• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 48 subscribers
  • Views 1087 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.

ADS1278: ADS1278

Eric Cui13
Intellectual 920 points

Part Number: ADS1278
Other Parts Discussed in Thread: ADS1178, ADS1271

Hi,

We used ADS1278, REF=2.5V,
Now when the P and N terminals of the differential input are shorted together, the output value of the AD is around 0.1mv, and the perceived accuracy is a bit low. At the same time, we measured the EVM evaluation board, also in this order of magnitude.
 Do not know what the reason is that the conversion accuracy is relatively low, how can we improve the conversion accuracy, the specific picture as shown below:

Thank you.

  • 0
    TI__Guru 69366 points

    Hi Eric,

    Thanks for your post!

    Please quantify what you mean by "low accuracy."

    Based on the few data points that you show in your post, the max variation in codes is only 121. This translates to approximately 36 uVpp, which is exactly the typical noise spec for High-Resolution Mode. Other modes generally have more noise, so 36 uVpp is pretty good.

    Also, please keep in mind that the inputs are floating with respect to ground. Our noise measurements are conducted while driving the inputs to a common-mode voltage near mid-supply.

    Best Regards,

  • 0 in reply to Ryan Andrews
    Intellectual 920 points
    hi Ryan,

    1. According to what you said, the difference is 121, that is, the 6 to 7 bits in the 24-bit ADC chip are in jitter, that is, the actual accuracy is only about 17 bits. We hope to achieve 22 to 23 digits of accuracy.

         2. We are also using ADS1178, a 16-bit ADC chip. The precision of this chip can reach 14 bits, that is, the last two bits of jitter. Why is the 24-bit accuracy so much worse?

         3, we have referenced the same type of product ADS1271, the chip's accuracy is higher than ADS1278? We feel that the data output of that product is very low noise

         4. Is there a reference design? Can you provide it?

    Regards,

    Eric
  • 0 in reply to Eric Cui13
    TI__Guru 69366 points

    Hello Eric,

    1. 6 to 7 bits of noise agrees with the typical datasheet noise specifications. The effective resolution of an ADC is calculated as log2 [VREF / sqrt(2) / Vn_rms]. For 5.5 uVrms, this equates to about 18.3 bits of effective resolution. It is common for an ADC to use n-bits to represent the conversion result while some of the least significant will toggle randomly due to noise.
    2. For a 16-bit converter, the weight of one LSB is much larger for a given reference voltage (2^8 times larger to be exact). If the same noise is present in a 24-bit ADC and a 16-bit ADC, the noise will affect more of the least significant bits in the 24-bit ADC. In other words, if the average noise is less than 2^8 times the LSB of the 24-bit ADC, the same noise will have almost no affect on the 16-bit ADC. This is why our typical noise spec is 1 LSBpp in the ADS1178.
    3. The ADS1271 is also a 24-bit ADC with slightly higher thermal noise than the ADS1278. For the same oversampling ratio and digital filter bandwidth, the ADS1278 should have less noise in the result.
    4. The only reference design that we have available for these devices are their respective evaluation modules. Please find them on the device product page.

    Best Regards,