Question about input amplitude of sigma delta modulator

[twteng]

Hi all,
I am designing a 2nd order 1 bit sigma delta modulator, the feedback reference is 0.6V, if set input amplitude to 0.6V(0dB), why the signal power in the spectrum which is not 0dB? Please see attached file.
Please give me a help if you know this issue, Many thanks for your help.

 

[YL.WANG]

Hi,

Have you checked that whether the output data on which you do fft is normalised to 1, i.e., if you have output data 1.2 and 0 (the supply), then fft will not give a 0dB, but the real level your input signal has..

 

[twteng]

Hi, thanks for your reply,
I think that I have do normolized, because the reference voltage is 0.6V~1.2V,
and input amplitude is 0.6V, then 0.6/(1.2-0.6)=1, the signal power should be 0dB, is that right?? please correct me if my concept is wrong.
Thanks a lot.

 

[YL.WANG]

You are welcome~~

Okay~~ let's do it step by step~~

What are the voltage levels of you data on which you do fft? i.e. the output of your modulator.

 

[twteng]

Hi,
My voltage level is two(one bit), and supply voltage is 1.8V
Thank you for your kindly help and reply.

 

[YL.WANG]

so you have 2 levels : 0V and 1.8V at the output, am I right?

Then you perform fft to this data? If yes, you cannot expect that fft gives you 0dB at the signal frequency, it will be the real level of you 0.6V signal, that's a minus value.

Am I right till now?

 

[twteng]

1.Yes
2.The real level is not 0.6V, its 0.9V (-0.915dB), please see attachment for more detial.

 

[YL.WANG]

Okay~ let's try this.

you normalise the output data: when it is larger than 0.9V, give it the value +1; if it is smaller than 0.9V, give it the value -1, then it has two levels +1 and -1, instead of 1.8 and 0, do the fft and plot it again to see what happens.

I am doing a DSM now, I also meet this problem before, my output has 1.2V and 0V, if I directly do fft on it, I cannot get the dB I expect; after I normalise it, it looks very good~

I dunt know how you do the fft, if you deal with the output data already as I write above, then it might be an other story...

 

[twteng]

Hi YL
Sorry, I can't get your explain about the "when it is larger than 0.9V, give it the value +1; if it is smaller than 0.9V, give it the value -1", because of the comparator output which is just high(1.8V) and low(0V).

To be brief, You means that I just use -1 and 1 to instead of 0 and 1.8 from my output data, and then do the fft again, is that right?
if yes, why use -1 and 1?

It would be appreciated if you could elaborate more detials,Thanks a lot.

 

[JoannesPaulus]

Just one quick question: why do you care about the full scale of the output in your fft? The only things that really matter here are the difference between your fundamental tone and your harmonics, and the total integrated noise in your band of interest.
You are doing the fft of a digital sequence, so you will see the power of that signal but you are actually interested in the meaning of that signal...

 

[YL.WANG]

Just one quick question: why do you care about the full scale of the output in your fft? The only things that really matter here are the difference between your fundamental tone and your harmonics, and the total integrated noise in your band of interest.
You are doing the fft of a digital sequence, so you will see the power of that signal but you are actually interested in the meaning of that signal...

Yes~that's really a good saying, whether normalising the output to the full scale or not gives nothing different in the parameters like SNR or so... it is just a different outlook..

Yes~I mean it, 1.8V is +1, 0V is -1. You can take a look at this question, why the power of the full scale is 0dBFS? In fact, for example, in your case, the FS is 0.6V, its power is not 0dB at all... it is just a relative value..when the FS is regarded as 1, you get the 0dBFS.

 

[twteng]

Hi YL
Let me tell you a good news, the issue has been sloved by following your comment.
I'm rellay thanks for your great help and so lucky to meet you, and also
thanks to the JoannesPaulus's reply.

 

[somyasinha]

Hi,

I need to design a sigma-delta ADC with the following specs:

1. Effective Resolution - 14 bit
2. Effective SNR - 80 dB
3. Bandwidth - 80 kHz
4. Dynamic Range - +/- 5V
5. Number of bits in comparator - 1 or 2

I want to use comb-filter or any other technique which is easy to understand. Where can I get some material which tells me how to design the ADC and digital filter and implement it in MATLAB?

Most of the papers and books I read only have theory and results. I cannot find out how to implement.
Please help.

Somya