Ellaine's nameplate

A desk nameplate for Ellaine Chou of the Astera Institute, built two ways. Each revision is a small experiment with its own question, method, result and files, published the day it was made, including the parts that did not work the first time.

This is Rev 1, loaded when it scrolls into view. The guided walkthrough starts on its own; the PCB, Schematic and Parts tabs are live. Open it full screen.

Two ways to put a name on a board. Rev 1 builds the letters out of 108 LEDs, a new circuit per name. Rev 2 keeps one standard LED-lit board for everyone and has the workcell's laser etch the name into it. Rev 2 is the plate Ellaine will hold. Both grew out of Steven's nameplate, whose first revision is where the 5 x 7 font lesson came from.

The two revisions

Rev 1: 5 x 7 LED dot matrixdesigned

108 white LEDs spell ELLAINE, every part a stocked Adom Basic Part with an MPN, about 460 mA from a phone charger. The circuit carried over from Steven's board as one line of code; the routed layout did not, and the sweep that found a clean one is tabled in full. 0 DRC errors, 112 of 112 nets connected.

Full write-up Plain text Live 3D

Rev 2: name laser-ablated into the silkscreenmade

The plate Ellaine receives. Adonis Lysandrou's standard blank board (Fusion 360, six side-firing LEDs), with "Ellaine Chou" on the first line and the Astera logo on the second, ablated into the soldermask by the workcell's UV laser from a Nameplate Generator job. Per-person work drops from a new PCB to a laser job.

Full write-up Plain text Live 3D of the plate Board designer's page

What changed, and why

	Rev 1	Rev 2
How the name is made	108 LEDs, 5 x 7 font	Laser-ablated silkscreen, lit by 6 edge LEDs
Board	154.4 x 44.4 mm	160 x 48 mm
Designed with	tscircuit 0.0.2375, AI	Fusion 360 Electronics, human
Parts	all stocked, with MPNs	OSRAM, GCT, Royalohm, with MPNs
New work per name	a new PCB (one line of code, plus a routing sweep)	a new laser job
Status	verified, not fabricated	made, handed over in DC
Lesson	DRC passes boards with dead nets; check connectivity geometrically	one board for everyone; the per-person work is a laser job

Open science

This page is a worked example, for the Astera Institute, of how Adom's AI publishes an experiment as soon as it exists. The full scorecard lists every rule and where this page meets it, meets it partly, or does not yet.

Astera asks for	Here
Publish when ready for reuse, not when a paper is written	both revisions published the day they were made
Code MIT or more open	yes, MIT
Data CC0	yes
Protocols and text CC BY	yes, CC BY 4.0; each revision has a reproduce section
Funder with ROR ID	U.S. NSF, ROR 021nxhr62
Negative results	every failed router layout, with its error and split-net counts
Register of outputs	outputs.csv with SHA-256 per file
DOI and Zenodo deposit	not yet, the gap our proposed pipeline closes
Code on GitHub	partial, git on the wiki, no GitHub mirror

Files and licenses

Board source and scripts (rev*/source/): MIT
Data (circuit.json, 3D models, gerbers, BOMs): CC0 1.0
Write-ups, photos, artwork: CC BY 4.0
Every output with its SHA-256: outputs.csv. How to cite: CITATION.cff. The brief and every request: plan.md

Provenance

Requested by John Lauer (PI, Adom Industries, NSF PCL Test Bed node, Fort Worth). Rev 1 designed by Claude (Anthropic) in Adom's Hydrogen desktop with tscircuit and adom-tsci. Rev 2 board designed and ablated by Adonis Lysandrou; laser job generator by Aditya Angajala. The node is funded by the U.S. National Science Foundation (ROR 021nxhr62). Made for the Astera Institute (ROR 00ydx1s47), which did not fund it.