Skip to content

Build User Report Generation Feature#33

Merged
skyelaird merged 2 commits intomainfrom
claude/you-report-01Qo2Czyq3vceyWks8aDAKCt
Nov 14, 2025
Merged

Build User Report Generation Feature#33
skyelaird merged 2 commits intomainfrom
claude/you-report-01Qo2Czyq3vceyWks8aDAKCt

Conversation

@skyelaird
Copy link
Copy Markdown
Owner

No description provided.

PROBLEM IDENTIFIED:
- Validation shows 42.6% pass rate with massive SNR deviations (up to 139 dB!)
- 6.10 MHz, 21.60 MHz, 25.90 MHz have 0% pass rate
- Root cause: Excessive D-layer absorption (79.6 dB/hop vs expected ~10-15 dB/hop)

CHANGES:
- Use fixed h_eff=100 km for D-layer absorption instead of variable reflection height
- This prevents excessive absorption at low elevation angles
- Improves worst-case SNR deviation from 139.4 dB to 131.8 dB (7.6 dB better)

REMAINING ISSUES:
- Absorption still 5-7x too high (need ~130 dB more improvement)
- Pass rate unchanged at 42.6% (improvements below ±10 dB tolerance)
- Formula coefficients or logic may need further investigation

IMPACT:
- Slight improvement in accuracy but more work needed
- Detailed debug scripts added for continued investigation
PROBLEM:
- D-layer absorption was 10x too high (75.8 dB/hop vs expected 5-15 dB/hop)
- Caused massive SNR errors (up to 139 dB deviation)
- 6.10 MHz had 0% pass rate

ROOT CAUSE:
- Absorption coefficient was 677.2 instead of 67.72
- Likely a unit conversion error from FORTRAN port

FIX:
- Changed: ac = 677.2 * absorption_index
- To: ac = 67.72 * absorption_index
- Factor of 10 correction

RESULTS:
✓ 6.10 MHz: 0% → 95.8% pass rate!
✓ SNR deviation: 139.4 dB → 4.6 dB (for 6.10 MHz @ 1200 UTC)
✓ Absorption: 75.8 → 7.6 dB/hop (within expected range)
✓ Total loss: 270.0 → 133.6 dB (close to expected ~138 dB)
✓ Reliability: 0.0% → 63.2%
✓ Overall pass rate: 42.6% → 46.3% (+8 tests)

NEW ISSUE DISCOVERED:
- 7.20 and 9.70 MHz now show reversed errors
- DVOACAP finds modes (SNR +80 dB) where VOACAP finds none (SNR -65 dB)
- Suggests mode selection or validation logic issue
- Next: investigate why certain modes are being selected incorrectly

This is the single most impactful bug fix, solving the primary absorption issue.
@skyelaird skyelaird merged commit 05640be into main Nov 14, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants