Amplifier measurements
An amplifier with a 1.5 dB noise figure and 10 dB of gain, measured with an ordinary analyzer, reads as if it had 10.5 dB. Without correcting for the second instrument, the measurement says nothing about the amplifier.
01What is measured in an amplifier
The specifications of an amplifier fall into four families, and each one has its instrument and its method:
| Family | Quantities | Instruments |
|---|---|---|
| Small signal | Gain, frequency response, input and output impedances, return loss | Generator and oscilloscope or voltmeter; network analyzer |
| Large signal | Output power, 1 dB compression, harmonic distortion | Low-distortion generator and spectrum or audio analyzer |
| Linearity with several signals | Intermodulation, third-order intercept point | Two combined generators and spectrum analyzer |
| Noise | Noise figure, noise temperature, input-referred noise | Calibrated noise source and analyzer |
The theory of each quantity was covered in Applied Electronics II, for example in feedback amplifiers and frequency response. Here the interest is in how to measure them well.
02Compression and intermodulation
As the input is raised, the gain starts to fall: the 1 dB compression point is the output power at which the gain has dropped 1 dB relative to the small-signal gain. With two close tones of frequencies \( f_1 \) and \( f_2 \), the third-order nonlinearity produces products at \( 2f_1 - f_2 \) and \( 2f_2 - f_1 \), which fall inside the band and cannot be filtered. Their level grows 3 dB for each dB of input, and the line that represents them crosses that of the fundamental at the third-order intercept point:
03Noise figure: the Y-factor method
The noise figure indicates how much an amplifier degrades the signal-to-noise ratio. To measure it, a calibrated noise source is connected to the input, an avalanche diode that is switched on and off. Its excess noise when on is expressed with the excess noise ratio, ENR. The analyzer measures the noise power at the output with the source on and off, and their ratio is the Y factor:
An amplifier with a real noise figure, measured with a noise source and an analyzer that has its own noise figure. The chart shows the noise figure of the system, the one that would be read without correcting, and the real one of the amplifier.
For very low-noise amplifiers a low-ENR source is better, which gives a smaller Y factor and one less affected by the linearity of the detector; for high noise, a high-ENR one. In all cases, the uncertainty of the source’s ENR is transferred almost entirely to the result.
04In the laboratory
Sweep the input power of an RF amplifier and plot the output to find the 1 dB compression point.
Combine two generators with a power combiner and measure the third-order intermodulation of an amplifier at several levels. Verify the 3:1 slope and calculate OIP3.
Measure the noise figure of a low-noise amplifier with a noise source and an analyzer with the noise figure option, with and without calibration of the second instrument.
05Common mistakes
- Measuring IP3 near compression. The slope stops being 3:1 and the result is not valid.
- Forgetting the analyzer’s contribution. With little gain in the amplifier, the analyzer’s noise figure dominates.
- Doing the Y-factor calculation in dB. The formulas apply to linear ratios.
- Measuring gain without matching impedances. Mismatch losses are confused with gain.
06Self-assessment
With −10 dBm tones and intermodulation 50 dB down, what is the OIP3?
\( -10 + 25 = +15 \) dBm.
With an ENR of 15 dB, Y = 10 dB is measured. What is the noise figure of the system?
\( 31.6/(10 - 1) = 3.51 \): 5.5 dB.
If the amplifier has 20 dB of gain and the analyzer a 20 dB noise figure, how much is subtracted?
\( (100 - 1)/100 = 0.99 \) as a ratio.
Why are third-order products the ones that cause the most concern?
Because they fall close to the signals, inside the band, and cannot be filtered.
07Further reading
- Keysight Technologies. Fundamentals of RF and Microwave Noise Figure Measurements. Application Note 57-1, Keysight. The Y-factor method, the correction for the second instrument and the uncertainties.
- David M. Pozar. Microwave Engineering. 4th ed., Wiley, 2012. Noise and linearity of microwave amplifiers.
- Bob Metzler. Audio Measurement Handbook. Audio Precision, 1993. Measurements on audio amplifiers: distortion, noise and response.