• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 4 answers
  • Subscribers 48 subscribers
  • Views 1476 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.

ADS1261: ADS1261

Ram Ch
Prodigy 120 points

Part Number: ADS1261

Hello,

we are planning to use ADS1261 in our design. we have a requirement to measure 3 sensor outputs using 3 differential channels. By considering various factors like input voltage, reference voltage, no. of channels and cost etc.. we are preferring ADS1261. In our older design we used 3 ADC's to do this operation simultaneously. Below are my questions.

Q.1) Is there a possibility to do the simultaneous sampling of 3 channels using ADS1261 with single start command?

Q.2) If not simultaneous, what is the time required for getting the 3 channel data? for example, if ADS1261 is configured with 50SPS then the time required to do 3 channel's conversion would be 3*20 ms theoretically.

        so it require 60ms to get 3 samples. Is my understanding correct?

Please advise.

Thank you.

  • 0
    TI__Guru 58911 points

    Hi Ram,

    Thanks for your inquiry about the ADS1261. Here are the answers to your questions:

    1. The ADS1261 is a multiplexed ADC, not a simultaneous sampling ADC, so there is no possibility of taking simultaneous measurements using 1x device.
    2. The conversion latency is described in section 9.4.1.3 in the ADS1261 datasheet, and is characterized in Table 8. The conversion time per channel depends on several factors, including the data rate, filter choice, the DELAY bits, and whether CHOP mode is on. For example, if you configured the ADS1261 as follows:
      1. Data rate = 50 SPS
      2. Filter = sinc 3
      1. DELAY = 0001
      2. CHOP mode = off

    • Then you can see in Table 8 that these settings correspond to a conversion time per channel 60.43 ms. If you needed to convert 3x channels, that would be a minimum of 3*60.43ms = 181.3ms. If you instead chose a sinc 1 filter, the conversion time is 20.43ms, for a total of 3*20.43ms = 61.3ms. Note that lower sinc filter orders provide faster settling time but slightly higher noise

    Let me know if this answers your questions.

    -Bryan

  • 0 in reply to Bryan Lizon86
    Prodigy 120 points

    Thank you Bryan. You clarified my doubts regarding the simultaneous sampling and conversion times.

    I want to operate this ADC for +/- 2.5V differential input range. Can I consider Section 8, Table 1 (RMS and P-P values) for calculating the ENOBs? Basically I am looking for more than 16 ENOB's.

    In older design 3 ADC's are operated at max.160SPS each. So, to meet this requirement I need to operate ADS1261 with higher sample rate (probably >= 3*160 SPS). I need to go for 1200SPS as per datasheet.

    So, my question is if I go for higher sampling rates (with GAIN =1, Filter=Sinc1), do I get the ENOBS > 16 for the +/- 2.5v differential input range?

    For 24- bit ADC, with 16 ENOB output means, the most significant 16 bits are effective and rest 8 bits may be considered as noise. Is my understanding correct? 

    Please advise.

    Thank you.

  • 0 in reply to Ram Ch
    TI__Guru 58911 points

    Hi Ram,

    Your system effective resolution (ENOB) is determined by the signal-to-noise ratio. So you would need to determine the max signal you will measure, and divide that by the total system noise. If you take the log(2) of this value, that gives you the system effective resolution.

    If you just look at the ADC noise for example, that might not represent the total system noise, so you might not be able to get as much dynamic range as the ADC datasheet would suggest.. But the ADC can certainly provide 16-bit ENOB at 1200 SPS, sinc 1 filter. Table 2 actually shows the effective resolution using a 5V reference, which is 22.1 bits at the given settings.

    -Bryan

  • 0 in reply to Bryan Lizon86
    Prodigy 120 points

    Thank you Bryan. I understood your point.

    Also, Could you please advise on my below question.

    For 24- bit ADC, with 16 ENOB output means, the most significant 16 bits are effective and rest 8 bits may be considered as noise. Is my understanding correct? 

  • 0 in reply to Ram Ch
    TI__Guru 58911 points

    Hi Ram,

    What you have described is basically the definition of noise-free bits (NFB). With effective resolution the last bit could be toggling due to the statistical distribution of noise since this is an RMS value.

    -Bryan

  • 0 in reply to Bryan Lizon86
    Prodigy 120 points

    Hello Bryan,

    We are Planning to use the Internal Temperature Sensor of ADS1261 and compute the temperature in °C. We needed your support in understanding about the Internal Temperature Sensor. Please provide your answers for below questions.

    1) What is the range of Temperature's that ADS1261 Internal Temperature Sensor can measure? I see that ADC1261 operating temperature range is –40°C to +125°C. 

        Our product operating temperature range is –40°C to +85°C. I would like to know whether it can support  for –40°C to +85°C.

    2) What is the reference voltage we need to connect for Internal Temperature sensor? Currently our application will be operated with ±2.5V.

    3) In Table 43, I see positive and negative input selections for Internal temperature sensor. Do we need to enable both for it's operation?

    4)  In Section 9.3.1.3, I see a formula for Temperature calculation in °C. What does "Temperature Reading (µV)" mean in formula? Is it voltage or ADC count?

    5) What would be the output voltage range and resolution of Internal Temperature sensor?

    6) In Figure 33, for 25°C, the Temperature Sensor Voltage range can be 2 mV (i.e., 123.4 - 121.4). But in Page 9, Temperature coefficient is 420 µV/°C, which means for 420µV change the temperature will

    be incremented by 1 °C. I am missing some correlation between these two statements.

    Thank you.

       

       

  • 0 in reply to Ram Ch
    TI__Guru 58911 points

    Hi Ram,

    Here are the answers to your questions:

    1. The integrated temp sensor tracks the ADC die temperature well, and therefore can operate over the ADC's specified temp range (-40C to +125C). The device should not be operated outside this range, or behavior is no longer guaranteed as defined in the datasheet. Any temp range within these limits is inherently supported. Also, the temp sensor is a coarse measurement with an accuracy on the order of +/-5C (not guaranteed though). So it is likely not the best option for high precision measurements e.g. CJC in a thermocouple application.
    2. You can see from the Electrical Characteristics table that at 25C, the temp sensor output voltage reading should be 122.4mV (pg. 9). Inputting this value into equation 2 on pg. 29 indeed yields a value of 25C. Therefore, use the VREF value on pg. 9 to get the correct values, which is 2.5V.
    3. You would select both MUXP and MUXN as 1011b to enable the temp sensor
    4. "Temperature reading (uV)" means the voltage the ADC reports back from the temp sensor measurement. The ADC will give you a code which you would then convert to a voltage in uV that you can then enter into Equation 2 to calculate the temperature.
    5. You can calculate the expected output voltage range using the temp sensor value at 25C and the temperature coefficient, both given on pg. 9, as well as your system operating temp range. The expected resolution/accuracy is discussed in #1
    6. I am not sure what conclusion you are drawing here? Figure 33 shows the output code/voltage distribution at 25C. You can see that the error is quite wide, which is the point of my statements in #1 that this is not really a precision measurement. Therefore, the value at 25C (or likely any other temp value) can vary by several mV compared to the nominal.

    -Bryan