• State Resolved
  • Locked Locked
  • Replies 12 replies
  • Answers 1 answer
  • Subscribers 40 subscribers
  • Views 5324 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

TINA/Spice/OPA627-DIE: OPA627

Jubayer Jalil
Prodigy 200 points

Part Number: OPA627-DIE

Tool/software: TINA-TI or Spice Models

Hi,

I have simulated the attached circuit (OPA627 in inverting mode), and simulation result shows the output noise of 61 nV/rt(Hz) at 10KHz. By hand calculation, I found 64.72 nV/rt(Hz) at 10KHz without considering OPA627 device/model input capacitance of 15pF (differential i/p cap. + CM i/p Cap from datasheet).  Now, for this configuration, if I take into account OPA627 device/model input capacitance of 15pF in addition with two external capacitors of the given circuit, my hand calculation shows 68.4 nV/rt(Hz) at 10KHz exhibiting significant difference from previously estimated hand calculation of 64.72 nV/rt(Hz) at 10KHz.

Now my question is:

1. Why does the hand calculation of 64.72 nV/rt(Hz) at 10KHz show  the good agreement with the simulation result of 61 nV/rt(Hz) at 10KHz, even though OPA627 device/model input capacitance of 15pF (differential i/p cap. + CM i/p Cap) is disregarded in the calculation?

Kind regards

Jubayer

6470.Inverting mode.TSC

  • 0
    TI__Mastermind 38065 points

    Jubayer,

    Without knowing the details of the hand analysis, it is hard to say why it does not agree with the TINA noise sims.
    In a past E2E post we confirmed the OPA627 model and TINA simulator give noise sim results which agree with a hand analysis.
    To review this discussion, please see that E2E thread:

    e2e.ti.com/.../2165030

    Regards,
    John

  • 0 in reply to John Miller - Sensing
    Prodigy 200 points

    Dear Mr. John,

    I have attached the file of the circuit and the equation of the output noise voltage spectral density, which is used for hand caluclation of OPA627. 

    Interestingly, if I consider CM capaciatnce (7pf) and differential capaciatcne (8 pf) for output noise voltage spectral density calculation of the attached circuit, it does not agree with the simulation result.

    Kind regards

    Jubayer

    Vout.docx

  • 0 in reply to Jubayer Jalil
    TI__Mastermind 38065 points
    Jubayer,
    How do the results of the calculation compare to the sim results if you leave out the differential capacitance and use just the single-ended capacitance?

    Regards,
    John
  • 0 in reply to John Miller - Sensing
    Prodigy 200 points

    Dear Mr. John,

    If I consider op amp's input capacitance of 15pF (differential i/p cap. + CM i/p Cap from the datasheet) for output noise density calucaltion, it exhibits  68.4 nV/rt(Hz) at 10KHz. Whereas TINA simulation result shows the output noise of 61 nV/rt(Hz) at 10KHz. Significant difference of hand and simulation  results due to considering input capacitance of 15pF. My question is why this discrepency arises for inverting mode?

    regards

    Jubayer

  • 0 in reply to Jubayer Jalil
    TI__Mastermind 32570 points

    Jubayer,

    In the word document you sent, the OPAMP is treated as a VCVS that:

    1, is noiseless

    2, has a gain of infinity.

    Neither condition is true with the actual OPA627 model used in your TINA circuit. As a result, the numbers from your calculation with those 2 assumptions would not be identical.

    If you include the internal circuit of the OPA627 model you use and redo your hand calculation, you will get a 100% match.

    Regards,
    JC

  • 0 in reply to JC Zhu
    Prodigy 200 points
    Dear Mr. ZC Zhu,

    You are correct. However, I have used OPA627E model for my simulation.
    I have mentioned earlier that considering input capacitance of 7pF for noninverting mode of OPA627E model agrees well with TINA simulation and hand calculation. But, considering input capacitance of 15pF for inverting mode of OPA627E model does not agree well with TINA simulation and hand calculation. Interestingly, if I disregard, this 15 pF for hand calculation, TINA sinulation and hand calculation agrees well.
    Regards
    Jubayer
  • 0 in reply to Jubayer Jalil
    TI__Mastermind 38065 points
    Jubayer,
    Where did your equation come from?
    John
  • 0 in reply to John Miller - Sensing
    Prodigy 200 points

    Hi Mr John,

    I have atthached the technical paper of TI from which I have devised the equation.

    Kind regards

    Jubayersignal-conditioning-sensors-1316043200.pdf

  • 0 in reply to Jubayer Jalil
    TI__Mastermind 38065 points
    Thank you Jubayer.
    I did a derivation and got the same equation.

    TINA sims of the inverting circuit gives an output noise of 50.65nV/rt(Hz) at 100kHz.

    In the analysis, if you make Cin equal to the sum of Cc, CA and the op-amp input capacitance (Cin = Cc +CA+ Copamp = 2pF + 2pF + 15pF = 19pF), the calculated output noise is 47.4nV/rt(Hz). So the calculated and simulated results are within a few percent of one-another.

    Regards,
    John
  • 0 in reply to John Miller - Sensing
    Prodigy 200 points

    Dear Mr. John,

    Thanks for your cooperation in this regard.

    My calculated result is 51.03 nV/rt(Hz) at 100kHz.

    I have considered 2.5 fA/rt(Hz) [max. value from the datasheet] as a current noise density.

    Now my questions are:

    1. How did you achieve low output noise in hand caluclation ?

    2. What should consider for current noise density : 2.5 fA/rt(Hz) or 1.6 fA/rt(Hz) (typ) calculating output noise??

    3. How to generate current noise density curve in TINA for OPA 627?

    Look forward to hearing from you soon.

    Regards

    Jubayer

  • 0 in reply to Jubayer Jalil
    TI__Mastermind 38065 points

    Jubayer,

    For my original calculation I used the Typical value for voltage noise density (4.5 nV/rt(Hz)).
    If the max value is used, the calculated output noise rises to 50.5 NV/rt(Hz).

    for this type of device with FET inputs, the current noise is so low that it probably won't matter if you use 1.6 or 2.5 fA/rt(Hz). For the circuit you have shown, the impact of current noise is several orders of magnitude less than the voltage noise. That is why these types of devices are popular for sensor interfaces.
    If you must include the current noise, then the value to use depends on the goal of the analysis. If you want a result that represents a worst-case, then use the max value. If you want a result that represents an "average" behavior, then use the Typical value.

    I tried generating a believable current noise curve for this model and have had no luck. The text header in the model netlist says the model was revised in the past to support this feature, but simulations don't seem to agree with that. The circuit is uploaded to this thread.

    Regards,
    John

    OPA627_Noise analysis_revised3.TSC

  • 0 in reply to John Miller - Sensing
    Prodigy 200 points

    Dear Mr. John,

    Thank you very much for your extreme support in this regard.

    Kind regards

    Jubayer