Перейти к содержанию

RF safety guide for SDR experiments

This page defines the minimum safety discipline for all RF-facing labs in the course. It is written for educational SDR experiments with Zynq/AD9363-class boards, RTL-SDR receivers, attenuators, coax cables and controlled low-power signals.

Warning

Do not connect an SDR transmitter directly to a sensitive receiver input unless the expected power level, attenuation and maximum input rating have been checked. When in doubt, add attenuation and measure with a safe instrument first.

Safety goals

The goal is to protect:

  • SDR receiver inputs from overload or permanent damage;
  • RF front-end components on the Zynq/AD9363 board;
  • test equipment and USB receivers;
  • other spectrum users;
  • learners from ambiguous or unsafe lab setups.

Minimum RF connection checklist

Before any RF experiment, record these items in the lab notebook or measurement report:

Item Required note
Signal source Board, generator, replay file, antenna, or synthetic data.
Center frequency Expected RF or complex baseband frequency.
Sample rate / bandwidth SDR sample rate and approximate occupied bandwidth.
TX gain / output level Configured gain or estimated output power.
RX gain Manual gain setting when possible.
RF path Coax, attenuator chain, splitter, antenna, or over-the-air path.
Attenuation Nominal attenuation in dB and power rating.
Receiver input limit Known safe input level or conservative assumption.
Observation tool HDSDR, GNU Radio, MATLAB, Python, spectrum analyzer, or logs.

Pattern A: Synthetic data only

Use this for early labs and CI validation.

Python / MATLAB model -> IQ file -> reader / analyzer -> plots / metrics

Risk level: low. No RF hardware is connected.

Pattern B: External observation receiver

Use this for the first practical RF observation.

Zynq/AD9363 low-power signal -> controlled RF path or weak OTA signal -> RTL-SDR -> HDSDR / IQ recording

Risk level: medium. Keep TX power low, use attenuation when using coax, and avoid unknown gain settings.

Pattern C: Conducted RF loopback

Use this only when attenuation is explicitly calculated.

TX port -> fixed attenuator chain -> optional DC block -> RX port / RTL-SDR -> analyzer

Risk level: high if attenuation is missing or incorrect.

Conservative attenuation rule

For first experiments, prefer excessive attenuation over receiver overload. A practical educational starting point is:

TX output -> 30 dB to 60 dB attenuation -> receiver input

The exact value must be adapted to the board output level, receiver limit and measurement objective. Never assume that a receiver input is protected just because the signal is generated by another SDR board.

Gain discipline

Use manual gain settings whenever possible. Automatic gain control can hide overload and make measurements non-reproducible.

Stage Recommended practice
TX gain Start low and increase only after observing the signal.
RX gain Start low or moderate; avoid maximum gain during first connection.
HDSDR / SDR software Disable AGC for measurement-oriented captures when practical.
Analysis scripts Record gain, bandwidth, sample rate and file format in metadata.

Overload symptoms

Stop the experiment and reduce input level if you observe:

  • a flat-topped waveform;
  • a spectrum filled with unexpected wideband products;
  • many harmonics or mirrored images that do not match the model;
  • unstable noise floor when gain is unchanged;
  • receiver software showing clipping or ADC overload warnings.

Spectrum and regulatory discipline

Educational SDR experiments should normally use shielded or conducted setups, dummy loads, attenuators, or very low-power short-range links. Do not intentionally transmit on frequencies or power levels that require authorization.

For public demonstrations, prefer:

  • synthetic IQ files;
  • recorded captures;
  • conducted loopback with sufficient attenuation;
  • receive-only observations of known allowed signals.

Lab report RF safety section

Every RF-facing lab report should include:

RF safety summary:
- Source:
- Center frequency:
- Sample rate / bandwidth:
- TX gain or estimated output power:
- RF path:
- Attenuation:
- RX gain:
- Receiver protection assumption:
- Observed overload symptoms: yes/no
- Corrective actions:

Definition of a safe course lab

A lab is safe enough for course publication when it has:

  • an explicit RF path diagram;
  • attenuation assumptions;
  • receiver protection notes;
  • gain settings;
  • a fallback synthetic-data mode;
  • clear warnings around conducted TX/RX connections;
  • metadata fields for frequency, bandwidth, sample rate, gain and file format.