On solder, and the people who trust it

Nobody has ever bought a synthesizer module because of its solder. This note is not going to change that, and it is not trying to. It exists because our joint inspection data, three years of it, tells a story that deserves a page before it becomes a habit nobody remembers the reason for.

Every board in every batch since 2023 has passed through manual joint inspection under magnification, and every finding is logged against the batch, the operator, the profile, and the day. That is 24 batches and, as of October, 41,300 modules. The data says something plain: the profiles and materials we chose in 2023, chosen partly because they were boring, have produced a joint failure rate in warranty service of 0.09%, and 60% of those failures trace to a single cause we have since engineered out.

What the data actually shows

Across the fleet, jack connector joints are the dominant failure site, not because our solder is weak but because a jack on the face of a panel lives a harder life than any component on the board. It takes the lever load of every cable insert, ten thousand times by design. The fix that took the failure rate from 0.19% to 0.09% was not a better solder alloy. It was a reinforcing loop in the jack footprint, added in batch 2025-03, that spreads the mechanical load across more copper before any of it reaches the joint.

The second finding is about people, not metal. Joint quality variance between operators on the same profile, same day, same alloy, is smaller than variance across days for the same operator. Fatigue, room temperature, and the hour of the shift show up in the data more than individual technique does. We responded with the least glamorous intervention available: inspection moves to the first two hours of the shift, when the data says attention is at its widest, and the QC pass that follows assembly checks joints again, at whatever hour the batch finishes.

Why we are telling you this

Because the community deserves the version of this note that a marketing department would never write: our solder is not special. It is a mainstream no-clean alloy from a profile we tuned once in 2023 and have touched twice since, both times with a changelog entry. What is special is that we log every joint finding, publish the fleet number, and redesigned a footprint when the data pointed at it. The alloy is replaceable. The habit is not.

There is also a practical point for anyone repairing their own module, which our license explicitly supports: your joint does not have to look like ours. It has to pass the same test ours does, continuity under load after thermal cycling, and the hardware documentation includes the profile parameters we use so your iron can be set to something close. The certificate in your box says who built your module and when. If you rework it, you are allowed to sign it too, and some of our favorite community patches have come out of racks that carry two signatures.