[Moved]: Sigma Delta ADC Linearity & 3rd Harmonic

[Shady Ahmed]

I am working on a 3rd order, 1 MHz Sigma Delta ADC with the following Specs
3 level quantizer and OSR = 32.

I am facing a linearity problem that i don't know how to fix it.
The system achieved max SQNR = 69 dB at Vin Full Scale and Max Frequency (i.e: Fin = Fb)

The following figure shows the PSD:



However, the simulation with Fin = F_bandwidth/3 ,, the max SQNR = 60.76 dB which means I am losing about 9 dB ..
and here is the PSD with the lower input frequency



I know that means that i have harmonics issue (2nd , 3rd ,, ) ,, How can I fix it ?

 

[pankaj jha]

Re: Sigma Delta ADC Linearity & 3rd Harmonic

I am working on a 3rd order, 1 MHz Sigma Delta ADC with the following Specs
3 level quantizer and OSR = 32.

I am facing a linearity problem that i don't know how to fix it.
The system achieved max SQNR = 69 dB at Vin Full Scale and Max Frequency (i.e: Fin = Fb)

The following figure shows the PSD:
View attachment 120323


However, the simulation with Fin = F_bandwidth/3 ,, the max SQNR = 60.76 dB which means I am losing about 9 dB ..
and here is the PSD with the lower input frequency
View attachment 120324


I know that means that i have harmonics issue (2nd , 3rd ,, ) ,, How can I fix it ?

That is a very practical problem...
1) If i take fin=fb, then while calculating the snr I am not including any harmonics (noise calculations take into consideration frequencies only upto FB)
2) So the best way is to choose a fin such that around 5*Fin or 9*fin is less than FB, so that we can include atleast 5/9 harmonics