What causes the gain of my amplifier to be lower than the datasheet specification?
Diagnosing Low RF Amplifier Gain
A systematic approach to diagnosing low gain starts with verifying the easiest things first (bias, temperature) before investigating more complex causes (impedance matching, layout parasitics).
Diagnostic Checklist
- Step 1: Verify bias. Measure the actual drain/collector voltage and current at the device pins (not at the power supply, which may drop through series resistance). Compare to the datasheet test conditions. Even 10% deviation can cause 0.5-1 dB gain change
- Step 2: Verify temperature. Measure device temperature (IR thermometer on package). Compare to datasheet conditions (usually 25C). Apply the temperature coefficient (typically -0.01 to -0.03 dB/C) to estimate the thermal gain reduction
- Step 3: Measure input/output match. Use a VNA to measure S11 and S22 at the circuit board's coax connector ports. Calculate mismatch loss: ML = -10 log(1-|S11|^2). A return loss of 10 dB corresponds to 0.46 dB mismatch loss; 6 dB return loss corresponds to 1.26 dB
- Step 4: Estimate PCB losses. Calculate the insertion loss of all PCB traces between connectors and device pads. At 10 GHz on FR4, loss is approximately 0.3-0.5 dB/cm. On Rogers 4003C, loss is approximately 0.1-0.15 dB/cm. Include connector launch losses (typically 0.1-0.3 dB per connector)
- Step 5: Check for oscillation. A partially oscillating amplifier may compress on its own oscillation signal, reducing available gain for the desired signal. Check the output spectrum for any unexpected signals
Return loss 6 dB: ML = 1.26 dB, RL 10 dB: ML = 0.46 dB, RL 15 dB: ML = 0.14 dB
Total loss budget: Gain_measured = Gain_device - ML_input - ML_output - L_traces - L_connectors
Thermal gain coefficient: delta_G ~ -0.015 dB/C x delta_T [typical for GaAs]
Frequently Asked Questions
How much gain loss is acceptable versus the datasheet?
In a well-designed circuit board implementation, the total gain should be within 1-2 dB of the datasheet S21 at frequencies below 6 GHz, and within 2-3 dB at frequencies above 18 GHz. The difference comes from PCB losses, connector transitions, and mismatch effects that are not present in the datasheet's on-wafer or reference test fixture measurement. If the gain is more than 3 dB below the datasheet, there is likely a design or assembly error.
Can the PCB substrate cause significant gain reduction?
Yes, especially at high frequencies. FR4 substrate has a loss tangent of 0.02-0.03, causing 0.3-0.5 dB/cm of trace loss at 10 GHz, which can add 1-3 dB of total loss for a typical circuit layout. Switching to a low-loss substrate (Rogers 4003C with tan_d = 0.0027, or Rogers 3003 with tan_d = 0.0013) reduces trace loss by 5-10x and is essential for circuits above approximately 3 GHz.
Does the amplifier's gain change with input power level?
Yes. As input power approaches the amplifier's P1dB compression point, gain decreases (gain compression). The datasheet small-signal gain applies only for input power levels well below P1dB (typically 10-20 dB below). Operating near compression causes 1 dB or more of gain reduction. Check that the input power level during your measurement is consistent with small-signal conditions.