SAR ADC系列4：SAR ADC简介+Matlab建模（7月11日更新版）

单端输入SAR ADC MATLAB 建模程序代码：

以下是一个单端输入SAR ADC MATLAB 建模程序代码，可使能输入Noise和电容Mismatch
（完整版代码看个人简介）

% ------------------------------------------------------
% Description:
%        This progran shows a Nbit-bit binary weighted SAR ADC behave model
%
% Input:
%    -Avin : the analog input signals
%           the magnitude of input signal should be [0,1]
%    -Nbit : ADC resulotion
%    -noise ：When high，would add random gauss noise to input signal and 
% Output:
%    -Dout : Nbit-bit ADC quantization output data
% Author：
%    xsjkk
% Modified data：
%    2023-7-10 original version。
% ------------------------------------------------------

function [dout,cap_redunt] = sar_adc_ideal(Avin , Nbit , noise , mismatch)

%=======================================================
%            Check Input Variable
%=======================================================
if nargin <4 , mismatch = 0;  end   % mismatch默认值为0
if nargin <3 , noise    = 0;  end   % noise默认值为0
if nargin <2 , Nbit     = 12; end   % 默认12位
if nargin <1 , Avin     = []; end   % 默认无输入，必须要外加输入信号，一般是sin波

if isempty(Avin)
    error('No input signal')    % 检测有无输入信号
end

if (Nbit < 2)
    error('Increase ADC resolution')    % 检测位数
end

%=======================================================
%            Add Noise to Input Signal
%=======================================================
% When enabled, this function is uesed to model ADC noise, like refer-
% -buffer, comparator, etc
% Noise is confined within 2LSBs
if (noise ~=0)
    sprintf('Noise feature have been enabled');
%   Missing codes
else
    fprintf('noise feature is NOT enabled, \n');
end

%=======================================================
%            SAR ADC Modeling
%=======================================================
cap = zeros(1,Nbit);        % 根据位数确定reference的长度
for cnt = 1:Nbit
%   Missing codes
end
cap_redunt = cap;

%=======================================================
%            Add mismatch to Cap
%=======================================================
LSB = 1/2^Nbit;          % 定义LSB
%   Missing codes
cap_mis = zeros(1,Nbit);
for cnt = 1:Nbit
%   Missing codes
end
% adding mismatch to SAR ADC if this feature is enabled
if (mismatch ~= 0)
    cap_redunt = cap_redunt + cap_mis;
end

%=======================================================
%            SAR ADC A-D Conversion
%=======================================================
%Avin = Avin + cap_r/cap_shift;
dout = zeros(length(Avin),Nbit);
%therm = cumsum(cap_redunt(1:7));
for j = 1:length(Avin)  
%   Missing codes
end
%=========SAR ADC Model Finished=======================

差分输入SAR ADC MATLAB 建模程序代码：

以下是一个差分输入SAR ADC MATLAB 建模程序代码，可使能输入Noise和电容Mismatch
（完整版代码看个人简介）

% ------------------------------------------------------
% Description:
%        This progran shows a Nbit-bit binary weighted SAR ADC behave model-
%        -with Vcm-based switching
%        -and obtain SNDR,ENOB,SFDR for a given sine input
%
% Note: The pre-requisite for this function to work properly is coherent
% sampling - in other words, input must be integer number of input cycles.
%
% Perform FFT and plot the figure
%
% Author：
%    xsjkk
% Modified data：
%    2023-7-11 original version。
% ------------------------------------------------------

%%
clc;
clear;
N=5; % ADC resolution
M=10^6; % unit: M
n=10^-9; % unit: n
m=10^-6; % unit: m
num=2^14; % FFT number
vdd=1.2; % supply voltage
Vfs=1.2; % signal swing
Vref=1.2; % reference voltage
Vcm=Vref/2; % common mode voltage
%    Missing Code
% Version 1：Without Noise
Vin_p=Vfs/2+Missing Code
Vin_n=Vfs/2-Missing Code
% Verison 2：With Noise
%    Missing Code

%%
sig_c=0.001; % capacitor mismatch
C_arr_1=Missing Code
C_arr_2=Missing Code
C_dev_1=Missing Code
C_dev_2=Missing Code
%    Missing Code

%%
for j=1:num
    Vxp=Vin_p(j);
    Vxn=Vin_n(j);
for i=1:N-1
%    Missing Code
end
%    Missing Code

D(j,:)=B;
end

Dout= D*2.^[(N-1):-1:0]';

%%
%%%%%%  FFT calculation  %%%%%%%%%%%
numpt=length(Dout);
spect=fft(Dout);
dB1=20*log10(abs(spect));
%    Missing Code

figure;
set(gcf,'Units','centimeters','position',[25,10,15,8]);
%    Missing Code
grid on;
zoom;
% axis tight;
set(gca,'linewidth',2);
set(gca,'fontsize',14,'fontname','Times New Roman');
set(Q,'linewidth',1.5);
xlabel('ANALOG INPUT FREQUENCY (MHz)','fontsize',14,'fontname','Times New Roman'); 
ylabel('AMPLITUDE (dB)','fontsize',14,'fontname','Times New Roman'); 
a1=axis;

%%
% 找出输入频率并计算功率谱
%    Missing Code
span=0;%输入频谱范围
%    Missing Code
%%
%    Missing Code

%%
%求动态参数
%    Missing Code
SNDR=10*log10(Ps/(Psum-Ps));
ENOB=(SNDR-1.76)/6.02;

%%
text(a1(2)/5,a1(3)/5, sprintf('SFDR=%3.2fdB\nSNDR=%3.2fdB\nENOB=%3.2fbit@%3.4fMHz',...
   SFDR,SNDR,ENOB,fin),...
    'EdgeColor','black','LineWidth',2,'fontsize',14,'BackgroundColor',...
      [1 1 1],'Margin',5,'fontname','Times New Roman');

SAR ADC MATLAB 建模Sim：

（7.12日先欠着，睡觉了，下次更新）

Matlab2019版本之后带一个SARADC的模型，理想的性能。

SAR ADC 动态指标计算MATLAB程序代码：

上文的差分SAR ADC中已经涉及到了，下面再给出两种方法：
（同样完整代码看个人简介）

clear;clc;
N = 1024;
M = 37;
fin = 1e6;

fs = N*fin/M;
Ts = 1/fs;
%    Missing Code
LSB = 2/2^10;
vin = sin(2*pi*fin*t);
%    Missing Codes
figure(1);
plot(t,vin);

s = fft(vin);
s = abs(s) + 1e-6;
s_dB = 20*log10(s);
%    Missing Code
figure(2);
%    Missing Code
plot(1:N/2,s_dB);

% 计算SNR
%    Missing Code
power_sig = s(sig_bin);
power_noise = sum(s) - power_sig;
snr = 10*log10(power_sig/power_noise)

clear
close all
%-----------生成测试文本-------------%
%---------- 用仿真输出文本取代---------------%
% fs = 1E6;  %采样率
% fsig = 10E3; %信号频率
% ts = 1E-3; %仿真时间
% snr = 1E3;
%     Missing Codes
% noise_source = randn(1,Ns);  %噪声源
%     Missing Codes
% fid1=fopen('E:\sin_wave.txt','w');                                                  
% count=fprintf(fid1,' %f \n',wave);     
% plot (tsim,wave)
% fclose(fid1);
%--------------------------------------------------------%


result = importdata('E:\sin_wave.txt'); %ADC仿真路径下结果
figure(1);
plot(result);
% result = result(11:138);
d_len = length(result);
%     Missing Codes

%***************** 画出功率谱 **************************************
%     Missing Codes
figure(2);
plot(xz,10*log10(pow_spec)) %画功率谱图，Y轴是对数刻度
title('ADC Output Spectrum') 
xlabel('fi/fs')
ylabel('Power (dB)')
axis([0,0.5,-120,0])
grid
%***************** 求SNDR *****************************************
%     Missing Codes
sndr = 10*log10(ps/(sum(pow_spec)-ps))  %求SNDR
%***************** 求SFDR *****************************************
%     Missing Codes
sfdr = -10*log10(max(pow_spec))   %现在SFDR就是最大的杂波幅度了

SAR ADC 静态指标计算MATLAB程序代码：

太晚了，明天更。

PS：

一般，动态性能和静态性能不同时测，因为：动态性能测试时—>想要动态性能好，就不能满量程输入；而静态性能测试时—>想要不缺码，就需要超量程输入。

校准：

校准前的INLDNL，可以看到DNL和INL0有好多超过0.5LSB的，INL还有超过1的。校准之后好很多