Agentic AI Designs an Analog Synthesizer
Agentic AI Designs an Analog Synthesizer
The author used Claude Code with Opus 5.5 to design an MCU-controlled analog synthesizer from a blank project folder, including a Juno-style DCO, a Moog-style transistor ladder filter and a VCA, with circuits simulated in ngspice and a Python behavioral model calibrated against the simulation. The PCB layout required the most human intervention, with two-layer routing failing before a four-layer board fully routed, and the design took 19 hours from first prompt to ordering. After the boards arrived, bring-up was automated via audio loopback and an oscilloscope, SPICE simulations matched the hardware closely, and the only hardware fix was removing a hand-soldered solder bridge on the MCU.
Source: Hacker News AI · cpldcpu.github.io