The high-speed ADC analog input interface must consider six conditions, how many do you know?

System design with high input frequency, high speed analog-to-digital converters (ADCs) is a challenging task. The ADC input interface is designed with six main conditions: input impedance, input drive, bandwidth, passband flatness, noise, and distortion. Look at the six conditions listed here, do you understand?

input resistance

The input impedance is the characteristic impedance of the design. The internal input impedance of the ADC depends on the type of ADC architecture that the ADC supplier will provide on the data sheet or on the product page. The voltage standing wave ratio (VWSR) is closely related to the input impedance and measures the amount of power reflected into the load within the target bandwidth. This parameter is important because it sets the input drive level required to achieve the full-scale input of the ADC. Maximum power transfer occurs when the source impedance is equal to the load impedance.

The high-speed ADC analog input interface must consider six conditions, how many do you know?

Figure 1. Input Z/VWSR on the network analyzer

The example shown in Figure 1 is the input impedance and VSWR curve of a front-end network measured using a network analyzer. The input impedance is the characteristic impedance of the design. In most cases, it is 50Ω, but a particular design may require a different impedance.

VSWR is a dimensionless parameter that reflects how much power is reflected into the load within the target bandwidth. This parameter is important because it sets the input drive level required to achieve the full-scale input of the ADC. Note that the higher the frequency, the greater the drive power or gain required to drive the ADC input to full scale.

Input driver

Input drivers are related to bandwidth characteristics and can be used to set the system gain required for a particular application. The input drive level should be determined before the front end design begins, depending on the selected front end components such as filters, transformers, and amplifiers.

Bandwidth and passband flatness

Bandwidth is the range of frequencies that the system will use. Passband flatness is the amount of fluctuation in a given bandwidth; the cause of the ripple may be the ripple effect or the slow roll-off characteristic of the Butterworth filter. Passband flatness is typically less than 1 dB and is critical to setting overall system gain.

noise

Signal-to-noise ratio (SNR) and distortion requirements are helpful for the choice of ADC and are therefore generally determined early in the design. The ratio of the amount of noise seen by the converter to its own amount of noise is the SNR. SNR is related to bandwidth, signal quality (jitter), and gain. Increasing the gain also increases the noise component associated with it.

distortion

The distortion is measured by the spurious free dynamic range (SFDR), which is the ratio of the rms full scale to the rms value of the peak spurious spectral components. SFDR is mainly controlled by two factors. The first factor is the linearity of the front-end equilibrium mass, which is primarily related to the second harmonic distortion. The second factor is the required gain and input matching. Higher gain requirements increase the difficulty of matching. In addition, high gain requirements compress the headroom of the ADC, increasing nonlinearity, and their nonlinearity increases as more power passes through the external passive components. This effect is generally considered as the third harmonic.

The high-speed ADC analog input interface must consider six conditions, how many do you know?

Figure 2. Noise floor of an ideal 12-bit ADC using a 4096-point FFT

Figure 2 shows the output of a 4096-point FFT of an ideal 12-bit ADC and some basic operations. The theoretical SNR is 74 dB. The noise is distributed throughout the Nyquist bandwidth. The FFT increases the processing gain because it deals with small "bins" whose width is equal to the sampling frequency divided by the number of FFT points. For a 4096 point FFT, the processing gain is 33 dB. This is like narrowing the bandwidth of an analog spectrum analyzer. The actual FFT noise floor is equal to the SNR plus the processing gain, as shown in Figure 2.

The FFT noise floor under the above conditions is equal to 74 + 33 = 107 dBFS. In some systems, the results of multiple independent FFTs are averaged, which does not reduce the FFT noise floor, but only reduces the amplitude variation of the noise component.

8Pt Low Voltage Switch Cabinet

General Electric Switchgear,Switchgear And Transformer,Eaton Low Voltage Switchgear,Epe Switchgear

Shandong Shunkai electrical equipment co., LTD. , https://www.chinasdsk.com