How to simulate the ENOB of a SAR ADC

[howardtang85]

Does anyone know how to simulate the ENOB of a Charge Redistribution SAR ADC based on cadence spectre simulator?

Thanks!

 

[palmeiras]

Hi!

You need to do a transient simulation and apply FFT in your output sine wave. After this, you can calculate the SINAD, and consequently ENOB, based on the following relation:

You apply a input sine wave at the input of your ADC, and then, you can use an ideal DAC (for instance, described in verilog-A) to convert your output bits to a sine wave. Then, you do the FFT of your curve. Save your points, and in a excel spreadsheet, you can calculate the SINAD. After this, the using the above equation, you find the ENOB.

 

[howardtang85]

Thanks for the solution!

In addition, may I know the ideal DAC should have an ideal transformation characteristic (eg. 1 V using 10-bit, so each LSB will be 1/1024), or same as the ideal transformation characteristic given by the ADC (assume there's a known gain or offset error in the ADC)?

 

[erikl]

I think the ideal DACs from the VLOGA or AHDL libs have a fullscale settable parameter. If not, you can always convert the analog value via a vcvs source from analogLib (sources).

 

[palmeiras]

Hi howardtang85,

If I undertood well your question, It is better that you DAC described in verilog-A acts as IDEAL without anynon-ideality.