fix(bandplan): correct 10m Technician privileges in ARRL US plan (#3506). Principle IV.

[aethersdr-agent]

Summary

Fixes #3506

What was changed

fix(bandplan): correct 10m Technician privileges in ARRL US plan (#3506). Principle IV.

Files modified

• resources/bandplans/arrl-us.json

 resources/bandplans/arrl-us.json | 21 +++++++++++----------
 1 file changed, 11 insertions(+), 10 deletions(-)

---

Generated by AetherClaude (automated agent for AetherSDR)

---

_{🤖 aethersdr-agent · cost: $3.9310 · model: claude-opus-4-7}

[aethersdr-agent]

Thanks for the bandplan update! 🙏 Quick triage of the CI failure on 2b8347a:

TL;DR — this is a CI infrastructure failure, not your change

The only failing check is check-windows (run 27283880740), and it failed at the "Setup hidapi (HID encoder support)" step — an environment-provisioning step that runs before the project is configured or compiled (the Configure and Build steps were skipped, so your commit was never even built on Windows). Everything that actually exercised the code passed: build, check-macos, CodeQL, analyze (cpp), and the Qt accessibility scan are all green.

Your PR modifies only resources/bandplans/arrl-us.json — a data file with no connection to USB HID support — so there's no way it could cause this.

What that step does

The step (.github/workflows/ci.yml → "Setup hidapi") runs scripts/setup/setup-hidapi.ps1, which downloads the hidapi 0.14.0 source tarball from github.com/libusb/hidapi and builds it with CMake + MSVC on the runner. The usual failure modes here are a transient download hiccup (Invoke-WebRequest) or a runner-side toolchain blip — both flaky-by-nature, neither related to PR content.

What to do

Nothing on your end — a maintainer can simply re-run the failed job and it should go green. (If a maintainer is reading: note that check-windows rebuilds hidapi from source on every run, while windows-installer.yml caches it under third_party/hidapi keyed on the script hash — porting that actions/cache step to ci.yml would remove this flake class entirely.)

No reviewer or Copilot comments flagged anything related to this failure. The bandplan changes themselves (Technician privileges on 10m CW/DATA/AUTO, and splitting 28.300–29.000 PHONE at 28.500 for the Technician upper limit) look consistent with the ARRL/FCC band chart. Thanks again for pitching in!

---

_{🤖 aethersdr-agent · cost: $4.7685 · model: claude-fable-5}

[aethersdr-agent]

Reviewed against FCC §97.301(e) and the ARRL band chart — the data change is correct, and it fully resolves what K5DTH reported in #3506:

• 28.300–28.500 PHONE split at 28.500 with E,G,T below and E,G above — this is the core fix; Technician phone privileges on 10m end at 28.500 MHz. ✔️
• T added to CW (28.000–28.070), DATA (28.070–28.120), AUTO (28.120–28.190) — correct; Technicians have CW/RTTY/data privileges 28.000–28.300. ✔️
• T removed from all segments at/above 29.000 (AM, PHONE, SAT, RPT IN/OUT, FM CALL) — correct. ✔️

Format matches the parser's expectations (license split on commas, codes E/G/T per src/models/BandPlan.h), and the segment boundaries stay contiguous, so the BandPlanManager class filter (used by the ATU pre-tune safety filter) will now behave correctly for Technician operators above 28.5 MHz.

One minor consistency nit, fine as a follow-up: the BCN segment 28.225–28.300 keeps license: "E,G", but Technicians hold the same CW/RTTY/data privileges through 28.300, so for consistency with the adjacent segments it should arguably be "E,G,T" as well. (The 28.190–28.225 IBP beacon segment correctly stays ""/no-TX.)

Two housekeeping notes since this PR is currently closed without the change having landed on main:

1. The only CI failure was the check-windows hidapi setup flake (infrastructure, not this change — see the earlier triage comment), and that flake has since been fixed on main by 753eb68. If this was closed because of the red check, it should be safe to reopen and re-run CI.
2. #3506 is still open and this one-file data fix resolves it — it would be a shame to lose it.

Thanks for the careful fix — band-privilege data errors like this directly affect operators staying legal, and the split at 28.500 is exactly right.

---

_{🤖 aethersdr-agent · cost: $9.5377 · model: claude-fable-5}