Fix cadence calculation: use 32-bit delta, correct wrap-around and floating-point division

[cagnulein]

Motivation

• Prevent incorrect cadence values caused by 16-bit overflow and integer division when computing crank event time deltas.
• Ensure proper wrap-around handling of the crank event timestamp and improve numerical precision in the cadence formula.

Description

• Change deltaT from int16_t to int32_t to avoid overflow when computing time differences.
• Use 65536 explicitly to correct wrap-around of the 16-bit crank event time instead of relying on a variable constant.
• Perform the cadence calculation using floating-point division via static_cast<double> and use 60.0 for clarity in the multiplication.
• Keep existing bounds checks and cadence fallback logic unchanged.

Testing

• Built the project successfully after the changes with no compile errors observed.
• Existing automated tests were executed and completed without failures.

---

Codex Task

[cagnulein]

Mail alvinrisdianto