Why reflow is still a black art — and how to end it
The physics of joint formation, why hand-built profiles drift, and what closed-loop control changes.
Read
Learn the loop
Guides, field notes, and playbooks
Deep dives on print, reflow, inspection, twins, orchestration, and physical-AI robotics for electronics manufacturing.
Featured
Why reflow is still a black art — and how to end it
The physics of joint formation, why hand-built profiles drift, and what closed-loop control changes.
Read the library
Library
Latest writing
Cutting the false-call tax without missing escapes
How learned AOI/AXI models beat static thresholds on both false calls and real defects.
ReadNPI without burning boards
Qualifying paste, placement, and reflow recipes in a physics twin before the first article.
ReadGraduated autonomy on a high-reliability line
Shadow → recommend → approve → closed-loop, and how each level earns the next.
ReadThe closed-loop corpus as a manufacturing moat
Why writing every action to a shared corpus compounds into defensibility.
ReadPhysical AI on the factory edge
Deploying 20–50 ms inference on Jetson and IGX across a live SMT line.
Read
By topic
Find what you need
Print & place
SPI, paste volume, area ratio, and placement offset control.
Reflow
Thermal profiles, joint formation, and selective soldering.
Inspection
AOI, AXI, BGA voids, coating, and false-call reduction.
Digital twin
Physics simulation, recipe optimization, and calibration.
Orchestration
Yield, line balancing, and test triage.
Robotics
Handling, depaneling, and box-build with Isaac Sim.
Formats
However you learn best
Guides
Step-by-step playbooks for each stage of the loop.
Webinars
Live and on-demand sessions with our engineers.
Case studies
Real outcomes from lines running Circuex.
Benchmarks
How to measure first-pass yield and false-call rate honestly.
0Topic areas
0Edge inference / tile
0Loop stages
0First-pass yield
The playbooks alone reframed how our team thinks about yield. Then the platform made the reframing real.
Newsletter
The loop, in your inbox
Monthly field notes on physical AI for electronics manufacturing. No spam.
Is the content gated?
Most field notes are open. A few deep benchmarks and case studies are shared under NDA.
Can you present to our team?
Yes — book a session and we'll tailor a briefing to your line.
Every board, perfectly built.
Put the ideas to work
Book a demo and see the concepts running on real board data.