• State TI Thinks Resolved
  • Locked Locked
  • Replies 1 reply
  • Answers 1 answer
  • Subscribers 48 subscribers
  • Views 806 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.

ADS1256: Noise Limitations using the REF5025

Rakxit Patel
Prodigy 90 points

Part Number: ADS1256

Hi,

 

My customer is using the ADS1256 with the REF5025 external reference on a DAQ demo board. He had issues with noise associated with the device at a full scale differential input of 4V, and the customer was looking for an ADC capable of <50uVp-p over a 600Hz input bandwidth.

 

Based on the REF5025 datasheet spec, the customer believes the reference noise would be 3uVp-p*2.5=7.5uVp-p over a 0.1-10Hz bandwidth. Ignoring 1/f noise, this would be equal to 158nV/rtHz. Now, over a 600Hz bandwidth this would be 23uVp-p. Based on these calculations, the customer believes the demo board should be <50uVp-p at 4V input and 600Hz.

 

Can you confirm? Do you know if the customer can reach these spec with the ADC and Ref he has, or do you recommend another part?

Regards,

RP

  • 0
    TI__Mastermind 44780 points

    Hi Rakxit,

    This is an excellent question, but I do need to make several clarifications...

    • In the 0.1 to 10 Hz region, most of the noise is 1/f noise, so it is difficult to estimate the actual broadband noise up to 600 Hz using this information...However, I think the broadband noise is somewhere around 75 nVrms/rtHz (at least it was on an earlier revision of the REF5025). With this in mind, the reference noise will probably be around       sqrt(  7.5uVp-p^2   +  [6 c.f. * 75 nVrms/rtHz * sqrt(1000 Hz)]^2  )  =  16 uVp-p    ...Using root sum of squares (RSS) addition of the estimated 1/f and broadband noise.

    • Also, keep in mind that the actual noise bandwidth will depend on the RC filter and the ADS1256's digital filter, so the total noise may be more or less depending on the amount of filtering applied. I used a 1000 Hz bandwidth in the calculation above to account for the noise bandwidth of the ADS1256's digital filter when running at 2000 Hz. At this data rate, the digital filter attenuation is only about -1.3 dB @ 600 Hz. For any slower data rate this attenuation will be much more significant.

    • Also, since the reference noise scales with the input signal magnitude, only about 4V / 5V or 80% of the reference noise will be observed with a 4V signal and a maximum 5V positive full-scale input. Therefore, out of a total 16 uVp-p reference noise, only about 12.8 uVp-p of reference noise will be observed with a 4V input.

    • Finally, the ADC's noise and the input signal's noise contributions must also be considered and added to the reference noise. The ADS1256's noise performance at 2 kSPS, and gain of 1 V/V is about 4.173 uVrms or about 25 uVp-p (for c.f. = 6), and for the input signal, I'll assume that this noise is much much smaller than the ADC and reference (or else there this discussion is of no use).

    In summary, if the ADS1256 and REF5025 are used together and the ADS1256 is configured for 2kSPS and gain of 1 V/V (allowing for a 4V signal with about 600 Hz bandwidth), then I would expect the typical noise performance to be the RSS addition of the ADC Noise (25 uVp-p) and Reference noise (12.8 uVp-p) which is about 28 uVp-p. This is probably about the best noise performance one can expect in this configuration, however, this figure may increase due to noise from the input signal, noise from the power supplies, or noise due to poor PCB layout; but it would still seem that < 50 uVp-p with a 4V input signal is quite achievable.

    Best regards,
    Chris