• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 47 subscribers
  • Views 1342 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.

ADS1248 temperature reading unstable

Igor Jevtovic
Prodigy 100 points

Hi,


Does anyone know what could be the cause of unstable last 2 bytes from 24bits RTD temperature reading. The first MSB from 24bits is stable and the other two bytes (Mid and LSB) change values rapidly(can be seen on an osciloscope). 

My set up is similar to TI 3-Wire RTD Measurement System Reference Design document but more basic and modest, i.e no decupling capacitors (only one between Vrefout and Vrefcom), no low pass filter on Vref. Other config details :

Sensor: PT100

Rref:  820ohms

Idac = 1mA

Vref = 1,64V

Gain = 4

Data rate =  1000SPS

clock used: internal

mode: continuous

calibration: none

  • 0
    Intellectual 300 points
    Please show the schematic to me. Please pay attention, the sigma - delta ADC, if you use the high speed acquisition rate, How to ensure the ENOB. please reference the data sheet , adjust the acquisitions rate meet your System requirement. Thanks.
  • 0 in reply to Gaven Liu
    Intellectual 300 points
    Please reference the data sheet, page 12 - 14. And add a software filter.
  • 0
    TI__Guru* 97515 points

    Igor,

    Gaven is correct in that we would need some sort of schematic to help figure out where the noise is coming from. I think that the ADC raw data helps in determining the source of noise as well.

    Common sources of noise are the longer wires that are picking up EMI and power supply noise that couples into the reference or inputs. If you have an oscilloscope, you might try looking at the circuit input and reference (and anything else on the board) to see if there is any noise.

    Joseph Wu

  • 0 in reply to Joseph Wu
    Prodigy 100 points

    Hi Joseph and Gaven,


    Thank You both for trying to help.

    Here are schematics, a photo of the set up and random row data for the simple room temperature reading.

    As said before the set up is basic with simple blackboards and as seen on the photo 4 PT100 are connected but at t he moment I'm using only one.

    Row Data, taken randomly :

    2201759

    2202025

    2203975

    2203345

    2206739

    2205761

    2204661

    2203991

    2207242

    2203897

  • 0 in reply to Igor Jevtovic
    TI__Guru* 97515 points

    Igor,


    I took the raw data for your measurements and in general, the measurement looks ok. I built a table of the results. In the first column you have the raw data in decimal. In the second column, I take the data and divide by the full scale (which is 2^23). In the third column, I assume that you use two 1mA currents to drive the reference resistor, while using 1mA for each sense side of the 2 wire RTD. In the last column, I calculate the temperature, using an ITS-90 polynomial coefficient. You can see the table below:

    Raw data Code Ratio PT100 res Temp
    2201759 0.26247 107.6128 19.53
    2202025 0.262502 107.6258 19.57
    2203975 0.262734 107.7211 19.81
    2203345 0.262659 107.6903 19.73
    2206739 0.263064 107.8562 20.16
    2205761 0.262947 107.8084 20.04
    2204661 0.262816 107.7546 19.90
    2203991 0.262736 107.7218 19.82
    2207242 0.263124 107.8807 20.22
    2203897 0.262725 107.7172 19.80



    There is some noise in the measurement but it does give you something reasonable. The temperature measurement has a standard deviation of 0.22C, which is the equivalent of a noise standard deviation of 0.0889Ohms in the RTD. That resistance, reacting with 1mA is about 90uV.

    If you want to clean this up, there are a few things to try. I would enclose this in something so there aren't air currents moving around to change the temperature. It might help reduce the ambient temperature variation. I'd also start to include the decoupling capacitors to the device. A capacitor on the reference input may also help.

    Some of the problem may be just part of the setup. Using a protoboard may introduce extra resistance in the leads, some of which may be inconsistent. The loose wires may also pick up EMI that would also add noise to the measurement. I'd also consider power supply issues. If you have some sort of DC-DC converter as the supply, you may have more problems than if you used an LDO.

    Regardless, there are plenty of things to try to clean up the measurement. Just remember that your input noise is in the microvolts. Cleaning that up requires a lot of attention to detail.


    Joseph Wu

  • 0 in reply to Joseph Wu
    Prodigy 100 points

    Joseph,

    Thank you very much for you help and your time.

    You have  pointed me in the good direction in solving temperature unstability problem.

    Everything you've said will be taken into account and hopefully the problem will be solved.

    I'll inform you about the results.

    Kind regards,

    Igor J.