Post Go back to editing

how do I know if an ADC is suitable for IF sampling, i.e AD9653 or AD9655?

I have some question on determining the IF sampling capability of ADC's :

I've checked the ADCs selection table for ADC > 10MSPS

  • some ADC datasheets explicitly mention that the device is suitable for IF sampling, i.e : AD9461, LTC2195, ...
  • some ADC dasheets also specify the max IF frequency, i.e. AD9268-125 says 'up to 300MHz', LTC2107 says 'up to 500MHz'
  • some ADC datasheets even give specs, i.e. AD9652 contains plenty of specs for Nyquist 1, 2 an 3 zones
  • some ADC datasheets don't mention IF sampling at all, i.e. AD9653 and AD9655

now I'm interested in these last 2 devices, because of their low-power and low-pin count. But I'm not sure if these are suitable for IF sampling. I'm not a high-speed ADC expert, so can anyone explain me what paramaters I need to check to know if the ADC is suitable or not for IF sampling? I guess the full power analog bandwidth and the sample and hold-time are important, but some explanation would really be appreciated!

  • Hi

    I completely can understand where you are coming from here. I apologize for the various "standards" and language used in these datasheets. Ultimately, the IF performance comes down to 

    - how much is the analog input (or -3dB) bandwidth of the sampler

    - how much jitter is in the system that could limit performance

    If you start to compare the two parts, we can compare the -3dB BW and Aperture jitter, for starters

    Part #        |  -3dB BW (MHz)         |  Aperture jitter (fs_rms)


    AD9655     |  500                             | 80

    AD9653     |  650                             | 135

    I dont know the IF frequency range you plan to use. But I can infer two things

    - the AD9653 has higher -3dB BW. So you can undersample higher frequencies compared to the AD9655

    - however, the AD9653 also has higher jitter, which means the SNR degradation due to jitter will also be higher for AD9653 compared to AD9655

    This is evident in the SNR (dBFS) table below. All this information is derived from the datasheet. 

    Freq (MHz)      |  AD9655 (VREF = 1.4V)         |  AD9653 (VREF = 1.3V)


    9.7                   |  79.6                                        | 80.1

    200                  |  75                                           | 72.1

    300                  |  55                                           |  ? possibly much worse

    This shows that the AD9655 is clearly much better than the AD9653 when it comes to undersampling applications. However, if the performance is good enough for your application, please go ahead and choose the part that fits your needs. 

    Hope this helps answer your questions.

    Thanks

    Umesh

  • - however, the AD9653 also has higher jitter, which means the SNR degradation due to jitter will also be higher for AD9653 compared to AD9655

    What is the effect of this jitter? What is the cause of the degraded SNR?

    • Does the amplitude decrease due to the change in the sinx/x curve, so related to the imperfect Dirac sampling?
    • Or will the noise be higher? Is there a noise contribution versus frequency?
  • What is the effect of this jitter? What is the cause of the degraded SNR?

    here is some information for you to peruse

    www.analog.com/an-756

    https://www.analog.com/media/en/training-seminars/tutorials/mt-007.pdf

    https://www.analog.com/media/en/training-seminars/tutorials/mt-008.pdf

    https://www.analog.com/en/analog-dialogue/articles/adc-modeling-tools-speed-up-eval.html

    the degradation in noise is an artifact of the jitter or uncertainty in the sampling window. 

Before You Switch


Switching languages will make ADI Explorer unavailable. Resume your session by switching back to English and reopening ADI Explorer.