Spectral inversion feature of AD9082

Hi SAH 

Thank you for using AD9082. I have moved this thread to the correct forum:  Mixed-Signal Front Ends (MxFE). Someone here should be able to support you.

Hi, there is no spectral inversion with default settings in hardware. But we'll review the document and get back to you shortly.

-YH

Hi, I did a quick simulation for the signal flow through FDUC with a CW single tone as follows and I confirmed that there is no signal inversion.

Single tone CW signal can be defined as,

• DATA_I   = Ai * cos(wb * t)
• DATA_Q = Bi * sin(wb * t)

Output signals through the quadarture mixer are as follows

• OUT_I = cos(wnco * t) * DATA_I - sin(wnco * t) * DATA_Q,
• OUT_Q = sin(wnco * t) * DATA_I + cos(wnco * t) * DATA_Q

SSB will be generated in this case, and the the frequency of the modulated signal become as follows, which confirms no signal inversion.

• Bi = Ai   : base band frequency = +wb, modulated signal frequency = wnco + wb
• Bi = -Ai :  base band frequency = -wb, modulated signal frequency = wnco - wb

Hope this helps, and let us know if any further clarification is needed.

-YH

Thank you YH for your reply. We agree on the output of the CDUC. However, when Mod Mux is configured to configuration0 (see table 105), only the real part of the CDUC output (OUT_I = cos(wnco * t) * DATA_I - sin(wnco * t) * DATA_Q) is sent to the DAC. Right?

Yes, only the OUT_I signal goes to a DAC in case of Mod Mux Configuration 0. Still I see upper side band only in the spectrum on both of (OUT_I + 1.0 j * OUT_Q) or (OUT_I) alone in case of Bi=Ai.

-YH

Thank you very much YH for the clarification. 
I also managed to simulate the default configuration, verifying that the spectrum is NOT inverted indeed.

I share the simulation code here, just in case.

clc
clear
close all

fs = 1e3; % sample rate
fc1 = 100; % freq. of the input signal
fc2 = 150; % freq. of the input signal
fw = 250; % NCO freq. of the mixer

n = 0:1:fs;
data = exp(2*pi*j*fc1.*n/fs) + 0.5.*exp(2*pi*j*fc2.*n/fs);
data =data(:);
%NCO = exp(2*pi*j*fw.*n/fs);
NCO = cos(2*pi*fw.*n/fs) + j.*sin(2*pi*fw.*n/fs);
NCO=NCO(:);

subplot(3,1,1)
[pxx,f] = pwelch(data,[],[],[],fs,"centered");
plot(f,pxx);grid on; xlabel('freq (Hz)');
title('input signal');

subplot(3,1,2)
[pxx,f] = pwelch(data.*NCO,[],[],[],fs,"centered");
plot(f,pxx);grid on; xlabel('freq (Hz)');
title('Complex output of the mixer');

subplot(3,1,3)
[pxx,f] = pwelch(real(data).*real(NCO) - imag(data).*imag(NCO),[],[],[],fs,"centered");
plot(f,pxx);grid on; xlabel('freq (Hz)');
title('Real part of the mixer output');

Thank you very much for confirmation and sharing the simulation code !