RF engineering FAQ

Use a stated free-space, time-gated, waveguide, coaxial or fixture method and record the reference reflector or standard, antenna or port geometry, distance, sample size, incidence, polarization, frequency grid, gating and uncertainty. Reflectivity and transmission loss answer different questions and must not be mixed.

After coupon screening, install the exact backing, adhesive, seams, corners and compression. Repeat the original enclosure, chamber, near-field or coupling measurement with unchanged cables, probes, lids and instrument state. Verify the entire target band, not only the deepest null.

Reject the result when the baseline is not repeatable, edge diffraction dominates the coupon, an open seam remains, or performance drifts after power, thermal, humidity, vibration or maintenance exposure. Retain raw traces, configuration photos, material lot and drawing revision so a later replacement can be compared at the same boundary.

Frequency, match, loss, power, PIM and DC behavior must be checked alongside impulse waveform, current, sparkover, residual voltage and follow-current limits. Gas-discharge-tube protectors are broadband and may pass DC, but their static and impulse sparkover, follow-current behavior, cartridge life and residual energy require coordination with the equipment port. Quarter-wave or shorting-stub designs can offer low residual voltage in a defined RF band, but bandwidth and DC behavior depend on the mechanical filter. Hybrid stages add finer clamping at the cost of more topology-specific limits. Compare nominal discharge current only when the waveform is identical: an 8/20 microsecond induced-surge rating is not interchangeable with a 10/350 microsecond partial-lightning-current rating. Also compare impulse sparkover, residual voltage, RF power, VSWR, insertion loss, PIM, environmental seal and replaceability.

Compare impulse current only when the waveform, polarity, number of shots and test class are the same. An 8/20 microsecond induced-surge value and a 10/350 microsecond partial-lightning-current value describe different energy boundaries; neither alone states the voltage that reaches the protected port.

The RFQ should pair operating and guard bands, return and insertion loss, transmit peak and average power, PIM, maximum line voltage and DC or control current with impulse sparkover, residual voltage, follow-current behavior, cartridge life, sealing and maker-defined acceptance tests.

No single topology suits every installation; choose by RF bandwidth, DC continuity, maximum line voltage, residual protection target, power, PIM and service strategy. Gas-discharge-tube protectors are broadband and may pass DC, but their static and impulse sparkover, follow-current behavior, cartridge life and residual energy require coordination with the equipment port. Quarter-wave or shorting-stub designs can offer low residual voltage in a defined RF band, but bandwidth and DC behavior depend on the mechanical filter. Hybrid stages add finer clamping at the cost of more topology-specific limits. Compare nominal discharge current only when the waveform is identical: an 8/20 microsecond induced-surge rating is not interchangeable with a 10/350 microsecond partial-lightning-current rating. Also compare impulse sparkover, residual voltage, RF power, VSWR, insertion loss, PIM, environmental seal and replaceability.

A replaceable GDT is useful when broad bandwidth and DC continuity are required, provided sparkover, follow current, residual pulse and cartridge service fit the protected port. Quarter-wave or shorting-stub structures can provide a low residual path in a bounded RF band, but their filter geometry determines bandwidth and DC behavior.

Hybrid protection can add a finer downstream clamp, yet every added stage changes RF loss, match, power, PIM and coordination. Reject any topology offered without a frequency response, line-voltage boundary, named impulse waveform, earth-path requirement and post-event service procedure.

Mount at the zone boundary with a short low-inductance bond, retain RF/DC/PIM baselines, inspect after an event and follow maker-defined retest or replacement limits. Use a conductive bulkhead or the shortest specified ground connection at the cable entry; a long pigtail adds inductive voltage exactly when surge current rises fastest. Keep protected and unprotected cable routes separated, respect port direction where specified, weather-seal the exposed interface, apply connector and mounting torque, and avoid cable strain. Coordinate the coax protector with antenna bonding, feeder grounding kits, mains and data-line SPDs and any downstream receiver limiter. Before service, record S11, S21, DC continuity or isolation and PIM where relevant. After a severe event, inspect for arcing, heat or connector damage and repeat breakdown or cartridge, return-loss and insertion-loss checks under the maker's procedure.

Commission with the final connectors, jumpers, mounting torque, weather seal and earth connection. Save photographs, hardware identity, S11/S21, DC pass or isolation, leakage and PIM when relevant, together with calibration and reference planes.

After a severe event isolate the protected equipment, inspect for arc, heat, corrosion, water and deformation, then apply only the maker-approved cartridge, sparkover, leakage and RF checks. Replace when history is unknown or any electrical, mechanical or sealing limit fails; continuity by itself is not a return-to-service decision.

Select the RF amplifier family from the stage's job in the signal chain, not from the largest gain or power number. Fix the input and output reference planes and complete operating band first. Then reject families that cannot meet the required noise, linearity, output, control, supply, mechanical or thermal boundary.

Use this decision order:

1. Low-noise amplifier (LNA): use near a receiver input when cascaded noise figure and sensitivity dominate; also verify blocker tolerance, input survivability, stability and gain.
2. Gain block: use for predictable fixed gain and broadband 50-ohm integration when moderate noise and output performance are acceptable.
3. Driver amplifier: use when the next stage needs a defined drive level; compare output P1dB or OIP3, gain flatness, stability and maximum input.
4. Power amplifier: use for the final delivered RF output; define average and peak power, waveform, back-off, efficiency, heat removal, mismatch and protection.
5. Variable-gain amplifier: use when closed-loop level control or dynamic-range management is required; include gain range, step or control law, settling time and linearity at every gain state.

Do not compare unlike conditions. Gain, noise figure, P1dB, IP3 and efficiency are meaningful only with frequency, bias, temperature, load, waveform and reference plane stated. Reject a candidate when full-band data, maximum-input behavior, mounting boundary or production acceptance evidence is missing.

Before requesting a model, provide: operating and guard bands; minimum and maximum input; required gain and flatness; average and peak output; PAPR plus EVM, ACPR or IMD limit; CW, pulsed or modulated duty; source/load VSWR; supply and control; ambient or baseplate temperature; cooling, outline, connectors, quantity and qualification needs.

Continuous-wave and pulsed power amplifiers cannot be selected from the same headline power number. CW operation is limited mainly by sustained RF output, DC input, heat flow and long-term gain stability. Pulsed operation adds peak voltage and current, pulse width, repetition frequency, duty cycle, rise/fall time, droop, recovery and protection behavior.

• CW: verify rated output at the stated frequency, baseplate or ambient temperature and cooling boundary until thermal equilibrium.
• Pulsed: verify peak output across the complete pulse, not only at its center; record pulse width, PRF, duty cycle, gating timing, droop and recovery between pulses.
• Modulated signals: add bandwidth, PAPR and EVM, ACPR or spectral-mask limits. Peak capability does not prove linear modulated performance.
• Mismatch: state the expected VSWR, phase coverage, duration and whether protection may reduce or interrupt output.

Planning example: a 100 W CW output with an assumed 40% DC-to-RF efficiency requires about 250 W DC and leaves roughly 150 W to remove as heat. A 1 kW pulse at 10% duty cycle also averages 100 W RF, but it is not thermally or electrically equivalent: the supply, transistor, matching network and protection still experience the 1 kW peak condition and pulse-to-pulse temperature cycling.

Reject a rating when the vendor does not state waveform, frequency, temperature, cooling, pulse conditions and measurement reference plane. For an RFQ provide average and peak output, pulse width, PRF, duty cycle, modulation and PAPR, linearity limit, burst length, expected VSWR, supply transient limit, cooling interface, ambient/baseplate range and required acceptance test.