• State
  • Locked Locked
  • Replies 6 replies
  • Subscribers 45 subscribers
  • Views 980 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.

ADS1220: Open thermocouple detection with ADS1220IPWR?

JaPaBo
Intellectual 310 points
Part Number: ADS1220
Other Parts Discussed in Thread: ADS124S08, ADS1248

I'm having some difficulty with open thermocouple detection using the ADS1220IPWR.  The system is presently working quite well for temperature readings using the on-chip temperature sensor for CJC measurement and using dual k-type thermocouple inputs.   I'd like to understand the issue to confirm that this is a good means for detection of the open thermocouple condition. 

When I enable the burnout current sources and thermocouples are not installed, I get full scale output as expected (+16mV).  If thermocouples are installed, results are not what I expect (e.g., 12mV or so when I expect closer to 2.5mV). I've attached a schematic snippet below.  My expectation when checking this thermocouple channel would be for 10µA to flow through (1) R104, (2) the thermocouple reistance (somewhere from 1-100Ω), and (3) R105.  This should be equivalent to an additional differential 2.5mV (10e-6A * 250Ω).  In one specific case with ~50Ω thermocouple I see 11.6mV indicated (on a bench with no temperature gradients or actual thermal votlages developed).

When attempting to take readings with the burnout sources enabled, registers are configured as follows:

Register 0: 0x0e or 0x5E (depending on which channel is measured)

Register 1:0x01

Register 2,3: default values

Comments appreciated!

  • 0 in reply to JaPaBo
    TI__Guru** 113765 points

    Hi Jack,

    Our usual recommendation for the TC measurement is shown in the ADS1220 datasheet in the application section.  Here a large value resistor is used as a pull-up on the AIN+ pin and an equal large value pull-down resistor on the AIN- pin.  This type of configuration will keep the TC in the correct common-mode input range when the TC is connected and will produce a full-scale result if there is an open.

    The burnout current sources will accomplish the same thing if the TC is open in that one current source attempts to pull the ADC AIN+ input to AVDD and the AIN- input to AVSS.  In the case of a connected TC with the circuit you are using you are setting a bias point of 1/2 of the supply on the AIN- pin.  What is not clear and what I suspect is the voltage drop at AIN1 is lower than expected due to the current source pulling the input toward ground.  As there is a current divider at the junction of R105 and R106, the voltage at that pin may be changing and the ADC is operating in a non-linear region.

    To verify I would suggest measuring the voltages on both sides of R105 to check to make sure that nothing unusual is taking place when the BCS is turned on as compared to when it is turned off.

    Best regards,

    Bob B

  • 0 in reply to Bob Benjamin
    Intellectual 310 points

    Hi Bob,

    I appreciate your input!  We have used the reference topology previously (and the DNP'ed resistor in schematic will still enable this approach) albeit with smaller bias resistors than you probably recommend.  We had observed inaccuracies in the thermocouple response caused by voltage divider action between the pull-up (PU) and pull-down (PD) resistors and the thermocouple resistance (e.g., PU = PD = 100kΩ and thermocouple ~20Ω causes ~8°C error for K-type thermocouple with 3.3V power supply).  It was thought that we could resolved by either (1) using  the circuit as shown in the schematic - still using the voltage divider to establish CM voltages - and using burn-out current sources to detect a broken thermocouple or (2) by increasing the PU/PD resistances drastically.  Because of concerns about drift/stability in such large resistors (potentially dirty industrial environment) and because we don't readily control thermocouple resistances/stability, it was thought that the approach relying on burnout current sources would be technically better.

    I will post R105 terminal voltages shortly as a separate reply.

    Is there a differential current specification for the 2 burnout sources (similar to figure 29 for the IDAC sources - not sure if this is actually a different drive circuit)?  Maybe we can reduce the magnitude of the PU and PD resistors to a point where the current difference in burnout sources will not push the circuit out of CM range of the ADC?  If so, some assistance in establishing the necessary PU/PD values would be appreciated.

    I did find it curious that TI seemingly didn't reccomend use of the burnout current sources integrated into their product and instead reccomends an approach that causes modest error in measurement and higher BOM cost.  Is there any particular reason?  Are there any app notes supporting the burnout sources as a means of detecting the thermocouple failure?

    -Jack

  • 0 in reply to JaPaBo
    TI__Guru** 113765 points

    Hi Jack,

    The BCS differ from the IDACs in that the IDACs are trimmed and matched where the BCS are not.  Unfortunately I have not had a chance yet to actually test your circuit out myself due to some other commitments, but I do intend on testing it.

    As to the pull-up and pull-down resistors, this is a legacy method used by many customers already and they are accepting of it.  Also, this configuration will also place the input within the PGA input range when the TC is connected. 

    You have used a different method for establishing the correct common-mode and have also added two resistors.  So the difference in BOM cost really doesn't apply for this case, in my opinion.  We do have other devices with an internal VBIAS (ADS1248, ADS124S08, et al) which will set the correct common-mode without external resistors.

    So yes, the pull-up and pull-down will add to the BOM cost for the ADS1220 and will add some error to the TC measurement.  Many customers have dealt with that error, probably because noise dominates in their system. One advantage of using the pull-up and pull-down is you don't have to use the BCS for open wire detection.  Why is that an advantage?  Because you don't have to worry about analog settling from the caps charging up from the BCS.

    In my experience more often than not, customers not only want to know if a sensor is open, but they are also concerned about shorts and degradation of the sensor.  I'm not aware of any application notes regarding the usage of the BCS although there have been discussions in the past on E2E.  We now have code for making the temperature measurements using various sensor types for the ADS124S08 and are now working on similar code for the ADS1220 family of devices.  However, none of the code deals directly with the BCS.

    I agree that an application note and/or code examples of how to best use the BCS would be a good idea.  Unfortunately I have about 20 other similar ideas that I would like to see accomplished as well.

    Best regards,

    Bob B

  • 0 in reply to Bob Benjamin
    TI__Guru** 113765 points

    Hi Jack,

    It has been a while since we last communicated.  If you have further questions feel free to respond to this thread or create a new thread with your question(s).

    Best regards,

    Bob B