Fuss-Free Browser Apps

A keyboard matrix arranges keys in intersecting rows and columns. Instead of dedicating one microcontroller pin per key, the controller drives rows and scans columns (or vice versa). This reduces pin usage from keys to roughly rows + columns, at the cost of slightly more complex scanning firmware.

For each possible row count, the tool calculates the minimum number of columns needed to fit all keys. It keeps combinations where rows × columns covers the key count and where rows + columns is within your pin budget. Plans are then sorted by:

1. Lowest total pins (rows + columns)
2. Lowest wasted positions (empty matrix slots)
3. Smallest overall matrix

The top entries are usually the neatest designs to route and scan.

Because matrix sizes are whole rows and columns, rows × columns is often slightly larger than your exact key count. Any extra cells are “wasted positions” — they exist in the matrix but have no physical switch connected. A small amount of waste is normal, especially for odd key counts.

In a plain matrix without per-switch diodes, certain three-key combinations can cause “ghosting” or “masking”, where phantom key presses appear or real presses are missed. Adding a diode in series with each switch allows current to flow only one way, which largely prevents ghosting and enables near-NKRO behaviour. The trade-off is extra parts and soldering.

If your pin budget is very tight, there may be no rows + columns combination that can cover all keys. The planner will then show the minimum pin count that would work, along with example matrices. At that point you can either reduce the key count, free up more pins, or consider I/O expanders.