Manufacturing — how people actually build related hardware today

There is no mass-market factory for programmable matter. What exists today is a ladder of starter paths: makerspaces, university labs, small-batch PCB houses, 3D printers, and oligo synthesis services.

The real ladder (idea → prototype → small batch)

Skip “scale to millions” fantasies until you can make ten reliable units.

Mechanical modules (3D print & laser-cut)

What beginners actually do:

1. Design a brick/cube/tile in FreeCAD, Fusion, or Onshape.
2. Add pockets for magnets, screws, or snap fits.
3. Print in PLA/PETG on a school/makerspace FDM printer.
4. Or laser-cut acrylic/plywood tiles for 2D capillary/magnet demos.

Tips that matter for self-assembly toys:

• Tolerance magnets carefully (press-fit vs glue).
• Chamfer edges so modules can slide into contact.
• Design polarity (N/S faces) into the CAD, not as an afterthought.
• Print a 1-unit test before a 20-unit batch.

Open starting points: modular robot papers’ supplementary CAD when released; classroom floatile repos such as FloaTiles ; your own Exp 7 modules from Experiments.

Connectors people use in labs

Electronics path (Arduino / ESP32)

Most student swarm prototypes are “a brain + radios/IR + motors + battery” repeated N times.

Typical student stack:

• MCU: Arduino Uno/Nano, ESP32, or STM32 Nucleo
• Motion: vibration motors, small DC gearmotors, or servos
• Sensing: IR reflectance, ToF, bump switches, IMU
• Comms: UART daisy-chain, ESP-NOW, BLE, or IR like Kilobots
• Power: LiPo with proper charging protection (makerspace supervision)

Firmware pattern: sense neighbors → apply local rule → actuate. Keep the rule short enough to explain on one sticky note.

PCB fabrication (small batch, not a fab)

You do not etch boards at home for serious revision control. Students usually:

1. Draw schematic + PCB in KiCad (free).
2. Export Gerbers.
3. Order from a prototype house (examples people actually use: JLCPCB, PCBWay, OSH Park).
4. Optionally order PCBA (assembly) for a few boards when the BOM is simple.

Watch a real factory tour to calibrate expectations (this is how boards are made, not a programmable-matter plant):

Source: Strange Parts — Inside a Huge PCB Factory  (JLCPCB tour context).

Cost realism (order-of-magnitude, 2020s hobbyist): bare 2-layer boards often a few dollars each at quantity 5–10; assembled boards cost more once parts + stencil + labor are included.

Kilobot & open modular lineages

• Kilobots — designed for automated assembly and low cost per unit for research swarms. Academic/commercial kit availability changes over time; check Harvard SSR Kilobots page  and educational vendors. Many learners start with simulators first.
• M-Blocks / SMORES / other MRR — research hardware; learn from papers and lab pages (SMORES-EP , MIT M-Blocks videos) rather than expecting Amazon kits.
• DIY vibration robots — feasible makerspace project inspired by Kilobot locomotion (not a clone of the research platform).

Molecular work: DNA as a service

You generally do not synthesize DNA oligos yourself as a beginner.

1. Design sequences (paper, cadnano, or lab software).
2. Order oligos / gene fragments from vendors (commonly cited educational/research vendors include IDT, Twist, etc.).
3. Fold / anneal in a supervised wet lab with buffers, thermal cycler, and gel/AFM access.
4. Image via university core facilities.

Home “DNA origami kits” that promise kitchen nanotech should be treated with extreme skepticism. See Supply chain for educational reagent boundaries.

Safety & cost realism

Try this (manufacturing mini-plan)

Write a one-page plan for a 5-module magnetic tile system:

1. Sketch faces and polarity.
2. List print settings + magnet size.
3. Estimate cost from Supply chain vendors.
4. Define a pass/fail test (“forms a chain of 5 within 60 s of agitation”).

Next

→ Supply chain for BOM thinking · or return to Beginners