DEVELOPMENTS
The seven organs built during and following the 1980's were all analogue instruments, using frequency-dividers for reasons of cost. This detail may discourage you from reading further, as such organs justifiably have a bad reputation on the score of a sterile sound. This serious deficiency was avoided in the case of the five church instruments, by drawing upon three divider strings fed by crystal oscillators detuned slightly from the nominal frequency of 2 MHz. What this meant in practical terms, is that if a note was played with 8', 4' and 2' stops all "drawn", then three independent notes (although approximately octave-related) would be produced. In other words, a true chorus effect was maintained. In the five church organs, diode switching of either square waves (for the stopped flutes) or rectangular waves (for the principals or trumpets) was implemented, followed by a multitude of mostly passive, but in some cases active, electronic filters to achieve the desired steady-state harmonic spectra. The carefully-designed switching system allowed the build-up of a true chorus, where the sound level of individual ranks could be easily adjusted, and quite good chiffs could also be produced. This meant considerable economy in terms of circuitry, yet the "missing note" problem and the difficulty of balancing a chorus (troubles well-known in unit-type organs) were both overcome.
The two organs built for private homes were an attempt to produce a cheaper version of the highly successful church line, but using 4016 CMOS keyers instead of diode gates, and drastically reducing the number of filters involved. Another expedient on the score of cost was to have only a single frequency-divider string. Early in my project (i.e. in the late 1970's) I had abandoned the use of 4016's as keyers, because there was no way of improving the on/off transients (attacks/decays), which were much too fast. However, by using the 4016 as a chopper rather than as an analogue switch, i.e. by feeding a square wave or rectangular wave (as required, and very easily obtained from frequency-divider circuits) to the gate of the device, and a D.C. voltage (switched from the keying contacts) to the input, the operation was transformed. The D.C. voltage could be applied through a simple RC network, allowing excellent and very controllable attack/decay times. The output from each 4016 section was now a square wave or rectangular wave signal whose envelope-shape could be controlled almost at will. Appropriate filters followed the 4016's.
These two organs have operated very acceptably now for about 15 years. I now own the first of them, its original owner having died. Although there have been improvements which have been made or are pending, my many years of playing this organ have convinced me (even with my very fussy standards) that the basic design holds great promise for an instrument which can be made by enthusiasts at home (as this one was, totally), which is reasonably cheap because you use your own labour, but which has virtually unlimited capability for musical excellence of a very satisfying standard (admittedly a subjective assessment), and where the builder can actually voice it according to his/her own tastes (and without having to have esoteric technical knowledge to achieve this).
Although the commitment to frequency-division may seem disappointing, the disadvantages of this economical route can, I believe, be successfully overcome, by the use of multiple phase-shift circuits in the output stages, and by the use of multiple divider strings if the builder so chooses. I also have some ideas for "cheap but stable" free phase oscillators for those who feel they must go down that route, but as they say in engineering, "Nothing works till it works".
My plans include circuitry to insert appropriate wind noise along with the tones, together with improved starting transients. My practice of distributing the tones across a number of audio channels is also to be further developed. With all the organs so far, the sound comes from three separate sources as one plays across the keyboard. I plan to have each six notes across the keyboard coming from its own source, adding greatly to the quality of the sound.