The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> Circuit Simulators >> Sigma Delta ADC FFT simulation setup
https://designers-guide.org/forum/YaBB.pl?num=1467616575

Message started by Irascible on Jul 4^th, 2016, 12:16am

Title: Sigma Delta ADC FFT simulation setup
Post by Irascible on Jul 4^th, 2016, 12:16am

Hello,

I've read around extensively about FFT setup for a sigma delta ADC and I'm not sure if I'm not setting it up properly .

my current setup for a first order sigma delta ADC

sampling frequency : 38.4 kHz (for the ADC and FFT setup)
FFT bins : 65536
input signal : (31*38400)/65536 = 18.1640625 Hz

I'm using Mentor Graphics Eldo for the simulation

start time : 0.578064516 s (which is 10 and a half cycles of the input)
stop time : 2.284735711 s (31 cycles of the input from start time)
window : hanning alpha = 0.5

below is the output FFT of the ouput however the noise floor is extremely high. the x-axis is in log scale. there seems to be noise shaping however the harmonics are quite high so I wanted to
[url="http://i.imgur.com/tPVjNAN.png"]http://i.imgur.com/tPVjNAN.png

[/url]

plotting the wave using matlab yields the same output

y = csvread('file4.csv');
d = y(:,2);
d = hann(65536).*d(1:65536);
z = abs(fft(d));
x = 20*log10(z);
semilogx(x);
[url="http://i.imgur.com/2UvzmHnm.png"]http://i.imgur.com/2UvzmHnm.png

[/url]

Thank you in advance for your help, extremely appreciated :)

Title: Re: Sigma Delta ADC FFT simulation setup
Post by Ken Kundert on Jul 4^th, 2016, 6:06pm

It is my experience that Fourier analysis in SPICE simulators is a mess. Fourier analysis with SPICE is inherently tricky, and the simulators and design environments add dozens of user traps that make it very easy to produce inaccurate results without giving a good indication what the problem is, or even if there is a problem.

As such, It is always important to apply your Fourier analysis first to the input signal, then to the output signal. Of course your input signal contains no imperfections. Any imperfections that you see in the Fourier analysis results of the input and due to your measurement setup. Do not expect any more accuracy on the output signal that you see on the input signal.

Once you are getting accurate Fourier analysis results on the input signal, then you can shift to the output signal.

-Ken