• State
  • Locked Locked
  • Replies 7 replies
  • Subscribers 46 subscribers
  • Views 708 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

ADC12D1800RFRB noise level

Ethan wu
Prodigy 220 points
Other Parts Discussed in Thread: WAVEVISION5, LMX2531, LMX2541

I have an ADC12D1800RFRB board. When I ground all input with 50 ohms terminators and calibrate ADC. I read background noise at the range of of 20 p-p (out of 4096), which basically take 5 bits precision away from the ADC. Is this the noise level I should expect.

  • 0
    TI__Expert 3940 points

    Hi Ethan,

    No; something is incorrect in the setup.  Did you calibrate the ADC?  Forgetting to calibrate often leads to incorrect performance. 

    After calibrating, with WaveVision5 in the FFT view, please send a screen shot with the dynamic performance parameters pinned open.  This will show if the performance is correct.

    Kind regards,

    Marjorie

  • 0 in reply to Marjorie Plisch
    Prodigy 220 points

    Hi Marjorie, I am not quite sure what you meant by "with the dynamic performance parameters pinned open". Here is a screenshot of FFT view attached. The ADC is set to DESI mode. With all the input pins grounded with 50 ohms terminator and I did calibration before acquisition. Everything else was in default status. I also attached one raw data screenshot for your reference.

    Thank you,

    Ethan

  • 0 in reply to Ethan wu
    TI__Expert 3940 points

    Hi Ethan,

    Thanks for the screen shots.  Sorry my request was unclear before; in the frequency domain view, there is a tab on the left-hand side called 'FFT Readouts'.  This tab will show the SNR, from which we can calculate the NSD and check whether it matcheds the datasheet.  Please either let me know the SNR or attach another screen shot with the 'FFT Readouts' tab visible.  On the top of the tab, you can 'pin' it open.

    By the way, can you see the clock on your board?  Is it an LMX2531 or LMX2541?  How long ago did you purchase the board?

    Kind regards,

    Marjorie

  • 0 in reply to Marjorie Plisch
    Prodigy 220 points

    Hi Marjorie,

    Here is the screenshot with the tab open. By clock I assume you are referring to the U12 on the board, which is actually labeled as VM14ABE3-311778EA.  the serial number of the board is SN89, I bought it last year. 

    Thank you very much

  • 0 in reply to Ethan wu
    TI__Expert 3940 points

    Hi Ethan,

    Since you bought the board last year, it has the LMX2531 on it; the LMX2541 is a recent update, but the LMX2531 actually performs better, so that is good for you.

    From the FFT Readouts tab, the SNR = 58.67dBFS.  Since the part is in DESI Mode, the Nyquist zone covers 1.8GHz, so the calculated NSD is:

    -SNR - 10*log10(NyquistBW) = -151.22dBFS/Hz

    From the datasheet, the typical NSD = -154dBFS/Hz.  2.8dB is more than I expect missing in chip-to-chip performance variation.  Out of curiosity, how does DESQ Mode perform? 

    Kind regards,

    Marjorie

  • 0 in reply to Marjorie Plisch
    Prodigy 220 points

    Hi Marjorie,

    DESQ mode give SNR=60 dBFS. Is this normal? I am not sure whether I am reading the datasheet correctly. The datasheet says the effective bit would be a bit more than 9 bits, so we expect to lose 3 bits from noise?

    Regards,

    Ethan

  • 0 in reply to Ethan wu
    TI__Expert 3940 points

    Hi Ethan,

    This yields a NSD of -152.55 dBFS / Hz, which is pretty close to what I expect and could be considered within the normal variation of the device. 

    Yes, the inherent noise of the chip means that the noise floor is at a level which effectively drops the ENOB to between 9 and 10 ENOB for low input frequencies. 

    Kind regards,

    Marjorie