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.

How meansure Loop Gain for composite noninverting amp in TinaTI

Other Parts Discussed in Thread: THS6012, OPA627, THS6022, LM4562

I can measure the Loop Gain here for such a scheme:

http://imgbox.com/lTHerVv3

But how to measure the Loop Gain such a scheme, but here are two opamp and two feedback:

http://imgbox.com/dkzd8Fm5

  • Hello Vitalij,

    My colleague, Tim Green, and I analyzed your composite amplifier circuit for stability. The process is fairly involved. Each amplifier is analyzed individually. The first amplifier is a more straightforward analysis because it has a single feedback path. The second amplifier's analysis is more complex due to the dual feedback paths.

    Attached you will find a .zip file containing TINA analysis for each of these stages and their resulting loop-gain and phase plots. The AD device model's closed-loop output impedance did not match the data sheet when tested. Adding a 100 Ohm resistor at the model's output and within the feedback loop brought it much closer to the data sheet graph. The THS6012 model's closed output impedance checked okay to the graph.

    The first stage phase margin appears to be about 69 degrees, while the second stage is about 26 degrees. That latter number may be a concern because it indicates the stage is marginally stable. The .TDR file provides the gain/phase plots for the analyses. The phase margin is attained by using TINA's markers to determine where loop gain goes to 0 dB. You start with the very low frequency phase and then find the phase at 0 dB; subtract the two. Then, subtract that difference from 180 degrees. The result is the phase margin.

    If you want to know more about the process for analyzing stability for an amplifier using two feedback loops, I've included a PowerPoint that Tim had prepared on the subject. It is enclosed in the .zip file.

    Regards, Thomas

    PA - Linear Applications Engineering

    Regards,

    THS6012.zip
  • Big thanks to Thomas, you have published some interesting information.
    But I have a few questions.
    Here's a way to breaking the op amp feedback loop and measure the Loop Gain whole amplifier FOR INVERTING AMP (U1 + U2):

    Just an example we see at 10 Hz loop gain ~ 177db, is the result of the U1 and U2 together (and their feedback too).

    --

    I need to measure the loop gain for the noninverting configuration is the same amplifier circuit.

    In the published you THS6012_stability_04.TSC pointer Vloop is the loop gain the whole amplifier (U1 + U2 + their feedbacks)?

    VOA is AVOL of U2 in THS6012_stability_04.TSC?

     

     

  • Hello Vitalij,

    I find the compound test circuit you posted interesting. Yesterday, I tried something very similar but wasn't sure if the circuit idea was correct; however, I came up with almost exactly the same results you attained. You can see the circuit and gain-phase plot below. The phase margin is 59 degrees at 28.9 MHz (0 dB). Note if your add the recommended 100 resistor in the AD amplifier's output that the phase margin becomes considerably less. You may want to try including it in your test circuit. 

    To answer your questions:

    In the published you THS6012_stability_04.TSC pointer Vloop is the loop gain the whole amplifier (U1 + U2 + their feedbacks)? Yes

    VOA is AVOL of U2 in THS6012_stability_04.TSC? Yes

    Regards, Thomas

    PA - Linear Applications Engineering

  • Thank you Thomas.

    I still interested in a method of measuring the loop gain non-inverting amplifier with the input circuit:

    --

    For example, for an inverting amplifier input circuits affect the loop gain:

    -------------------------------------------------------------------------------------------------------------

    ----------------------------------------------------------------------------------------------------------------

    Maybe you even answer of how well the filter (rezizstor + ferrite bead) at the amplifier output for headphones (netlist for BLM31PG601SN I downloaded from Murata simsurfing):

  • Hello Vitalij,

    I will have to consult with one of my colleagues who is presently on vacation until next week regarding the gain/phase for the non-inverting compound amplifier. I am not nearly as expert on this subject as he is.

    Regarding the Murata SMD inductor; it exhibits an impedance Z of about 600 Ohms at a peak frequency a little over 100 MHz. Even though it is an inductor and one would suspect that the inductive reactance would be the dominant component of the impedance Z = (R2 + jX2)1/2 for the these particular inductors the R term is the dominant factor. It is not a dc R, but instead a loss resistance presented by the ferrite material to an ac signal in the specific frequency range.

    Since the inductor's impedance peaks around 100 MHz that is where it will provide the most opposition to ac current flow. That falls off above and below 100 MHz. Therefore, it is only effective as an RF "choke" in that frequency range.

    Regards, Thomas

    PA - Linear Applications Engineering

  • Thank you Thomas.
    On one forum I found an example of the amplifier circuit:


    As you can see a diagram of closed-loop Bode feedback. Is it with the help of this you can evaluate the stability of the amplifier? As I understand the resistance should be evaluated only when the open loop feedback?

  • Hello Vitalij,

    This is fine for a closed-loop analysis. I want to consult with my colleague when he returns next week so that we correctly set up the non-inverting compound amplifier circuit for the stability analysis. I will let you know when we get that underway.

    Regards, Thomas

    PA - Linear Applications Engineering

  • thanks for your time, I will wait for a response

  • Hello Vitalij,

    I consulted with my colleague regarding methods for analyzing stability for of the composite amplifier when the non-inverting input of the first stage amplifier is driven. He said that two analyses are necessary to determine if the amplifier is stable. Here is what he said to do:

    Evaluate the first amplifier stage (OPA627)

    • Break the loop going to the OPA627 inverting input by inserting a 1 T (terra) Henry inductor between the input pin and the R4/R5 node
    • Add a 1 T Farad capacitor between the non-inverting input and an input ac signal.
    • Add a voltmeter at the OPA627 output
    • Perform an ac sweep and obtain the loop gain and phase

    Evaluate the two stages together

    • Use the dual input drive setup included in THS6012.zip. Specifically, the THS6012_stability_04.TSC file
    • Remove the voltmeter at the OPA627 and place one at the THS6022 output
    • Perform an ac sweep and obtain the loop gain and phase

    Regards, Thomas

    PA - Linear Applications Engineering

  • """Evaluate the first amplifier stage (OPA627)""" - this is the stage to me is clear.

     

    """Evaluate the two stages together""" - this stage I do not understand.Should be so : ???

    ---------------------------------------------------------------------------------------------------

    ---------------------------------------------------------------------------------------------------

    ---------------------------------------------------------------------------------------------------

    Maybe you will be interested. I changed my headphones-amplifier circuit:

    In this scheme, a little less than the loop gain (100dB at 20kHz though clearly not enough), while it is absolutely stable. I am confused whether it is possible in this way to set the volume control (potentiometer 10K) to the input of an inverting amplifier? In the simulation phase margin is not changed and the DC voltage is always less than 1mV, everything seems in order.

  • Hello Vitalij,

    When I check the loop gain and phase for the latest circuit you proposed my analysis shows that it is unstable . The first stage has plenty of phase margin > 70 degrees, but when combined with the remainder of the circuit the phase margin is -12 degrees. There is gain at 0 degrees indicating the circuit should oscillate.

    Note that I used the LM4562 simulation model available from the TI website so I am not sure if it is the same as the one you used. It may be that only the symbols differ.

    You can find the analyses in the .zip file.

    Regards, Thomas

    PA - Linear Applications Engineering

    THS6012_B.zip
  • I use the model LM4562 site TI, just with this symbol it is more convenient to use.

    ------------------------------------------------------------------------------------------------------------------------------------------------

    Thomas I thank you for your time, but let's be clear yet.

    Please publish a method (as screenshot or archive) finding loop gain of such a NONinverting scheme:

     ---------------------------------------------------------------------------------------------------------------------------------------------------

    About my new scheme inverting headphones-amplifier I do not agree with you.

    Its phase margin is about 80 degrees, I attached the file with the simulation for this post.

    3678.composite THS6012 Shaq888 OPEN Loop.zip

  • Hello Vitalij,

    I am using the methods outlined in the PowerPoint prepared by my colleague Tim Green. I have analyzed the most recent OPA627/THS6022 composite amplifier and received the results seen in the image shown below. When the loop gain is 0 dB, the phase margin is just about 0 degrees. I have attached my TINA circuit for your review.

    You indicated that you are observing a phase margin of about 80 degrees. We have a technical disconnect occurring and are receiving different results. Can you please post the TINA schematic for the specific circuit you have shown above so that I may see how you went about analyzing the loop gain and phase and received the 80 degree result.

    Regards, Thomas

    PA - Linear Applications Engineering

     

    THS6012_stability_08.TSC
  • Hi Vitalij,

    We are reexamining our loop gain phase analysis technique used for the composite amplifier configuration. Please send us the test circuit schematic you used for the composite amplifier analysis. Please send the TINA file too.

    Regards, Thomas

    PA - Linear Applications Engineering

     

  • Thomas, let's take two different amplifier circuits and shall understand.


    The first scheme is an inverting amplifier ---- This circuit has a phase margin of about 80 degrees.
    The second scheme is an non-inverting ---- I want to know how to measure the phase margin of this circuit.


    Amplifier on the first scheme has already been assembled and checked my friends wideband oscilloscope - no oscillations unambiguously.
    Amplifier on the second scheme I built just for example, to find out how to measure the loop gain for the non-inverting amplifier.

    1411.The first scheme (inverting amplifier).TSC

    5342.The second scheme (non-inverting).TSC

  • Attached you will find a detailed PowerPoint of breaking the loop to check for loop gain phase margin for each of the circuits you posted in the last post.  In these composite amplifiers it is always best to check the Loop Stability of each amplifier separately.  They must both be stable and with good phase margin of greater than 45 degrees to account for manufacturing tolerances of op amps and external component tolerances. For any engineering problem there is more than one way to analyze the situation.  I prefer Loop Gain and  1/Beta on Aol approach to op amp stability problems.  This technique is best documented by Gerald Graeme and in the 4 part tutorial available at the link below:

    http://e2e.ti.com/support/amplifiers/precision_amplifiers/w/design_notes/2645.solving-op-amp-stability-issues.aspx

    Op Amp Stability Scheme 1 and 2.ppt
  • Tim Green1 and Thomas Kuehl - thank you very much for your time and explanations, I found answers to their questions. The result is a great headphone amplifier using op-amp Texas Instruments.