# Claude Opus Touchsynth: Agentic Analog Synth From Blank Folder

> Source: <https://explainx.ai/blog/claude-opus-touchsynth-agentic-analog-synthesizer-pcb-2026?utm_source=rss&utm_medium=feed&utm_campaign=blog>
> Published: 2026-11-16 00:00:00+00:00

**Tim Gerhard's November 14, 2026 write-up [Agentic AI Designs an Analog Synthesizer](https://cpldcpu.github.io/2026/11/14/touchsynth/?ref=explainx) documents Touchsynth — a mixed-signal board where Claude Opus 5.5 inside **Claude Code** handled circuit architecture, **ngspice** benches, **KiCad** layout scripts, firmware, and SSH-controlled lab bring-up.** Hacker News surfaced the story as one of the first **public, reproducible** examples where an agent crossed **schematic → simulation → PCB → audio** with only **floorplan** and **touch-key** fixes left to a human.

If you build with **[Claude Code commands](https://explainx.ai/blog/claude-code-commands-complete-reference-guide-2026)** daily, Touchsynth is the hardware mirror of **[Cinecode film skills](https://explainx.ai/blog/cinecode-claude-code-skill-films-from-code-three-js-ffmpeg-mit-github-trending-2026)** — agent ownership of **multi-file artifacts**, not a single README summary.

Claude Opus Touchsynth PCB 3D render of agent-designed analog synthesizer board
*Image: [Tim Gerhard](https://cpldcpu.github.io/2026/11/14/touchsynth/?ref=explainx)*

## TL;DR

| Question | Answer | 
|---|---|
| MCU | **CH32V003** RISC-V | 
| Architecture | **Juno-style DCO** ,**Moog ladder VCF** ,**VCA** , touch keys | 
| Tooling | **Claude Code (Opus 5.5)** ,**KiCad 10** ,**ngspice** ,**Freerouting** , Python glue | 
| Calendar time | **~19 h** prompt → PCB order (per author log) | 
| Human edits | **PCB floorplan** ,**4-layer routing strategy** ,**touch sensitivity** | 
| Primary source | [Gerhard blog](https://cpldcpu.github.io/2026/11/14/touchsynth/?ref=explainx) and[TouchSynth GitHub repo](https://github.com/cpldcpu/TouchSynth/?ref=explainx) | 

## Why this post matters for AI builders

Most **agentic coding** benchmarks stop at **LeetCode** or **web apps**. Touchsynth forces **physical constraints**:

- **JLCPCB free parts** cap op-amp and capacitor choices — the agent could not spec a fantasy OTA.
- **ngspice** either passes or**smokes** the ladder filter — no hiding behind TypeScript mocks.
- **DRC and routing** fail loudly when analog blocks crowd digital touch sensors.

That is the same maturity jump **[Edward Witten lattice paper with Opus](https://explainx.ai/blog/edward-witten-claude-opus-5-5-lattice-chiral-fermions-arxiv-2026)** hints at in **math**, but here the proof is **audio on a scope**.

Claude Opus Touchsynth agent flow covers analog PCB design from ngspice to KiCad routing

## Gerhard methodology: spec-first, not patch-first

Gerhard states he follows a **V-model** discipline:

1. Let the agent propose concepts from the **JLCPCB parts CSV** (1,586 lines pulled via JSON API).
2. Lock a **written specification** (Juno DCO + ladder VCF).
3. Generate **ngspice** test benches**S1–S7** and a**simulation report** .
4. Build a **Python behavioral model** , calibrate against SPICE, render**audio demos** .
5. Only then export **KiCad** sheets and**Python placers** +**Freerouting** .

He argues **iterative netlist hacking** fails because context windows drown in partial edits — a lesson that transfers to **[loop engineering for coding agents](https://explainx.ai/blog/loop-engineering-coding-agents-claude-code-guide-2026)** when you let agents rewrite half a monorepo without tests.

## Circuit highlights the agent chose

Gerhard's review (without changing nets) calls out:

**DCO integrator** — **TL072** integrator, **10 nF X7R 50 V** cap, **LM324** V-I stage, PWM pitch control — chosen for leakage and linearity within free parts.

**Transistor ladder filter** — **24 dB/oct** **Moog-style** ladder built from **matched NPN pairs**, tail current exponential converter, **instrumentation amp** output pickoff — Gerhard links classic Moog ladder explainers for readers who want theory beyond the agent report.

**VCA** — differential pair with **74HC595 + R-2R trim DAC** because PWM outputs were exhausted after filter and resonance control.

**Touch keys** — **CD4051** muxes and **74HC595** chains because **CH32V003** pins ran out — classic **agent pin-budget** story.

## Layout: where humans still steer agents

Gerhard's largest intervention was **PCB floorplanning**:

- Initial agent placement **bunched analog blocks** ; routing failed on**two layers** .
- Switching to **four layers** plus**functional-block floorplan** unlocked**Freerouting** .
- Gerhard moved blocks to open **routing channels** between**analog sensitive** sections and**digital touch** strips.

That matches how **[Claude bughunter skills](https://explainx.ai/blog/claude-bughunter-red-team-skills-bug-bounty-github-install-2026)** still need **human threat models** — agents propose, humans enforce **separation of concerns**.

## Bring-up: agents closing the loop on real hardware

After JLCPCB assembly (~two weeks), Gerhard:

- Connected a **lab PC over SSH** to Claude Code.
- Fed **audio loopback** into the MCU**ADC** .
- Read **Rigol MSO5074** traces over Ethernet while the agent stepped calibration scripts.

SPICE benches **S1–S4** matched **Board1** captures closely — Gerhard notes surprise because **second-source transistor models** often diverge.

Only fixes: **manual MCU solder bridge removal** and **touch recalibration** (keys too hot).

Demo tracks on the project site are **sheet-music-driven**, multitracked by recording **one voice at a time** — honest about **monophonic** hardware limits.

Touchsynth mixed-signal PCB design with Claude Code spans simulation validation and automated oscilloscope bring-up

## What developers are saying on Hacker News

The **[Hacker News thread](https://news.ycombinator.com/item?id=50512933&ref=explainx)** was thin at fetch time (single early comment), so there is no substantive **What developers are saying** section yet. Worth revisiting if hands-on builders post distortion measurements or BOM cost breakdowns.

## Limits and safety notes

- **Analog design agents** can still**over-componentize** when the parts library is narrow — Gerhard notes**excess Moog-era circuitry** because**no OTA/JFET** sat in the free list.
- **Touch keys on bare PCB** are**hard to play** ; Gerhard drives via**MCU scripts** instead of performance MIDI.
- **Export and ITAR** are not in scope — this is**hobby/education** scale, not defense electronics.
- **Do not** point agents at**export-controlled EDA flows** without legal review; Touchsynth stays**open BOM + JLC** .

## Toolchain specifics worth copying

Gerhard's post lists concrete versions — rare in hype threads:

- **KiCad 10** on Windows with**WSL** for Linux CLI tools.
- **ngspice** test benches named**S1–S7** with downloadable reports.
- **Freerouting** attempted on**two layers** , abandoned until**four-layer** stack-up.
- **CH32V003** firmware built with the**RISC-V** toolchain Gerhard already uses on other projects.

The **19-hour** log table is a realistic **agent calendar cost**: most hours were **unattended autorouters**, not **token streaming**. Budget similar **wall time** if you replicate — and compare that to **human-only** ladder-filter weeks from analog forums.

Audio demos came from the **behavioral Python model** before boards arrived — exactly the **digital twin** pattern **[Sophon realtime voice](https://explainx.ai/blog/sophon-realtime-v1-open-full-duplex-voice-vllm-2026)** pursues in software, but here grounding is **SPICE**, not **mel spectrograms**.

## Safety, export, and lab discipline

Touchsynth is **bench-scale**, but teams in **defense-adjacent** EDA should note:

- Automated **SSH scope control** is powerful — treat lab PCs like**CI runners** with**network segmentation** .
- **Agent-written firmware** still needs**manual power-on** smoke tests before**loudspeakers** face humans.
- **JLCPCB** assembly is**not ITAR-aware** ; this BOM is**consumer hobby** , not**munitions** .

For software-only readers, the moral is identical to **[Claude desktop Linux telemetry fixes](https://explainx.ai/blog/claude-desktop-linux-freeze-telemetry-fd-leak-deafbeef-2026)**: agents touch **real side effects** once they leave the chat box.

## How to try the pattern yourself

1. **Mirror Gerhard repo structure** — separate`spec/` ,`sim/` ,`pcb/` ,`fw/` folders so Claude Code can resume deterministically.
2. **Force simulation gates** before layout — refuse KiCad export until**ngspice reports** attach to the commit.
3. **Give the agent instrument APIs** — SSH to scope and programmer like**[connect fitness MCP patterns](https://explainx.ai/blog/connect-myfitnesspal-apple-health-strava-to-claude-chatgpt-mcp-2026)** but for**lab gear** .
4. **Budget human hours for floorplan** — agents still miss**return current paths** on two-layer budgets.

Compare to **[NVIDIA SkillEvaluator for agent skills](https://explainx.ai/blog/nvidia-skillevaluator-agent-skills-three-tier-cli-guide-2026)** — different domain, same meta-lesson: **evaluate end-to-end outcomes**, not single-file diffs.

## BOM economics and reproducibility

Gerhard constrained parts to **JLCPCB's free list** — a **contrived** but **repeatable** teaching lab. Readers replicating should expect:

- **BOM cost** in**tens of dollars** at hobby scale,**plus****MCU programmer** and**scope** .
- **Four-layer** boards beat**two-layer** for mixed signal, but**raise****Fab** minimums — the agent only succeeded after layer bump.
- **Sound examples** on the project site were**pre-silicon** from**behavioral models** — audition those before ordering boards to hear**filter sweeps** .

Compare total calendar time (**~19 h agent** + **two-week fab**) to hiring a **consulting EE** for ladder filter tuning; the post's value is **documentation density**, not guaranteed **cheaper** hardware.

## Failure modes Gerhard does not glamorize

The blog is candid about **touch keys** proving **unplayable** at human scale and about **excessive analog circuitry** because **JLC** lacked **OTAs**. Those are the wrinkles marketing threads skip:

- **Agents over-spec** passives when**pin mux** runs out — watch**BOM creep** .
- **Autorouters** burn hours without**human floorplan** hints — budget**calendar** , not just**tokens** .
- **SSH scope control** can**damage gear** if scripts**clip supply rails** — keep**manual estops** .

Treat Touchsynth as **R&D lab notes**, not a **product roadmap** for **music hardware startups**.

If you file **Issues** against **Claude Code hardware skills**, cite **Gerhard's V-model** commit boundaries — future maintainers will thank you when **KiCad netlists** stop drifting from **SPICE goldens**.

**Handoff tip:** archive **ngspice log hashes** beside **Gerber ZIPs** in release tags so the next agent session cannot claim **DRC clean** without reproducing **S3 filter gain sweeps** first. Hardware CI is **sparse** today; treat these artifacts as **your regression suite** until **open EDA** agents mature. Screenshot **scope FFTs** too — audio regressions hide in **harmonics**, not **Git diff** stats.

## Related on explainx.ai

*Hardware claims follow Gerhard November 14, 2026 primary post. Audio and BOM live in the linked project repository.*
