This is maybe the one to many FV1 design. But, if you care to know how it works in more details please consider reading further.
The idea of designing an effect based on the FV-1 chip came when I realised that many reverb effects are built around that chip. And this is due to the amount of DSP processing that is required to create pleasant reverbs.
With this design I wanted to achieve a few things:
- loads of memory for many DSP algorithms
- a single rotary encoder to scroll through all algorithm
- CV control over all parameters including sample rate of the DSP
- attenuverters on all inputs
- wet/dry blend output
- stereo signal path
The stereo input on the top left of the main schematic sheet is wired so that the Left signal is duplicated on the Right channel if nothing is connected. Right after the input a resistive divider and a band pass of the first order filters out DC, high frequencies and make the input levels to the DSP input requirements. At the DSP input the signal is forked to a couple of buffer to feed the dry/wet mixer on the other side while maintaining the high input impedance. The dry/wet mix is done by mean of a resitive divider and suming op-amp that uses a double wiper potentiometer to act on both channels at the same time. The mixed signal is raised back up to Eurorack level at the same time.
The bank selection of the DSP works by looking through the pages of a 24LC32 chip. In order to increase the number of banks two chips are present on board. In order to address them sequentially a coded rotary encoder scrolls through the algorithm with its 3 least significant bits through the program selection pins of the DSP. While the most significant bit is used to toggle the address of the two 24LC32 chips. Thus, on half of the full encoder range the top memory is addressed while on the rest of the encoder range the second memory is addressed.
In order to use CV inputs in place of the potentiometers, for which the DSP chip was initially designed, they have to be rescaled. The schematic above shows, from left to right, attenuverting of the CV, summing with common mode potentiometer and scaling to 0-3.3V range. The atennuverting makes interaction with modulation source easier by allowing a rescale and phase inversion of the input modulation. The common mode potentiometer allows to ajust the effect without modulation sources plugged-in, and offset the DC level of the modulation, when one is present.
The power supply features three functions. It takes the input voltage from the Eurorack power connector (keyed, but not further protected, one just has to be careful). It creates 3.3V, in two rails for analog circuits (op-amp, filters) on one and digital circuits (DSP, memory, clock) on the other. Finally it creates a -10V reference that is used for scaling the common mode control the the Eurorack CV levels.
The sampling rate of the DSP is directly derived from the crystal resonating frequency. An algorithm can get low-fi properties that are interesting from modulating the sampling rate. Thus the option of a voltage controlled clock was added thanks to the LTC6990 chip to replace the crytal for which the Spin FV-1 was originally designed.
For the Spin ASM IDE to recognise the EZ-USB FX2LP and the module as being a dev-board, this 4066 based isolator allows the separation of the I2C bus of the module from that of the devboard. This makes the Spin software recognise first the EEPROM from the EZ-USB which uppon detection disables the isolation and allows reprogramming of the EEPROM that holds the alorithms.
And with that we have covered the design of this Eurorack effect module.