• State TI Thinks Resolved
  • Locked Locked
  • Replies 16 replies
  • Answers 2 answers
  • Subscribers 48 subscribers
  • Views 6085 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: Connecting three 3-Wire TPM100

Noam Weidenfeld1
Intellectual 780 points

Part Number: ADS1248
Other Parts Discussed in Thread: ADS124S08, TS3A4751

Hi,

Can I connect three 3-Wire TPM100 to a single ADS1248?

Is there any connection scheme for such application

Thanks

Noam

  • 0
    TI__Guru** 100785 points

    Noam,

    I'm not sure what a TPM100 is, but you are referring to an PT100 RTD? If so then, yes you could measure up to four of these devices. However, there are limitations. I'll explain how this is setup, what some of the limitations are and what other options you have.

    Using the ADS1248 you could measure up to four 3-wire RTDs. Here is the setup that you would use:

    Looking at the diagram, you can see each RTD element in a dashed box. The three extra resistors shown represent the lead wire resistance (shaded in gray). When measuring each RTD, you would source current with the IDACs into each RTD individually. The lead resistances reacting with the IDAC currents cancel assuming that the IDAC currents and the lead resistances match. The RTD sees the current from IDAC1. The reference voltage be RREF*(IIDAC1+IIDAC2). This becomes a ratiometric measurement where the ADC output code would be proportional to RRTD/(2*RREF)

    However, many people will not use this topology because noise may be a problem and input filtering may help. Using input filtering, only two 3-wire RTDs can be measured. This next figure shows how this topology is put together.

    With this topology, we inject the IDAC currents into the RTD with external pins, bypassing the filter. In the previous example, we injected current into the RTD from IDACs connected to the input pins. With an external filter, it's extremely unlikely that the filter resistances will match enough to directly cancel out the lead resistances.

    I would note there is an extra set of input pins that can be used to make an alternate measurement. If there is something else that needed to be measured, then AIN6 and AIN7 would be available.

    If you need the filtering and would still like to use a single device, one other possibility would be to use the ADS124S08. This device has more pins connected to the input multiplexer. You would connect the device shown below.

    Unlike the ADS1248, in the ADS124S08, there are no dedicated IDAC output pins (although on both devices, unused input pins can be used for the IDAC).

    I hope this answers your question. If you have additional questions, or if I've completely misunderstood your question, please feel free to post back.

    Joseph Wu

  • 0 in reply to Joseph Wu
    Intellectual 780 points

    Joseph Hi,

    Thank you for your detailed answer.

    You are correct it is PT100 (TPM is from another project related to security chips)

    Can I use an external switch, for the current sources, instead of the internal switch i.e. switching the IDAC1 and IDAC2 using the dedicated pin with an external switch?

    If yes any suggestion for a switch.

    Noam

  • 0 in reply to Noam Weidenfeld1
    TI__Guru** 100785 points
    Noam,


    I'm sure that you could use some sort of 2:6 multiplexer, but I'm well versed on which to use. I'll do some searching around and post back.

    In the meantime, let me describe what might be important in the recommendation. First, you shouldn't care much about what the mux series resistance is. Low resistance is great but in the end, the mux only routs the current, and series resistance won't be part of the measurement. The important thing is that the mux leakage must be very low. Any leakage takes away current from the RTD and reference resistance. This can cause an apparent gain error. Also, if you need to consider the operating temperature of the circuit. Often leakage currents are higher at high temperatures.

    There are two other things to note in the diagrams that I showed you. First, the accuracy of the reference resistor is important. Any error in this also appears as a gain error in the measurement. With the reference resistor, you will want a precision resistor with good accuracy and low drift. Second, this reference resistance also acts to level shift the input measurement. The PGA used in the ADC may not have a large input range of operation (and it may get smaller as the PGA gain is larger). The reference resistor is often used to shift the input signal up so that it is within the input range of the PGA.


    Joseph Wu
  • 0 in reply to Joseph Wu
    Intellectual 780 points

    OK, I will look for a switch while waiting for your suggestion.

    Another thing, what will happen if I will try to measure an unplugged RTD? in another words, how can I detect a problem in the RTD (shortcut or disconnection).

    Noam

  • 0 in reply to Noam Weidenfeld1
    TI__Guru** 100785 points

    Noam,

    The ADC may give unusual results when the RTD is removed. Looking at the first diagram, with IDAC1 going to AIN0 and IDAC2 going to AIN1, the injected current will force the inputs high, pulling both inputs to the positive rail. In this case, the input would be basically 0.

    At the same time, neither current from IDAC1 nor IDAC2 will reach RREF, so the equivalent reference voltage would also be 0. This is outside the range of operation for the ADC.

    If you do get this type of unusual result, you may need turn off the IDAC currents, and turn on the burnout current source. Using this technique, you would pull the inputs apart to test if the RTD is burned out or open. Remember that you would use this as a test case and that when not testing the inputs the burnout current sources should be turned off.

    Joseph Wu

  • 0 in reply to Joseph Wu
    Intellectual 780 points

    Josheph,

    Thank you very much for the support.

    Assuming I will use the ADS1248 for two RTD, will I get better results if I will use different reference inputs instead a common one for both?

    Noam

  • 0 in reply to Noam Weidenfeld1
    TI__Guru** 100785 points

    Noam,

    You should be able to get the same results. It shouldn't matter if you are making one measurement or two. The one issue is that there may be a need for extra settling time when swapping excitation for the different RTDs. If you have extra capacitance that requires settling in the measurement, you'll need to add some delay to wait for the RTD/RREF to settle.

    As for the analog mux/switch, I was pointed to this thread:


    It appears that the TS3A4751 may work, but note that the absolute maximum voltage is 4V. Also at an operating voltage of 3V, it seems that the leakage current is 2nA max. Normally I'd prefer that to be much less than 1nA, so you may be able to find something that would operate better.

    Joseph Wu

  • 0 in reply to Joseph Wu
    Intellectual 780 points

    So if my AVDD is 5V I can't use it?

    Noam

  • 0 in reply to Noam Weidenfeld1
    TI__Guru** 100785 points
    Noam,


    That is correct. The TS3A4571 has an absolute maximum of 4V for the supply. A supply greater than that will risk damage to the device. If you used this device, I'd stick with a 3.3V supply.


    Joseph Wu
  • 0 in reply to Joseph Wu
    Intellectual 780 points

    Joseph Hi,

    Thank you very much for your support.

    BR

    Noam

  • 0 in reply to Joseph Wu
    Intellectual 420 points
    Joseph

    In these shared R-ref multi-channel configurations, there is an error source introduced that does not exist in a high-side R-ref topology. With R-ref located at the low return end, it sees the summation of IDAC1 and IDAC2, whereas in a high-side topology, the current through R-ref is always identical to the RTD current and thus cancels any drift and accuracy error in the current source. The low-side approach needs very precise matching of the current sources; datasheet reveals a matching tolerance approaching 1% at 100uA and lower, unacceptable for measurements of greater than about 8-bit precision. Can you suggest a calibration method to measure and compute the current mismatch, which then can be applied to the RTD measurements? Seems to me it may be possible through a second measurement with IDAC2 OFF.
  • 0 in reply to Rick Hille
    TI__Guru** 100785 points

    Rick,


    There are two different things you can do, but as you've mentioned, both will require two measurements.

    In the first method, you can make make the measurement as normal, injecting IDAC1 into the first RTD lead, and injecting IDAC2 into the second RTD lead. This will give you one measurement. Then you would swap the IDACs, injecting IDAC1 into the second RTD lead, and IDAC2 into the first RTD lead.

    You should be able to work it out, but the average of the two measurements will cancel the IDAC1/IDAC2 mismatch.

    With the second method, you would use a single IDAC, but the same basic setup. You would inject IDAC1 into the first RTD lead, driving the lead resistance with the RTD and the reference resistor and take the measurement. Then you would inject IDAC1 into the second RTD lead, which drives the lead resistance and reference resistance. Note this would represent the negative of the lead resistace measurement, so you would add the first measurement to the second measurement to cancel them out.

    I would also note that since you are only using one IDAC, your reference will appear half as large as in the previous method. This may require a change in reference resistance, IDAC current, or PGA gain.


    Joseph Wu

  • 0 in reply to Joseph Wu
    Intellectual 420 points
    I like the IDAC swap method, which I completely overlooked. That (elegantly) maintains the exact same electrical conditions for each measurement. Can one assume then that the mismatch arises completely from the internal current sources themselves and not due in any part to leakage variations amongst the Mux IO channels?
  • 0 in reply to Rick Hille
    TI__Guru** 100785 points
    Rick,


    Yes, the mismatch is primarily doe to the internal current sources themselves. With chopping, I know that you can certainly reduce the amount of error associated with the IDAC mismatch.

    I don't have any specific knowledge about leakage with individual multiplexer channels. However, the normal leakage is small, and since the individual channels are constructed the same way, I image any residual mismatch would be much smaller.


    Joseph Wu
  • 0 in reply to Joseph Wu
    Intellectual 420 points
    Could we assume that the absolute and differential input current specifications for the AINs (+/-2nA, +/-1nA, PGA enabled) represent the maximum offset contributions of not just the IO pad structures but also the mux channels they feed? What I interpret from this specification is that a pair of ports can contribute as much as 4nA difference. At 10uA excitation, 0.04%, so yes, much smaller than the IDAC mismatch, but limited to about 11-bits accuracy. At 100uA, this jumps to 14bits, assuming the IDAC chopping cancels the mismatch by better than 1000:1. Am I correct with these numbers or missing something?

    Also, I notice our conversation here is not showing when I open the original thread. Where is this exchange accumulating? Should I open a new thread with suitable title to where this thread can be moved? I'm actually using the ADS124S08 but the same principles apply.
  • 0 in reply to Rick Hille
    TI__Guru** 100785 points
    Rick,


    I'd say your analysis is accurate. However, the numbers from Table 8 are typical values, not maximum values. I'm not familiar with the characterization data that generated this table, but I'd say these numbers could be a single standard deviation.

    In respect to the post position, you posted a reply to an existing thread and that thread already was on the third page. To see the posts, you'd need to scroll to the other pages, which can be done with the arrow buttons at the bottom of the page. Normally, we'd like people to post their own threads (and reference others if necessary by pasting a link into post if necessary). I can split conversation into its own separate post to help find our conversation easier.


    Joseph Wu