Skip to content

Verify/lora phy campo - #8

Open
arlethmichellopez wants to merge 2 commits into
mainfrom
verify/lora-phy-campo
Open

arlethmichellopez wants to merge 2 commits into
mainfrom
verify/lora-phy-campo

Conversation

@arlethmichellopez

Copy link
Copy Markdown
Contributor

Verify RFSAM-LORA-PHY-01 with real US915 reception and decoding sessions

Replace the [FILL: measured CRC-OK frame yield, hard vs soft-decision, at a recorded SNR for one capture] placeholder in ## Field case with actual over-the-air measurements conducted at Electronic Cats, Aguascalientes.

What changed

  • ## Field case: Added measured statistics comparing soft vs. hard-decision LoRa demodulation yields across two 30-second US915 capture sessions (FlatSat baseline at 916.0 MHz vs. STM32 verification at 915.0 MHz).
  • Frontmatter:
    reviewStatusverified,
    lastResearched → 2026-09-11.

Measured results

Parameter / Metric Session 1 (FlatSat Baseline) Session 2 (STM32 Verification) Delta / Performance Impact
Center Frequency 916.0 MHz (US915 Ch 8 / Sub-band 2) 915.0 MHz (US915)
Bandwidth (BW) / SF 125 kHz / SF7 250 kHz / SF11
Estimated SNR ~45.0 dB (Strong / Saturated) ~6.0 dB (Near Sensitivity) -39.0 dB
Total Visible Frames 12 (30 s IQ capture window) 5 (30 s IQ capture window)
Soft-Decision Yield 12 / 12 (100% CRC-OK) 3 / 5 (60% CRC-OK) -40.0% under low SNR
Hard-Decision Yield 12 / 12 (100% CRC-OK) 1 / 5 (20% CRC-OK) -80.0% under low SNR
Soft vs. Hard Advantage No difference (Saturated) +40 percentage points Soft-decision advantage

Conditions

  • Location: Electronic Cats, Aguascalientes
  • Setup Spacing: 1-meter fixed line-of-sight bench test
  • SDR Receiver: RTL-SDR (RTL2832U) with stock telescopic antenna
  • RF Capture Settings: rtl_sdr -s 1000000 -g 40.2 (30-second raw uint8 I/Q output)
  • Spectrum Analyzer: Gqrx 2.17.7
  • Flowgraph & Decoder: gr-lora_sdr (GNU Radio Companion flowgraph lora_RX_916.grc) & offline Python script (lora_decode_SF11_fix.py)
  • Transceivers Used: Laboratory FlatSat (916 MHz) & STM32F446 + DX-LR30 transceiver (915 MHz, application CRC intentionally corrupted every 2 packets)

Evidence

Session 1 — FlatSat Baseline (916.0 MHz, BW 125 kHz, SF7)

  • Gqrx spectrum waterfall showing 916.0 MHz reception:
Screenshot From 2026-09-08 11-32-10
  • GNU Radio flowgraph configured for Soft-Decision (soft_decoding: True):
Screenshot From 2026-09-10 10-54-43
  • GNU Radio flowgraph configured for Hard-Decision (soft_decoding: False):
Screenshot From 2026-09-10 10-54-05
  • Demodulation terminal execution (12/12 CRC-OK, SNR ~45 dB):
Screenshot From 2026-09-10 10-54-37

Session 2 — STM32 Verification (915.0 MHz, BW 250 kHz, SF11)

  • Gqrx spectrum waterfall showing 915.0 MHz reception:
Screenshot From 2026-09-09 12-31-17
  • Terminal output log for Soft-Decision execution:
Screenshot From 2026-09-10 11-41-13
  • Terminal output log for Hard-Decision execution:
Screenshot From 2026-09-10 11-41-33
  • Metric summary and frame yield validation (3/5 soft vs 1/5 hard):
Screenshot From 2026-09-10 11-41-22
  • Detailed payload logs verifying controlled CRC corruption:
Screenshot From 2026-09-10 11-41-39

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.

1 participant