I opted for musical capabilities similar to the classical MiniMoog analog synthesizer, with fully analog sound generation and signal path and also an analog control signal and modulation path.
The analog sound part is built from different physical modules:
The control voltage section which effects contour and modulation control is built from
There are further two auxiliary modules
The modules are interconnected in a fixed manner like in the MiniMoog.
I did not implement the MiniMoog "external input" and 440Hz test tone features though.
Unlike the MiniMoog there is no separate "audio mixer" module as this function is implemented in a different way.
The analog part does not have any physical potentiometers or switches or buttons, rather the analog part is fully voltage controlled.
The control voltage lines are inputs only and are standardised as two types:
These control voltages are generated by an electronic interface.
The Digital Interface links the Analog Control Interface to the outside world. It contains analog to digital converters and logic control outputs and digital storage for one set of settings (patch).
The Digital Interface is very basic. It contains just enough functionality to receive MIDI control messages and generates 32 analog voltage outputs with and 32 bits of logic control signals (CMOS level) to control the analog sound generation part.
I followed four principles:
The main technology used for the analog part is bipolar transistors, either as single devices or in OpAmps (well I also use some with JFET inputs). CMOS devices are mostly used in electronic switches for analog signals or a digital circuitry.
A major design consideration is signal voltage range.
Given the transistor technology I decided for a signal range of -5áV < Vsignal < +5áV symmetrical around 0áV ground for both the audio signal and the analog control voltage. The audio signal may not use the full 10áVpp range. In fact I often work with a signal current in the range of -0.5ámA < Isignal < +0.5ámA. Flowing into a load resistor of 10áKOhm this will yield the mentioned voltage range. Working with current sources has the benefit to be able to easily sum audio signals. I use this also for modulation and contour signals.
The -5áV < Vc < +5áV range applies also to the control voltages that replaced the manual potentiometers. These control voltages are generated by digital to analog converters (DACs) in the Digital-In module.
There is another type of control voltage Vs that is digital and either 0áV or +5áV. Those control voltages are generated by CMOS type digital circuitry in the Digital-In module.
Classical VCOs usually contain wave form generation, amplification and level shifting.
Classical VCA designs also employ level shifting and usually attenuation to employ current controlled transistor pairs for "amplification" control.
I decide to pull the VCA functionality into the VCO module to save some of this amplification/ attenuation.
Classical synthesizers have the VCA after the VCF in the signal path. Pulling the VCA into the VCO thus reverses the order of signal processing. In an ideal linear world this would not matter. But these circuits are not perfectly linear.
Thus the UBZR1 will behave differently than a classical synthesizer when pushed to the extremes: effects to be discovered!
For the signal voltage range I also considered a signal range of 0áV < Vsignal < +5áV symmetrical around a notional 2.5áV. That would allow for use of standard DAC circuits running from +5áV only.
The power supply could then simplify to +10áV, +5áV and -5áV with further reduced power consumption.