• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Subscribers 49 subscribers
  • Views 875 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

LMP91000EVM with H2S MICROceL HS

Trilogis
Prodigy 170 points
Other Parts Discussed in Thread: MSP430F249, LMP91000EVM, LMP91000, LPV521, MSP430F2252, ADS1110, LMP91002

Hello,

I have an LMP91000EVM interfaced with a msp430f249 board.

I can fully control the LMP91000, read and write registers.
I can read the ADC conversion result, which fully matches VOUT voltage on the board.
External Vref 2.5V is OK and stable.

I have mounted a MICROceL HS sensor from CITY Technology. The sensor is in a dedicated connector, wires are as short as possible.

With the following configuration, Rload = 10Ohm, TIA Gain = 350K, Vref = 2.5V, Vref divider = 67%, I get VOUT = 1.697 V. Is this result realistic ???
Is this calculation correct: Iwe = (VOUT - 0.67 x 2.5) / 350K =  0.0628uA which is H2S = -1.6ppm

Do you have any experience with this MICROceL HS sensor ?

Is it possible to have a higher gain with TIA ???

Best Regards, Stéphane

  • 0
    TI__Genius 9655 points

    Hi Stéphane,

    the senor you are using generates a smaller current than an electrochemical cell sensor. So any offset of theTIA or of the internal zero have a  weight in the measurement. I sugegst to take a differential measurement between C1 and C2 pin. Check the input impedance of the ADC, if less than 1Mohm I suggest to buffer the C1 pin.

    In the DS we do not reccomend to increwase the RTIA becasue the output equivalnt capacitance of the electrochemical in conjunction wiht the RTIA (high value) could generate a zero/pole in the transfer function of the TIA which could decrease the stability of the TIA itself.

    In your case with MICROcel, it could be possible that the equivalent output capacitance of the sensor is lower than the one of the electrochemical cell so you could try to increase the RTIA with an external resistance.

    In this case you do not remove the offset of the TIA and the offset of the internal zero, but you mitigate their effects.

    regards,

    Domenico.

  • 0 in reply to domenyko
    Prodigy 170 points

    Hi Domenico,


    Thanks for you answer.

    If I want to make a differential measurement between C1 and C2, can I re-use the ADC161S626CIM which is present on the LMP91000EVM ?
    I would unsolder R4 and connect C1 directly to input IN-.

    Regards,

    Stéphane

  • 0 in reply to Trilogis
    TI__Genius 9655 points

    Hi Stéphane,

    you can use the ADC1616S626 present on the EVM. I suggest to connect C1 to the input of the ADC trhough a resister, same value of R4 in order to have a balanced input.

    regards,

    Domenico

     

  • 0 in reply to domenyko
    Prodigy 170 points

    Hi Domenico.

    I have connected the ADC1616S626 as suggested. C1 goes to IN- through a 20 Ohm resistor.

    I have the following configuration:
    RTIA = 350K
    Rload = 10R
    Vref is External (2.5V)
    Internal Zero is 67%

    With no sensor connected, the ADC gives very stable values, +/- 3LSB around 0.

    With a MICROcelHS connected, ADC output becomes noisy, +/- 80 LSB, but still no DC component.

    If I touch C1 with an oscilloscope probe, I get a cleaner signal, around 600mV.
    This value is not really what I expected, it means 60ppm of H2S.

    What is wrong with C1 ? Is it because it is not buffered ?
    Why touching C1 with the oscilloscope changes the behauviour of the system ?
    Any clue how to enhance ?


    Regards, Stéphane

  • 0 in reply to Trilogis
    TI__Genius 9655 points

    Hi Stephan,

    the problem is the driving capabiloty of the C1 pin. You can buffer that pin or increase the reistance between the C1 pin and the ADC (then you have to change also the the R8), If you increase the resistance case you have to lower the sample rate in order to give enought time to the filter to set.

    I think that when you conenct the scope you generate a DC path vs. GND which sources the current required by the ADC's input.

    P.S. in order to maximize the output dynmaic of the LMP91000 you should select an internal zero equal to 20% VREF.

     

  • 0 in reply to domenyko
    Prodigy 170 points

    Hi Domenico,

    I really need to save power as I will be powered by single 3V coin cell battery.
    Would the lpv521 Opamp from TI be a good candidate for buffering C1 ?

    What would be the max sampling rate ?
    We target 1 sample/second, which should be sufficient to detect H2S in the air and give alarm.
    Can you suggest a candidate for the A/D converter ?

    Regards, Stéphane

  • 0 in reply to Trilogis
    TI__Genius 9655 points

    Hi Stéphane,

    the LPV521 burns very low current but has poor performances in terms of capacitance driving capability. In order to drive the ADC + filter it requires tens of Kohm between the op amp and the input of the ADC.

    Since you need precision, you will have to perform some acqusitions in 1 second and then avergae them using the SAR. Altertativley you could use a sigma delta ADC wiht integrated PGA (which acts as a buffer) instead of SAR ADC + OP AMP.

    The Sigma delta will take more time to acquire a data, but it is already averaged.

    A solution based on ADS11xx family will simplify the design, impove the performances and minimize the power consumpion which seems to be the main constraint in your battery powerd application.

    regards,

    Domenico

  • 0 in reply to domenyko
    Prodigy 170 points

    Hello Domenico,

    Thank you for your responses !

    For next step, I'm currently designing a new test board with lmp91002, ads1110 and msp430f2252. It will be powered with a CR2032 coin cell.

    Do you have any advice for ultra low power voltage reference ?

    Regards, Stéphane

  • 0 in reply to Trilogis
    TI__Genius 9655 points

    Hi Stéphan,

    sorry for the delay in the answer, but in this period we have summer vacations in Europe.

    Your question cannot be answered in an easy way. As you know higher ower consumption means low noise voltage reference, and in som case better intial accuracy and low drift.

    Which is the temperature range of your application ?

    This tool http://www.ti.com/lit/snvd001 helps in the selection of the right part.

    Regards,

    Domenico