Stats

  • Locked Locked
  • Replies 2
  • Subscribers 126
  • Views 14777
  • Members are here 0
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

vsin frequency precision, and tips for coherent sampling

msharma

I am doing basic SNDR simulations on the SNDR of a 10 bit SAR ADC. I noticed that at moderate input sinusoid frequencies, the SNDR starts degrading. To eliminate any errors in my simulation setup, I ran a test simulation with an analogLib vsin source. The sinusoid frequency is 600e3/1024*23 = 13476.5625 Hz (600e3 = ADC sampling frequency, 1024 for FFT, 23 for coherent sampling/getting all ADC codes). I take FFT of the sinusoid output, with different period of the sinusoid. If the sinusoid perid is 1 cycle, the spectrum is as expected a single contribution at ~13476.5625 Hz (reported 13476.56 Hz. The FFT looks good upto time duration of 10 cycles. But as the FFT period is increased, non-fundamental frequency components appear in the plot. For FFT duration of 23 cycles, the fundamental (13476.5625 Hz) is about 580 mV (insted of 1 V), and there are two neighbouring contributions at 300 mV each (12.89062 kHz and 14.0625 kHz). I increased the precision of the FFT start and end times, but that did not improve the FFT.

Is this a spectre/fft precision issue, or some real phenomenon? Any related tips on cohereent sampling of ADC with spectre?

Thanks,

Mohit

  • auran
    Hello Mohit,

    One of the error contributors in FFT computation is the interpolation error, because when there is no evaluated point at the sampling instant, an interpolated value between the nearest evaluated points is used.

    To force evaluation with regular time intervals, the 'strobe period' transient option can be used. If the strobe period is set equal to a factor of the sampling period, the samples considered for FFT will now be evaluated points. I know that it is possible to get only your fundamental, with ~-300dB elsewhere in your FFT with this method.

    For more detailed info, you can check these documents at support.cadence.com:
    - How to Utilize a Windowing Technique for Accurate DFT
    - Getting accurate results when using the calculator DFT and the Fourier component.

    Hope it helps,

    Arda
  • msharma
    Thanks for the tips and pointers, Arda. I will try these.

    Mohit

Community Guidelines

The Cadence Design Communities support Cadence users and technologists interacting to exchange ideas, news, technical information, and best practices to solve problems and get the most from Cadence technology. The community is open to everyone, and to provide the most value, we require participants to follow our Community Guidelines that facilitate a quality exchange of ideas and information. By accessing, contributing, using or downloading any materials from the site, you agree to be bound by the full Community Guidelines.