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.

XTR111 additional filter

Other Parts Discussed in Thread: INA330, XTR111, OPA2188, REF5050

Hello,

My application concerns exciting and reading temperatures from 12 different NTC's, converting readings to 16 bit data and sending it to 28335 DSP.

Thermistor have +-0.05ºC accuracy spec, global accuracy should be within +-0.1ºC. It is a tight spec and requires careful planning and design.

I'll split the application circuit in different threads, so I can detail what I need and organize it better.

Signal first comes from previous circuits (not shown): NTC Therm -> INA330 -> LPF @ 2Hz -> XTR111

There are independent circuitry and PCB's for each of the 12 NTC's , INA330, LPF and XTR111 setup.

INA330 is supplied  by LP2989AIM-5.0 (+5V), which in turn is powered by a +12V DC external isolated supply and all circuitry of all PCB's share the same (isolated) ground. LPF is an OPA2188, tuned to 2Hz, 2 poles, powered directly by the +12V DC external isolated supply.

Each of these PCB's are separated by about 50 cm, and are disposed in a noisy environment, being physically close to ~4500 W SMPS.

Signal from each of the 12 the LPF's need to be transmitted to a MUX, next comes ADC.

The distance from each of the PCB's and the MUX is ~ 1 meter. Although it may be considered a short distance for current transmission, we have chosen 4-20 mA transmission, so there should be no surprises, particularly about EMI and noise.

For the 4-20 mA transmission, we need a 3 wire system, since MUX is single ended, loaded with a precision grounded 250 ohms low ppm resistor (for plain simple current to voltage conversion, back to +1 to +5V) and share the same ground as all circuits so far. We have decided for XTR111.

So attached there is our setup for XTR111. XTR Regulator is not used. +5VRef comes from REF5050. "V_LPF to XTR" is ths signal from the LPF, ranging from 0 to +5VDC.



I did not understand the "additional filter" suggested in XTR111 datasheet, pg 16, figure 41, shown below. It was poorly explained and circuit is somewhat confusing.


Is the 500 ohms resistor in the circuit supposed to replace the R6 and R7 (15 ohms each) found in the gate controlled current limiter, also suggested in datasheet, pg 14, figure 37a (below) ??



I suppose the only current path would be through the 500 ohms grounded resistor, supposedly before the load. If that is supposed to work like that (correct me if this is not the case), I can't afford it, since I am working with a 12V supply. This 500 ohm + the 250 ohm load would need 15V @ 20 mA, which I don't have.

Otherwise, is the 500 ohms the load itself??? It should be indicated in the circuit...

Why there is a left open connection at the right of the circuit, after the 10nF filter capacitor ? Is it supposed to connect anywhere??

Another question: I am connecting the output of the resistive divider (between +5VRef and "V_LPF to XTR") directly to VIN. By doing this, I can't use a resistor in series with VIN matching RSET and cancelling bias current. I thought of buffering resistor divider output (+1 to +5V) and connecting it to a 2k5 resistor before VIN. For that, I'd need a new op amp in the circuit. Is it really worth?? What is the actual gain by doing it??

Please check the rest of the circuit, I have never worked with XTR111 before and we are a bit short in time.


One more thing: I can't find any MUX in TI portfolio that works comfortably with 20 mA load, Vsupply >= 12V, single ended, 1:4 or 1:8 channels. I am missing something or TI really doesn't offer it?

I am considering ADG1608, it seems to do the job. It would be, however, the only part that is not TI... :(

EDIT: Also please confirm that EF and REGS pull up can be tied to +5V, instead of +12V. It is easier to interface directly with the MUX at +5V.

  • Hello Mr. Moraes,

    Thank you for providing so much information about your application, it makes it much easier for us to answer your questions.

    Your choice for a current-loop based data transmission makes sense in a noisy environment, but beware that current loops will reduce, but not completely eliminate, sensitivities to EMI/RFI.  The XTR111 is a good choice for this application and is very accurate, durable, and easy to use.

    I agree with many of your datasheet suggested improvements and I will suggest them if we update the datasheet in the future.  

    Although it's not labeled, the 500 Ohm resistor is the system load resistor in Figure 41.  Using 250 Ohms as you've chosen will not affect anything.  Perhaps they didn't want to confuse subjects, but these components are in addition to the recommended 10nF (C3)+ 15 Ohms (R7) shown in the output protection section and Figure 37.  

    In your system, the current will flow out the BJT, through the 15 Ohm output protection resistor, then through the 250 Ohm load resistor to GND forming the desired 1 - 5V output for the ADC.  Rf and Cf form a simple low-pass filter on the 1 - 5V output that you could set as you desire.  

    While bias current matching is not a concern in most newer linear circuits (see blogs below for fun reads) it does appear to be single-ended and flowing into the inputs of the XTR111.  The parallel combination of the input divider results in 10k || 40k = 8k of source impedance at the non-inverting input while the inverting input sees RSET = 2.5k  So you have at most a 8k - 2.5k = 5.5k deviation in input resistance multiplied by the maximum input bias current of 25nA = 137.5 uV, which is still only 1/10th of the maximum input offset voltage of 1.5 mV.  If you believe that to be a concern then you may want to feature the additional components you've mentioned to try to cancel the input bias current related offset more effectively.  

    Sadly I do believe you are correct about the current MUX selection in our portfolio and you may need to consider another solution.  Vishay also offers some very high performance analog MUX devices that you may want to take a look at for your application.

    Finally, yes it is okay (and recommended as shown in the XTR111EVM) to pull-up the digital outputs to +5V.  Here's the XTR111EVM User's Guide if you'd like to see the schematic which does also feature the output filter you had questions about.  http://www.ti.com/lit/ug/sbou048c/sbou048c.pdf

    Bruce Trump Blogs on Input Bias Current:

    Input Bias Current of CMOS and JFET amplifiers:
    http://e2e.ti.com/blogs_/archives/b/thesignal/archive/2012/11/08/input-bias-current-of-cmos-and-jfet-amplifiers.aspx

    Bias Current Cancellation Resistors, Do You Really Need Them?
    http://e2e.ti.com/blogs_/archives/b/thesignal/archive/2012/04/11/input-bias-current-cancelation-resistors-do-you-really-need-them.aspx

    Internal Input Bias Cancellation:
    http://e2e.ti.com/blogs_/archives/b/thesignal/archive/2012/04/17/internal-input-bias-current-cancellation.aspx

    Temperature Effects on Input Bias Current:
    http://e2e.ti.com/blogs_/archives/b/thesignal/archive/2012/11/14/temperature-effects-on-input-bias-current-plus-a-random-quiz.aspx

  • Collin,

    It is me who has to thank you very much for your answer!!

    Really helpful!!

    Best regards,

    Thiago