a LFO and add its output signal to the feedback mixer. This will emulate the effect of PWM on the clipped sawtooth, triangle and sine waveforms, making them sound more lively.

As you can hear a subtle phasing effect is added to the sound. What happens to e.g. a sawtooth waveform can be seen in the next illustration.

The green lines are the original sawtooth waveform as fed into the mixer. The slowly rising blue line is part of the LFO signal. The feedback raises the result of the original sawtooth and the LFO to the high level blue sawtooth signal. The Clipper module cuts off the signal at the dashed grey lines and the red signal is resulting signal at the output of the Clipper module. As you can see the clipped sawtooth varies in width at the rate of the LFO. The sound still has the audible characteristics of a sawtooth but with a bit more energy in the fundamental and a soft PWM-like effect added. Still the modules used to create this effect only use 2.7% DSP-resources.
The LFO level shouldn't be too high, a level setting on the feedback mixer around 90 to 100 seems to yield the best results.

When adding a suboctave to the patch an oscillator section cheap in DSP-resources, but with a full and lively sound is created.

This basic oscillator section is very useful for bass patches or for polyphonic patches where you need a lot of polyphony. Adding a simple filter and ADSR will give you a simple but versatile 'analog' patch.

Instead of the OSC B the Spectrum OSC module can be used like in the next patch. The Spectrum OSC is especially useful for sounds that should be played in the higher registers.

The SH-101 waveform mixer features sliders for a PWM-wave, a sawtooth wave and a SUB OCT which follows a switch that chooses between three suboctave waveforms. The suboctave can be a squarewave one octave or two octaves down, or a 25% pulsewave two octaves down. To get both the sawtooth waveform and the PWM-waveform an extra slave oscillator is needed next to two suboctave dividers. Here is the patch.

The master oscillator is hardsynced to the inverted slave sawtooth oscillator to get the right phase-relationship between the waveforms. Controlknob1 on the output mixer sets the level of the pulsewave, controlknob2 the level of the sawtooth and controlknob3 the level of the suboctave. The original SH-101 features three types of suboctaves, a squarewave one octave down, a squarewave two octaves down and a 25% pulsewave two octaves down. In this patch another type is added, the pulsewave one octave down that inherits the PWM generated on the main oscillator. The 4-1Switch module lets you choose between the four types of suboctaves.

A pair of two sinewave oscillators where one receives linear FM from the waveform output of the other has become known as a FM operator. In the FM workshop there is lots more on linear FM. Although FM has only become popular when super stable digital sinewave oscillators became available, it is worth exploring how such an operator performs when used as a substitute for a traditional analog oscillator in a traditional VCO -> VCF -> ADSR patch. A disadvantage of a two sinewaves FM operator is that the waveform lacks sharp transients. And as filters like transients, it is good to add some controllable clipping. The basic patch looks like this:

This type of oscillator seems particular suited for bright bell-like sounds.

FM modulation on the grey Mst input gives a FM-type sound where you can sweep a strong formant area through the soundspectrum. Again, adding a clipper can add some pleasing brightness to the sound. In the following patch the modulation on the grey input is switched by a suboctave pulsewave between a grey Slv output and a sinewave output of the master oscillator. Keep in mind that modulation of a grey input by the output of another oscillator will make the modulated oscillator inherit the pitch of the modulating oscillator.

Dual oscillator sections

There are several reasons why a second, independently controllable, oscillator is found on analog monosynths. The most commonly known reason is to fatten the sound by slightly detuning the oscillators to give a nice chorusing effect. Detuning the oscillators by an octace gives the suboctave effect without the need for a build-in suboctave generator. As the suboctave can also be slightly out of tune the sound is more lively than that of the suboctave generator, which always has a tight phase relationship to the main oscillator.
But having two oscillators also gives the possibility of modulating one waveform by the other waveform. Two types of "inter"-modulation are commonly found on dual-oscillator analog synths, RM (Ring Modulation) and FM (Frequency Modulation). RM works on the level of the waveforms, it actually multiplies the two waveforms together. FM works on the "timeline", by actually pushing the waveform back and forth in time on the rhythm of the other waveform. Both effects can introduce strong inharmonic frequency-components to the modulated sound, so they should be treated with care. FM is even capable of transforming a sound into noise at very high modulation levels.

A ringmodulator is actually a circuit that performs an arithmetic multiplication of the two input signals. As the sign of the signals is included in the multiplication the circuit is also known as a balanced- or fourquadrant multiplier. Although the Modular has a discrete Ringmodulation module, there are several modules that actually feature ringmodulators. In fact the Gain Controller module is the real traditional ringmodulator. All modules that feature an AM (Amplitude Modulation) input actually perform RM on this input, as the input will accept signals with a minus sign and invert the output if the sign is minus. So, when connecting a Constant module to the AM input on a 'OSC Slave A' module will not only control the output level, but invert the waveform if the Constant module is set to a "negative" value between 0 and -64 units.
The difference between AM and RM is that AM does not accept negative values on one of the controlling inputs. If the value is negative it is treated as a zero value and the signal on the other input is suppressed. Although the Gain Controller module has a button to shift up one input, that doesn't really make the module an AM module as values lower than -64 units do not silence the output. To get a real AM module a Diode module from the Audio tab has to be added in front of one of the inputs.
So, the designation AM on several modules is not really proper, it should be RM, only it is not expected that every synth-player on the planet knows the technical difference between AM and RM. What you need to remember is that the AM inputs on the Modular can be used for both AM and RM, depending on whether the controlling signal is unipolar (a signal swing between 0 and a maximum of 256 units) or bipolar (a signal swing between a maximum of -256 units and 256 units). Also remember that 64 units is arithmetically 1, or in audio terms 0dB or unity gain.

In the "Logic modules" workshop is an extensive explanation of the signal levels in the Modular and how they relate to the units and knobsettings used on the modules, so go there if you want to know more about this subject. But here the goal is strictly musical, so no more of this technical stuff. The only reason this is mentioned is because AM inputs and the Gain Controller module will be used for RM in the next examples. So, when the name ringmodulator is used, it might in fact be the Ringmodulator module, the Gain Controller module or an AM input on a module that features one.

The musical effect of RM is that, when two waveforms of different pitch are ringmodulated, new frequency components will appear in the signal. When feeding two sinewaves of frequencies 300Hz and 400Hz into the two inputs of a ringmodulator, the output signal will no longer have these two frequencies present, but they are transformed into the following sinewave frequency components:
400 - 300 = 100Hz
300 - 400 = -100Hz (100Hz with reversed polarity)
400 + 300 = 700Hz
300 + 400 = 700Hz

When toggling the button on the "Mix | RM" switch module the audio will change between the 300/400Hz chord to a 100/700Hz chord. The 100Hz and the -100Hz component don't cancel each other as there is some phaseshift between them that compensates for the cancellation.

There are two rules to keep in mind when you want your sounds to remain harmonic when using both RM and FM. The first is to keep the frequency ratio between the two oscillators simple, ratios of e.g. 1:2, 2:3, 1:1.5, 1:1.25, etc. are expected to give good results. The second rule is to use hardsync if the the ratio is complex like 1:1.58634, which in fact forces the ratio back to 1:1.

To see how a dual oscillator section might be patches the Korg MS-20 will be used as an example. This monosynth features switches to choose between the waveforms on both oscilators, however the waveform switch on oscillator B has a setting for the ringmodulated product of oscillator A and B. No extra waveshaping circuits or suboctave dividers are present on the MS-20. Both oscillators can be detuned separately. On the Modular the routing of the signals is slightly different as the waveform switches are on the oscillators themselves, but for practical purposes the patch can be considered the same.

The output of OSC A is inverted so, if hardsync is used, OSC B will sync to the transient of the OSC A sawtooth waveform and not follow possible PWM on the OSC A pulse waveform. A 4-1 Switch module switches between the inverted output of OSC A and a fixed level of -64 units. The output of the switch is routed to the AM input on the slave oscillator. This switch is used to toggle between RM on the output of OSC B or the standard, but inverted (-64), waveforms of OSC B.
A single toggle switch between the inverted output of OSC A and the Sync input of OSC B is used to set hardsync on or off.
Note that a single toggle switch cannot be used to toggle RM on or off, as the output of the switch module is 0 units if it is off. That would effectively silence the output of OSC B, which is not what you would want, of course.
FM is very simply implemented by directly connecting the inverted output of OSC A to the FM input on OSC B.

In a complete patch several provisions have to be made to allow modulation signals to reach their destinations. 4-1 Switch modules are very suited for this task. The next patch is a more or less complete emulation of the MS20. Try if you can recognize the Sync, RM and FM routings as were described above.

The additions in the oscillator section are a 4-1 switch to control the FM input on OSC B. It chooses between [1] a fixed amount of detune for chorusing purposes, [2] FM from the LFO, [3] an upward frequency sweep from the modulation AD generator or [4] FM from OSC B through the modulation AD generator. Use controlknobs 2, 3 and 4 on the switch module to set the level of modulation.
Another 4-1 switch is added to control the AM input on OSC B. It chooses between [1] a fixed volume, [2] amplitude modulation from the LFO, [3] the blue envelope signal from the modulation AD generator or [4] RM from OSC A. Use controlknobs 2 and 3 to set the level of modulation.

Pressing the toggle button on the KEY|SEQ module at the upper right hand corner makes the patch switch between a polyphonic keyboard patch and a sequencer patch clocked by the TEMPO module in the lower left hand corner.

Some analog synthesizers feature three or four oscillators. A good example of a triple oscillator section monosynth is the Minimoog. This synth was designed to be a relatively cheap performance synth compared to the big modular Moog synthesizer systems of the late sixties and early seventies. So its no surprise that many of the Minimoog features are decendents of those big systems. On the old Moog System 900 (around the year 1967) there is basically no distinction between a VCO and a LFO. One would typically find three VCO's connected to an Oscillator Control module. The control module handles common modulation signals for the VCO's and each VCO has its own detune and octave control switch. The octave switch has a special setting LOW, which turns the VCO into a LFO, tracking the Oscillator Control module.

The octave and LFO switching is done by manipulation of the grey Slv signals. Such a VCO section is particularly suitable for all sorts of strange effect patches.

The Minimoog oscillation section is quite similar, three VCO's from which one can be set to the LFO range by a switch. On the Modular however it hardly pays to provide for a VCO that can be switched into the LFO range, as the LFO A module can do anything and more (control of the phase after a reset) than a VCO would. So for many of your Minimoog sounds you could use the following section.

The Minimoog shares with the Sequential ProOne the possibility to route the signal of a VCO to the cutoff parameter of the filter. This is a really nice feature that I would recommend to explore on the Modular as well. Learn more about this in the Filter Section workshop.

The analog monosynth with the most elaborate oscillator section must be the four VCO Korg MonoPoly. This synth can be played with four voice polyphony where each oscillator will follow its own key. However the filter and the VCA/ADSR must be shared amongst the four voices. But when used in monophonic mode the oscillator sections performance is unsurpassed. The oscillators can be used as four independent VCO's, where the last three can receive hardsync, RM and FM from the first VCO. The VCO's can also be split in two independent sections where VCO2 receives its audio modulation signals from VCO1 and VCO4 receives them from VCO3. To patch the MonoPoly oscillator section involves quite a bit of miscellaneous modules to make up for easy switching between all the modes that the MonoPoly can operate in.