Though the LFO is not the most important unit of the synthesizer, it is described at the second place because together with the keyboard-interface the first printed circuit board is built up by it in the chain of the PCBs counted from left to right.
An LFO needs different wave forms than a VCO because the wave forms are audible only indirectly by modulating the VCOs and the filter in their frequency. So it's for example not only important that we have a ramp (sawtooth) function but we also have to distinguish between a falling and a mounting ramp.Beside these two functions, the LFO introduced here has an additional triangle function, but no square wave output.
IC1 and IC2 work together as a proven ramp-oscillator: The loading speed of the integrator's capacitor (C1) is controlled by an adjustable input voltage. IC2 works as a trigger with hysteresis which discharges the capacitor via a diode if the integrator's output voltage is mounting above a certain value. By the way this circuit also could work as linear VCO in a synthesizer if you use the CA3140 for both IC1 and IC2, but as it is complicated to change its output waveform into a triangle shape, I used another solution for the VCOs. IC6b and IC6c work as signal inverters and care for the right offset and amplitude adjustment at the same time.
First set the amplitude of the first ramp with P2 (4 V for a first test) and then set the offset of the signal in a way that the lower tip of the ramp just touches the zero line (oscilloscope necessary).
This is important because only this setting guarantees that the VCO starts mounting linearly from its present pitch if you turn the modulation potentiometer from zero to the right. Otherwise, the pitch of the VCO would go down in this case, which would lead to irritations and an unpleasant sound.
Repeat the same procedure with P4 and P5 and keep in mind that you always have to adjust P2 and P3 first as the second ramp is also influenced by these potentiometers.
The triangle function is not created by a waveform conversion of the ramp generator output because this would take too much additional chips. In opposition to a VCO where the frequency must stay exactly the same after changing the waveform, this is not important at an LFO because the human ear cannot recognize a small difference in the frequency of a modulation source. Therefore the triangle oscillator works as independent and different circuit, whose frequency is approximately the same as the one of the ramp generator - which is simplifying the LFO a lot.
The circuit, realized by IC3 and IC4 works similar to the one used for the ramp oscillator, but the integration capacitor (C2) is not discharged by a diode but switched by IC5 (part of the CMOS-switch 4066) to the inverted input voltage value, which leads to a continuous and linear charge and discharge of C2 and which results in a triangular waveform.
IC6a works as output buffer. IC7 is responsible for the amplitude (P8) and the offset (P7) of the output signal. The offset adjustment is different here: the triangle must be symmetric to zero. This is necessary because of the fact that (in opposition to the ramp signal where the synthesizer is "abused" as a siren) the triangle can be used to bring a little "life" into a played note when turning the modulation potentiometer of the VCO a little to the right at the end of a long note - making the melody sound a little "spooky" sometimes. If there would be an offset in the triangular shaped modulation signal, the pitch of the played note would mount or sink the more you turn the modulation potentiometer (or stick) to the right, which would sound like playing in a wrong key then.
If you turn the LFO frequency potentiometer (on the front panel) completely to zero, there will occur a DC voltage at the output which will detune the VCO in case the modulation potentiometer is not completely turned to zero. Using a capacitor at the LFO output would avoid this little blemish, but would also create a new problem: the distortion of the linear ramps of the low frequent signals. So I recommend to put a small resistor (depending on the value of P1) between P1 and ground. Because of a lot tolerances in the LFO behavior, I recommend an individual test to find out the right value of Rx.
To get a bigger variety in modulation frequency, the next version of the PCB will probably have a space for two additional integration capacitors to allow even lower-LFO frequencies.
The maximum amplitude of the modulation signal can be set by individual taste. Please keep in mind that, if the amplitude is too high, the frequency mounting or even the sound will be cut off when reaching the upper VCO input range by pressing a key on the right of the keyboard or be selecting the highest octave. The same goes for the VCF frequency. So its necessary to always find a good individual compromise by checking different settings and judging the results.