P33

📖 What Pearson does

Chapter 14 is the boss fight. It walks you through building the Dogbotophone MK1 — a “massive electronic orchestra complete with multiple voices, drums, and self-patching sequencers that compose on the fly”. The build has 15 numbered steps:

  1. An audio-rate oscillator. A 4093-based square-wave VCO.
  2. A sequencer clock. A 4040 binary counter, dividing the VCO into three square waves at three octaves of frequency. These drive the three sequencers.
  3. A 4017 overtone synth. An overtone synth that can chirp out melodies that can’t play out of tune.
  4. A 16-step sequencer for the overtone synth. Walks through 16 steps driven by the 4040’s three octaves of binary.
  5. A 4017 undertone synth. Plays undertones — frequencies below the audible range.
  6. A 16-step sequencer for the undertone synth. Like step 4.
  7. A 16-step sequencer for drums. Triggers the drum voices.
  8. A tunable square wave. A 4046-based square wave VCO, played manually.
  9. Twin-T voice 1. A harmonic drum voice, tunable from kick to tom to woodblock.
  10. Twin-T voice 2. A second harmonic drum voice.
  11. A cymbal (XOR voice). An inharmonic cymbal built from a 40106 (six oscillators) + a 4070 (four XOR gates) + a diode VCA.
  12. Diode AND-gate sequencer reset. A logic gate that resets the drum sequencer.
  13. An optocoupler-controlled sequencer. (Optional.) Uses an optocoupler to trigger the sequencer.
  14. (Optional) Active high-pass and low-pass filters. Sends voices through filters.
  15. An op-amp mixer. Mixes all six voices.

The chapter concludes: “If you’ve made it this far, congratulations! Your room is likely a mess of wires, but you’ve done it.”

🎓 Background: the Dogbotophone vs. our device

The Dogbotophone MK1 is the closest analog to our project in the entire book. It has:

  • Multiple oscillators (analog) ↔ multiple AMY voices (digital).
  • Three 16-step sequencers (hardware) ↔ our chain of patterns (software).
  • Two sequenced melodic voices (overtone + undertone synths) ↔ our melodic slots.
  • Three drum voices (Twin-T + cymbal) ↔ our drum slots.
  • A tunable square wave (the 4046) ↔ our tweak Tone + Knob A.
  • A mixer ↔ AMY’s built-in voice allocator.
  • Active filters ↔ our tweak Filter mode + punch-in FILTER_SWEEP.
  • Optocouplers ↔ no analog.

Our firmware does almost all of what the Dogbotophone does, with much less soldering.

🔧 Try it on the device

The complete Dogbotophone experience, in our firmware, can be built up over a few minutes:

Step 1: three sequencers.

Create three patterns, each with a different “instrument”:

> pattern_clear 1
> step_set 1 1 9 60
> step_set 1 5 9 64
> step_set 1 9 9 67
> step_set 1 13 9 72
> pattern_clear 2
> step_set 2 1 10 67
> step_set 2 5 10 71
> step_set 2 9 10 74
> step_set 2 13 10 79
> pattern_clear 3
> step_set 3 1 11 60
> step_set 3 5 11 65
> step_set 3 9 11 67
> step_set 3 13 11 72

Three patterns, each with its own slot (9, 10, 11), each with its own melodic content. Now chain them:

> chain_clear
> chain_append 1
> chain_append 2
> chain_append 3
> bpm 120
> play

You hear three patterns play in sequence. That is the three sequencers of the Dogbotophone, in software.

Step 2: the overtone synth.

The overtone synth plays melodies that “can’t play out of tune” — they are harmonics of a fundamental. In our firmware, you can approximate this by recording a harmonic-rich sample (e.g. a square wave) and using slot_play on a melodic slot.

Step 3: the drum voices.

Record three drum samples (kick, snare, hat) into slots 1, 2, 3. Write a pattern that triggers them on different beats:

> pattern_clear 4
> step_set 4 1 1 1
> step_set 4 3 3 1
> step_set 4 5 2 1
> step_set 4 7 3 1
> step_set 4 9 1 1
> step_set 4 11 3 1
> step_set 4 13 2 1
> step_set 4 15 3 1

This is the Bossa Nova example from Pearson’s chapter 12 — but applied to our project.

Step 4: the mixer.

AMY automatically mixes all six voices (3 melodic patterns + 1 drum pattern). You don’t have to wire a mixer.

Step 5: the active filters.

Add tweak-filter values to your melodic steps:

> edit
> tweak_filter_set 1 1 50 100
> tweak_filter_set 2 5 100 50
> tweak_filter_set 3 9 200 0
> write_off

When the patterns run, the filter sweeps each step differently. That is Pearson’s “active high-pass and low-pass filters on individual voices”.

🛠 Code reference

  • The pattern step data with filter values — main/sequencer/pattern.h, pattern_step_t. The filter_cutoff and filter_resonance fields.
  • The tweak filter UI — main/ui/knobs.c. Maps ADC readings to step-level filter values.
  • The mixer — components/amy/src/amy.c. AMY’s voice allocator mixes up to N concurrent events into the output buffer.

🚫 What we can’t simulate

  • The undertone synth. Frequencies below the audible range. Our firmware doesn’t generate them.
  • The XOR cymbal. The XOR-based inharmonic cymbal is a beautiful analog circuit. We can’t synthesize that timbre without a custom wavetable.
  • Optocoupler-controlled sequencing. An optocoupler is an electrically-isolated analog component. We have no analog of electrical isolation.