[cpp-RCP-09] ADC endpoint uses an invented "ADC_NO_SIGNAL" error and 32-bit values instead of the spec's 16-bit width #77

Location: include/rcp/adc.hpp — AdcErrc::no_signal message text (~line 91) and the uint32_t-typed sample/value plumbing throughout AdcEndpoint/compute_average (~lines 113-190).

Finding: Two separate divergences from the governing specification in the same header. First, the ADC endpoint's no-signal condition renders as "ADC_NO_SIGNAL" in its error message; the spec's numeric error-code table has no such code — the only *_NO_SIGNAL entry in that table belongs to the PWM_IN endpoint, not ADC. Second, ADC sample/result values are modeled as uint32_t throughout this header, while the spec's ADC endpoint is defined for resolutions up to 16 bits, with responses carrying 16-bit values.

Recommendation: Remove the invented "ADC_NO_SIGNAL" identifier from the error message (use one of the spec's real error codes, or none, rather than a fabricated one), and narrow the ADC sample/result representation to 16 bits to match the spec's resolution ceiling.

Verification: Independently re-read include/rcp/adc.hpp against current HEAD (commit 4b9449a, 2026-07-29), and cross-checked against the governing specification's error-code table (confirming PWM_IN_NO_SIGNAL is the only *_NO_SIGNAL entry) and its ADC endpoint section (confirming the 16-bit value width and response figure). Confirmed both divergences.

Filed from the 2026-07-29 ecosystem audit register; independently re-verified against current HEAD before filing.