RF Isolation and Attenuation for 5G SDR Testing

[Pavel]

Short answer: Calculate a conducted RF budget that keeps receiver input below its safe and linear level and prevents unintended radiation. Use fixed attenuators, DC blocks, combiners or couplers rated for the band, then verify leakage with a separate monitor.

Before you start​

• Run private cellular experiments in an authorized, shielded or tightly conducted RF environment with sufficient attenuation between radio ports.
• Record RAN, core, UHD, firmware and operating-system versions together; a working configuration is a matched software and hardware set.

Step-by-step method​

1. Step 1: Find worst-case SDR TX output, UE input tolerance and every receiver maximum. Add enough fixed attenuation for safety plus adjustable attenuation for link testing.
2. Step 2: Account for combiner, splitter, cable and adapter loss, and ensure each component handles DC and power correctly. Terminate unused ports.
3. Step 3: Start with software TX gain at minimum, measure the level at the receiver-side connector and increase only within the written budget.
4. Step 4: Monitor the room spectrum with the conducted path active and enclosure closed. Stop if leakage exceeds the lab limit or other systems are affected.

Concrete example​

If SDR output could be +10 dBm and the receiving port should see no more than -30 dBm for the test, design at least 40 dB path loss plus uncertainty margin. A 30 dB attenuator alone is insufficient.

How to judge the result​

The RF path is valid when measured receive level matches the budget, every port stays protected and external leakage remains below the defined threshold.

What to record​

• Troubleshoot registration in order: RF synchronization, broadcast system information, random access, RRC, authentication, security mode and PDU session.
• Monitor CPU scheduling, USB or network throughput, clock state and RAN logs during a sustained cell test.
• Use spectrum analysis and power calculation to confirm that leakage outside the controlled path remains below the lab requirement.

Common mistakes​

• Changing several hardware, software or RF variables at once, which removes the controlled comparison needed to identify the cause.
• Treating one autoscaled screenshot or one unusually good result as proof without recording the settings and repeating the test.
• Software gain is not a safety device because configuration or restart can change it; fixed hardware attenuation provides the required baseline protection.

Final check​

Cold-start the core, RAN, SDR and test UE in the documented order. A private-5G baseline is complete only when attach and traffic repeat inside the verified attenuated RF environment.