The Impact of A/D Converters on the Dynamic Range of Spectrum Analyzers and Signal Analyzers

ADC dynamic indicators
Signal-to-noise ratio
For an ideal ADC, the quantization error within the Nyquist bandwidth is a white noise random signal with a quantization variance . Where q=2-N is the quantization interval of the A/D converter, and N is the A/D word length N bits.
The signal-to-noise ratio of the quantization noise is:

SNR=6.02N+1.76+101g(fs/2B) (1)

Where N is the number of bits of the ADC, fs is the sampling frequency, and B is the bandwidth of the analog input signal. The third term on the right side of the above equation indicates that increasing the sampling frequency (oversampling) can improve the signal-to-noise ratio.

Effective number of bits

In fact, the error of ADC is manifested as static and dynamic nonlinear error, and the dynamic error increases with the increase of input signal slew rate. Therefore, the actual measured signal-to-noise ratio is smaller than the theoretical one. The effective number of bits (ENOB) can be calculated by solving equation (1) for N.
ENOB(N)=[SNR-1.76-101g(fs/2B)]/6.02 (2)

Evaluation of ADC performance in spectrum analyzers and signal analyzers

Signal analyzer indicators affected by ADC dynamics

The typical circuit block diagram of a spectrum analyzer is shown in Figure 1. The ADC samples and converts the intermediate frequency (or video) signal after the intermediate frequency signal. The performance of the ADC affects the noise and signal-to-noise ratio of the spectrum analyzer, as well as the distortion indicators of the spectrum analyzer, including harmonics, spurious and intermodulation distortion.

For modern signal analyzers, the processing of intermediate frequency and video signals basically adopts digital technology and vector analysis (IQ analysis) technology. Therefore, the impact caused by DAC is very important.

Taking Rohde & Schwarz FSQ as an example, the A/D conversion and data analysis parts are further analyzed, and the


block diagram is shown in Figure 2.

· ADC quantization noise and signal-to-noise ratio theoretical analysis

For the narrowband IQ analysis module of FSQ, the ADC (marked ② and ③) used is 14-bit 81.6MHz sampling. According to formula (1), the theoretical value of the signal-to-noise ratio normalized to 1Hz bandwidth (B=1Hz) is:

SNR1(1Hz)=6.02N+1.76+101g(fs/2B)=6.02×14+1.76+101g(81.6×106/2)=162dBc/Hz

Therefore, for the narrowband IQ analysis module, the theoretical value of the ADC quantization noise is -162dBc/Hz.

For the FSQ wideband extended IQ analysis module (FSQ-B72), the ADC (marked ①) used is 8-bit 326.4MHz sampling. According to formula (1), the theoretical value of the signal-to-noise ratio is normalized to 1Hz bandwidth (B=1Hz):

SNR2(1Hz)=6.02N+1.76+101g(fs/2B)=6.02×8+1.76+101g(326.4×106/2)=132dBc/Hz

Therefore, for the broadband IQ analysis module, the theoretical value of the ADC quantization noise is -132dBc/Hz.

Relative to the actual signal analyzer, based on the filter bandwidth BW before the ADC, the actual noise that the analyzer can achieve under the corresponding bandwidth can be calculated. The noise calculation formula is:

N=-SNR(1Hz)+101g(BW/1Hz)

For the narrowband IQ analysis module, when the filter bandwidth is 10MHz (the measured signal bandwidth is less than 10MHz), the theoretical value of the ADC quantization noise is

N=-162+101g(BW/1Hz)=-92dBc

For the broadband IQ analysis module, when the filter bandwidth is 60MHz (the measured signal bandwidth is less than 60MHz