• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Answers 1 answer
  • Subscribers 46 subscribers
  • Views 704 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.

ADS1118

William Denale
Prodigy 50 points
Other Parts Discussed in Thread: ADS1118

I setup a case with TI on 12/18 and have not had a response on this question.  The ADS1118RUGR application report SBAA189 - September 2011 states on  page4 in the last paragraph:  "For extremely long thermocouples, these 1M ohm resistors may impact measurment accuracy.  Increasing the resistance can alleviate these effects.  Alternatively, two 1M ohm resistors connected connected as a resistor divider to one of the inputs can maitain the midpoint bias without affecting measurement results."

Using the reference diagram Figure 3 on the same page, with AIN2 and AIN3 as the inputs, is there a procedure or formula to detemine the resistor values for a particular length of thermocouple?  If an off the shelf handheld thermocouple with appropriate connector with a 2 foot coiled cord were attached to this circuit how is it determined to pick the resistor values?

 

  • 0
    TI__Guru* 98025 points

    William,


    The point that the article was making is that the two 1M resistor will add some current that will react with the series resistance of the thermocouple wire. With extremely long lengths of wire, this might be an issue (imagine two 1M resistors across 3.3V generating 1.65uA of current). I don't think a 2-foot coiled cord will have enough series resistance to appreciably affect the measurement.

    Note that the ADS1118 is a 16-bit ADC, so the measurement resolution isn't going to be as good as our 24-bit ADCs. This part will work best with something like a k-type thermocouple where the resolution is ~ 41uV/C.

    I talked to the author and he said that he tested thermocouples with wires about 1m in length and didn't have any issues.


    Joseph Wu

  • 0 in reply to Joseph Wu
    Prodigy 50 points
    Thanks for your reply Joseph. The Thermocouple is a T type off the shelf commercially available. It sounds like the length of the wire does not make a huge difference. I measured the wire and the coiled part is only 1 foot while coiled with about 1 extra foot of this just straight for ease of use. The probes, wire and connectors are all matching for T type. I didn't choose the part but I do understand the higher bit parts would be more accurate. It is good to know the 1M wire did not have any issues.
  • 0 in reply to Joseph Wu
    Prodigy 50 points
    Joseph,
    If we changed the RPU and RPD resistors to 10 megohm, because we no control over the length and resitance of the wire a customer may choose for a Thermocouple, we need to make sure our accuracy stays the same. Do you think this is an issue if we used 10Megohm resistors and what other changes would be required if we did this? The reason for 10 megohms is to make the offset smaller than the ADS1118 resolution. Does TI see any issue with changing the resistors to 10 megohm?
  • 0 in reply to William Denale
    TI__Guru* 98025 points
    William,


    No problem. If you have any other questions, just post back.


    Joseph Wu
  • 0 in reply to William Denale
    TI__Guru* 98025 points
    William,


    I think 10M would be fine. There would be resistor noise, but I don't think you'd see it with the resolution of the ADS1118. If this were for a wider bandwidth 24-bit ADC, it may be a consideration.


    Joseph Wu
  • 0 in reply to Joseph Wu
    Prodigy 50 points
    Joseph,
    If the thermocouple were disconnected, what type of response would the ADC have using 10 megohm resistors? When the thermocouple is disconneted will it go to rails to detect open circuit? Is the 10Meg ohm stiff enough to supply that part of the thermocouple circuit?
  • 0 in reply to William Denale
    TI__Guru* 98025 points
    William,


    Short answer: Not sure, it might be enough.

    Longer answer: It's true that you'd want them to pull apart on an open circuit. It won't completely, but it might be enough to get you even out of the full scale range. The ADS1118 has an equivalent input impedance that would probably hold the inputs together. The input impedance is described on page 15 of the datasheet.

    I believe that if the thermocouple opened up, you'd have the pull-up, the input impedance, and then the pull-down all in series. Additionally, I think you'd have the equivalent common-mode input resistance in parallel with the pull-up and pull down. If you're using the ADS1118 and set the range to +/-256mV, the differential input impedance would only be 750k, but there's no given common mode impedance for that range, so I don't really know.

    If I had to guess, it would probably be just enough to get you outside 256mV (I think the 6M of common-mode input impedance holds through the different ranges). It would take a little experimentation to be sure. If you do test it, and are using a multimeter, just remember the meter likely has 10M of input impedance as well.


    Joseph Wu
  • 0 in reply to Joseph Wu
    Prodigy 50 points
    Joseph,
    Is it possible for yourself and the TI chip designer to have a confercne call tommorrow on this? I would like to have myself and one other person form my company so we can discuss this. Please let me know. Both of us are on the East Coast (New York and Florida). We would like to try an understand this more in depth. I understand your input but none of us seem sure exactly on this issue. Please let me know. We are available 8am-5pm East coast time.
  • 0 in reply to William Denale
    TI__Guru* 98025 points
    William,


    I'll contact you through email that you provided as a login. We'll take this thread offline.


    Joseph Wu