• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 1 answer
  • Subscribers 46 subscribers
  • Views 372 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.

AFE7225: Input voltage/current noise around 1kHz

JMac
Intellectual 2270 points
Part Number: AFE7225
Other Parts Discussed in Thread: AFE7222,

Hello team,

I am working with a customer that wants to use the AFE7225 or AFE7222 as the ADC for their scope application (they also need the DAC, and would prefer to go with a single integrated option). We are looking for some additional information regarding what the input noise would look like at 1kHz. I know high-speed ADCs don't typically include a lot of low-frequency noise information in their datasheets, but I was hoping we might have some additional data that would provide some insight into the expected noise. Even if you can't provide an actual measured value, any comments would be helpful - for example if you expect the noise performance to worsen as high/early as 100kHz, then that helps, or if you expect noise in the 10s of uV that would be another red flag. Thanks in advance!

Regards,

Jon

  • 0
    TI__Mastermind 18095 points

    Jon:

    You are correct, noise performance at such a low frequency offset is not typically measured on a high speed ADC.  The device will have some 1/f flicker noise that will degrade noise performance at very low offsets.  I do not have a quantifiable number off-hand.

    Check out the datasheet in figure 6-1 as an example.  You can see a hint of the 1/f noise rising up near DC.  Note there is also a true DC component "spur" there as well.  The RMS noise floor is roughly -105 dBFS in that plot.  The 1/f noise rises up a bit, to roughly -102 dBFS.  The DC component is at about -92 dBFs for that capture.  In short, there will be some impact, but 100 kHz is still high enough that I do not think that there will be a huge contribution from 1/f noise.

    --RJH

  • 0 in reply to RJ Hopper
    TI__Intellectual 2270 points

    Hi RJ,

    Thanks for the reply. I do want to clarify one thing, however:

    "100 kHz is still high enough that I do not think that there will be a huge contribution from 1/f noise."

    What about at 1kHz itself? I apologize for the confusion, I was really just using 100kHz as an example to say if it's already bad there it will probably be much worse at 1kHz, which is their target frequency. The scope will experience a wide signal range, from DC to 1MHz, but it's important to the customer that they have as little noise as possible as early as 1kHz. I believe they have some very specific use case in mind. I'm looking into very low noise JFET front-end solutions for the signal chain, but I wanted to make sure that high 1/f ADC noise won't render it all moot.

    -102dBFS -> magnitude = 0.000007943282347242822 of FS, multiply by 2V full scale range and you get 15.886uV of noise. Would you then just divide by sqrt(1kHz) to get ~502nV/rtHz of noise? Or am I approaching this the wrong way?

    Thanks in advance,

    Jon

  • 0 in reply to JMac
    TI__Intellectual 2270 points

    Hi RJ,

    Bumping this thread again, can you just confirm that the above calculation is proper? If the AFE7225 1/f noise is too high we need to begin exploring other ADC options ASAP.

    Thanks,

    Jon

  • 0 in reply to JMac
    TI__Mastermind 18095 points

    Jon:

    I think that the calculations are off because the bin size of the FFT is not accounted for from the values in the graph.  It is difficult to know for sure the bin size of the FFT from the datasheet graph so I attacked the calculations based on the SNR performance at 65 MSPS.

    I calculated the NSD normalized to a 1 Hz BW.  From there, I converted to a voltage based on the full scale voltage.  The calculations reveal an RMS noise floor of about 37 nV/rtHz.  This the average noise floor.  Let's assume that the 1/f noise degrades SNR by 3 dB localized to the very low frequency.  Then the noise voltage degrades to around 52 nV/rtHz.

    --RJH

  • 0 in reply to RJ Hopper
    TI__Intellectual 2270 points

    Thanks for the response and calculations RJ! I will share this information with the customer.

    Regards,

    Jon