rf_baseline_observer Ham Band Interference Monitoring System (RTL-SDR Blog V4 + Python)

This project establishes a high-resolution RF noise floor baseline on Windows 11 to quantify electromagnetic interference (EMI) from a nearby data center across all US amateur radio bands — HF through microwave.

Project Architecture

• Sensor: RTL-SDR Blog V4 (R828D2 chip, 500 kHz – 1.766 GHz; HF via direct sampling)
• Host: Windows 11
• Data Logger: rtl_power — dual sweep per cycle: HF (1.8M–28M, direct sampling) + VHF/UHF (28M–1.766G)
• Processor: check_floor.py — per-band noise floor analysis (Python/Pandas)
• Scheduler: run_continuous.ps1 — headless continuous logging loop (PowerShell)
• TUI: rf_observer_tui.py — interactive terminal UI with live band table (replaces run_continuous.ps1)
• TX Toggle: toggle_rf.ps1 — kill/resume logging when going on air
• Visualization: SDR++ — realtime spectrum and waterfall display

Monitored Ham Bands

Band	Frequency Range	Sweep
160m	1.800 – 2.000 MHz	HF (direct)
80m	3.500 – 4.000 MHz	HF (direct)
60m	5.330 – 5.404 MHz	HF (direct)
40m	7.000 – 7.300 MHz	HF (direct)
30m	10.100 – 10.150 MHz	HF (direct)
20m	14.000 – 14.350 MHz	HF (direct)
17m	18.068 – 18.168 MHz	HF (direct)
15m	21.000 – 21.450 MHz	HF (direct)
12m	24.890 – 24.990 MHz	HF (direct)
10m	28.000 – 29.700 MHz	VHF/UHF
6m	50.000 – 54.000 MHz	VHF/UHF
2m	144.000 – 148.000 MHz	VHF/UHF
1.25m	222.000 – 225.000 MHz	VHF/UHF
70cm	420.000 – 450.000 MHz	VHF/UHF
33cm	902.000 – 928.000 MHz	VHF/UHF
23cm	1240.000 – 1300.000 MHz	VHF/UHF

Folder Structure

C:\
├── Projects\
│   └── rf_baseline_observer\          ← this repo
│       ├── check_floor.py             ← per-band noise floor analyzer
│       ├── rf_observer_tui.py         ← terminal UI for live monitoring
│       ├── run_continuous.ps1         ← main continuous logging loop (use this)
│       ├── start_logging.ps1          ← single-run background logger + Task Scheduler setup
│       ├── toggle_rf.ps1              ← pause/resume logging when transmitting
│       ├── setup.ps1                  ← one-time dependency installer
│       ├── LICENSE
│       └── README.md
│
├── Tools\
│   └── rtl-sdr-blog-v4\              ← RTL-SDR Blog V4 Windows driver package (extracted here)
│       ├── rtl_power.exe             ← sweep tool used by run_continuous.ps1
│       ├── rtl_test.exe              ← device verification tool
│       ├── rtl_fm.exe
│       ├── rtl_sdr.exe
│       └── *.dll                     ← required runtime libraries (must stay with the .exe files)
│
└── Users\
    └── <your username>\
        └── rf_logs\                  ← runtime output (auto-created on first run)
            ├── hf_<timestamp>.csv    ← HF sweep data (1.8M–28M, direct sampling)
            ├── vhf_<timestamp>.csv   ← VHF/UHF sweep data (28M–1.766G)
            └── rf_history.log        ← JSON Lines history (one record per cycle)

PATH requirement: Add C:\Tools\rtl-sdr-blog-v4 to your system PATH so rtl_power.exe is accessible from any terminal. The .dll files in that folder must stay alongside the executables — do not move the .exe files out on their own.

Installation & Setup

1. Install Python

Download and install Python 3.x from https://www.python.org/downloads/

Check "Add Python to PATH" during installation.

2. Install RTL-SDR Blog V4 Drivers

The V4 requires the RTL-SDR Blog driver package — not the generic osmocom build.

1. Download the RTL-SDR Blog Windows driver package from https://www.rtl-sdr.com/rtl-sdr-blog-v4-dongle-initial-release/
2. Extract and add the folder containing rtl_power.exe to your system PATH:
   • Search "Environment Variables" in the Start menu
   • Edit the Path variable under System Variables
   • Add the path to the extracted folder (e.g., C:\Tools\rtl-sdr-blog-v4)

3. Install the WinUSB Driver (Zadig)

1. Plug in your RTL-SDR Blog V4 dongle
2. Download Zadig from https://zadig.akeo.ie/
3. In Zadig: Options > List All Devices
4. Select Bulk-In, Interface (Interface 0) — the V4 shows as RTL2832UHIDIR
5. Choose WinUSB and click Install Driver

V4 note: Do not select the wrong interface. If you see two entries for the device, pick Interface (Interface 0).

4. Run the Setup Script

Open PowerShell as Administrator and run:

Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope CurrentUser
.\setup.ps1

This installs the required Python packages (pandas, numpy) and verifies your rtl_power installation. If you plan to use the TUI (rf_observer_tui.py), also run pip install textual afterward.

Operation Workflow

Phase 1: Continuous Logging (Recommended)

Run the continuous logger for long-term baseline collection. Each cycle runs two back-to-back 30-minute sweeps — HF bands using direct sampling mode (-D 2 required by the V4 for frequencies below 28 MHz), then VHF/UHF with the normal tuner. After both sweeps, it auto-fetches the Kp index from NOAA, analyzes results across all 16 bands, and appends a JSON record to the history log.

.\run_continuous.ps1

Optional parameters:

.\run_continuous.ps1 -OutputDir "D:\rf_data" -FreqHF "1.8M:28M:100K" -FreqVHF "28M:1.766G:100K"

Press Ctrl+C to stop. Each completed cycle is already saved and logged before the next begins.

Pausing Logging While Transmitting

The RTL-SDR and your transceiver share the USB bus — you cannot transmit while logging. Stop the logger before keying up:

.\toggle_rf.ps1

Run toggle_rf.ps1 again to resume, or simply restart run_continuous.ps1. Each resume starts fresh CSVs with new timestamps.

Phase 2: Manual Analysis

Analyze any collected CSVs at any time. Pass one or both sweep files from a cycle. Use --json for machine-readable output.

Space Weather Flagging

Geomagnetic activity independently raises the noise floor on HF/VHF bands. Always include the Kp index to distinguish solar interference from local EMI.

# Human-readable, auto-fetch Kp from NOAA (requires internet)
python check_floor.py "$env:USERPROFILE\rf_logs\hf_<timestamp>.csv" "$env:USERPROFILE\rf_logs\vhf_<timestamp>.csv" --fetch-kp

# Manually specify Kp
python check_floor.py "$env:USERPROFILE\rf_logs\hf_<timestamp>.csv" "$env:USERPROFILE\rf_logs\vhf_<timestamp>.csv" --kp 1.7

# JSON output with Kp
python check_floor.py "$env:USERPROFILE\rf_logs\hf_<timestamp>.csv" "$env:USERPROFILE\rf_logs\vhf_<timestamp>.csv" --json --fetch-kp

# Single-sweep file (HF only or VHF/UHF only)
python check_floor.py "$env:USERPROFILE\rf_logs\vhf_<timestamp>.csv" --fetch-kp

Human-readable example:

--- RF NOISE REPORT: hf_20260101_120000.csv, vhf_20260101_123000.csv ---
Space Weather:  Kp=1.7  [UNSETTLED]
Band     Avg (dBm)  Peak (dBm)  Samples  Status
------------------------------------------------------------
160m        -88.40      -74.10      512  NOMINAL
80m         -85.22      -70.60      820  NOMINAL
60m         -84.10      -69.80      102  NOMINAL
40m         -83.55      -68.30      492  NOMINAL
30m         -82.90      -67.50       82  NOMINAL
20m         -81.20      -66.40      574  NOMINAL
17m         -80.75      -65.90      164  NOMINAL
15m         -79.60      -64.20      738  NOMINAL
12m         -78.30      -63.10      164  NOMINAL
10m         -83.12      -71.40     1024  NOMINAL
6m          -81.55      -68.90      412  NOMINAL
2m          -76.20      -62.10      492  WARNING
1.25m       --          --            0  no data
70cm        -79.45      -60.80     1860  NOMINAL
33cm        -65.10      -55.20     1638  WARNING
23cm        -82.77      -70.10     3720  NOMINAL

JSON example:

{
  "timestamp": "2026-01-01T12:00:00Z",
  "source_file": ["hf_20260101_120000.csv", "vhf_20260101_123000.csv"],
  "space_weather": {"kp_index": 1.7, "condition": "UNSETTLED"},
  "bands": {
    "160m": {"avg_dbm": -88.40, "peak_dbm": -74.10, "samples": 512, "status": "NOMINAL"},
    "2m":   {"avg_dbm": -76.20, "peak_dbm": -62.10, "samples": 492, "status": "WARNING"}
  }
}

Kp Condition Scale

Kp	Condition	Effect on HF/VHF
0–1	QUIET	Clean baseline
2–3	UNSETTLED	Minor ionospheric variation
4	ACTIVE	Noticeable HF impact
5	G1-MINOR	HF degradation possible
6	G2-MODERATE	HF unreliable
7	G3-STRONG	HF blackouts possible
8	G4-SEVERE	Wide HF blackout
9	G5-EXTREME	Total HF blackout

Baseline tip: Use only Kp ≤ 1 (QUIET) runs as your reference baseline when comparing pre- vs. post-construction readings.

Phase 3: Review History Log

Each cycle appends one JSON record to %USERPROFILE%\rf_logs\rf_history.log (JSON Lines format — one complete object per line). Tail it live:

Get-Content "$env:USERPROFILE\rf_logs\rf_history.log" -Wait

Terminal UI (rf_observer_tui.py)

rf_observer_tui.py is a fully interactive terminal application that replaces run_continuous.ps1 with a live display. It runs the same dual-sweep cycle (HF direct sampling + VHF/UHF) but shows results in real time as each cycle completes.

Additional Dependency

The TUI requires the textual package, which is not installed by setup.ps1:

pip install textual

Running the TUI

python rf_observer_tui.py

# Custom output directory or second dongle
python rf_observer_tui.py --output-dir D:\rf_data --device 1

Layout

┌─ RF Baseline Observer — RTL-SDR V4 ──────────────────────────────────────┐
│ ┌─ HAM BAND RF NOISE FLOOR ──────────────┐ ┌─ SWEEP STATUS ─────────────┐ │
│ │ Band   Avg(dBm) Peak(dBm) Samples Status│ │ HF Sweep (elapsed: 12:34)  │ │
│ │ 160m   -88.40   -74.10      512 NOMINAL │ │ Cycle: 3                   │ │
│ │ 80m    -85.22   -70.60      820 NOMINAL │ │                            │ │
│ │ ...                                     │ │ SPACE WEATHER              │ │
│ │ 2m     -76.20   -62.10      492 WARNING │ │ Kp: 1.7  [UNSETTLED]       │ │
│ │ 33cm   -65.10   -55.20     1638 WARNING │ │                            │ │
│ └─────────────────────────────────────────┘ │ CONTROLS                   │ │
│                                             │ [ Start ]                  │ │
│                                             │ [ Stop  ]                  │ │
│                                             └────────────────────────────┘ │
│ CYCLE HISTORY                                                               │
│ [12:04:11] Cycle 3 | Kp=1.7 [UNSETTLED] | NOMINAL:14 WARN:2 CRIT:0       │
│ [11:34:08] Cycle 2 | Kp=1.5 [QUIET]     | NOMINAL:16 WARN:0 CRIT:0       │
└─ q Quit  s Start/Stop  c Copy History ───────────────────────────────────┘

Keyboard Shortcuts

Key	Action
s	Start / Stop the sweep loop
q	Quit
c	Copy cycle history to Windows clipboard

Notes

• Writes to the same rf_history.log as run_continuous.ps1 — both tools share the same log file and output directory.
• Status colors: green = NOMINAL (avg ≤ −75 dBm), yellow = WARNING (> −75 dBm), red = CRITICAL (> −60 dBm).
• The sweep phase timer resets at the start of each HF and VHF/UHF sweep.
• Press Stop (or s) to let the current sweep finish cleanly before the loop exits. Pressing Stop mid-sweep sends a terminate signal to rtl_power immediately.

Realtime Visualization

For a live spectrum and waterfall display, use SDR++:

1. Download from https://github.com/AlexandreRouma/SDRPlusPlus/releases
2. Extract and run sdrpp.exe — no install required
3. In SDR++: set Source to RTL-SDR, click Play

Note: SDR++ and rtl_power cannot share the dongle. Run .\toggle_rf.ps1 to stop background logging before opening SDR++, and run it again to resume when done.

Troubleshooting

• SDR not found: Run rtl_test.exe -t in a terminal. If it fails, re-run Zadig and reinstall the WinUSB driver on Interface 0.
• rtl_power not recognized: Ensure the RTL-SDR Blog V4 bin folder is in your system PATH and restart the terminal.
• HF bands show no data: Direct sampling requires the RTL-SDR Blog V4 driver build. Generic osmocom builds do not support -D 2 correctly. Confirm with rtl_power -d 0 -D 2 -f 7M:7.3M:1K -i 10s -e 10s test.csv.
• Inaccurate readings: Place the antenna away from PCs and routers to reduce local noise.
• Task Scheduler errors: Run PowerShell as Administrator when executing start_logging.ps1.