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.

XTR116: Capacitive loading and stable operation

Part Number: XTR116

I have built and deployed many boards out into the field using the XTR116 in an isolated configuration.

This design has always worked well and is stable on the bench, but now I being told that the the 4-20 output occasionally "drifts" up to the full 20mA rail.
I cannot replicated the stated problem on the bench.
Only now am I realizing the significance of drawing 5 on page 8.
In reading the document there was no discussion on the why, when and if of this drawing.
What is un-stable operation?
Why are there two capacitors?
What are the operational differences of the two drawings?   
I might be able to "hack" the board and use a 10 ohm resistor on JMPR4.  This is very close to the top drawing in figure 5.
Should I increase or decrease the capacitance of a or all capacitors?
Should I remove a capacitor so that I only have two?
With a more serious "hack" I could separate the I2C rail from the MCP4725 IC and just use the VREF pin for that.
That would leave me with only one 100nF cap on the VREF pin.  Would I need to add another?
If I separate the two I2C and VDD rails, and  remove the capacitor, does that mean I don't need the 10 ohm resistor
Thoughts and comments are urgently needed.


  • HI John,

    Essentially the XTR116 voltage reference requires the specified load/compensation capacitance and compensation resistors shown on Figure 5 for reliable or stable operation.  If the reference does not have the proper compensation, the reference circuit will not have enough phase margin to guarantee stability. Without the compensation, the reference can easily became unstable exhibiting behavior such as oscillations, or the reference voltage could become unpredictable, or drift to the supply rails. 

    When analog circuits are only marginally stable, it is not uncommon to see devices manifest the instability as temperature or voltage conditions change, or due to the expected subtle variation across devices and across different lots; or the issue could occur over the life of the product.  

    There are two different options for compensation as shown on Figure 5.

    Option (A) requires CLF capacitor in the range of 2.2uF to 22uF.  The RISO 10-Ohm resistor is required as well.  The CHF can be the rest of the bypass capacitance required on your circuit; and this capacitance can be anywhere in the range from 10pF to 500nF.   For example, if you choose option A, you will need to add CLF (at least 2.2uF) and RISO (10-Ohm) as shown below:


    Option (B) requires CLF capacitor in the range of 2.2uF to 22uF.  The RCOMP 50-Ohm resistor is required as well.  Same as above, the CHF can be the rest of the bypass capacitance required on your circuit.


    In summary, one possible modification is to use compensation option (A) in your circuit.

    You could leave the C44, C48 unmodified, and add the compensation option (A) with CLF (2.2uF capacitor replacing C49) and add RISO 10-Ohm on JMPR4 as shown below.


    Thank you and Kind Regards,


  • Louis,
    Thank you for verifying my assumptions.  You have hopefully save a lot of field people much grief.
    Last question(s):
    The Option A drawing has Clf with a range of 2.2 - 22uF.  
    Would a value higher than 2.2uf be better or have a greater chance of stability?
    What if any negative effect does a larger cap have on the operation of the circuit?
    Thank you.

  • HI Kovach,

    As long as the compensation capacitance (with the recommended Rcomp or Riso) is inside the range 2.2uF to 22uF, the circuit offers a conservative phase margin, and circuit is stable per the datasheet recommendation.

    One effect of having a larger capacitor is longer settling time after power-up; and possibly a larger capacitor footprint may be required for the larger value capacitor.  However as long as you use the components inside the recommended range on Figure 5, the circuit will still be reliable.

    Thank you and Regards,


  • Hi John,

    Many Thanks,

    Kind Regards,