• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Subscribers 51 subscribers
  • Views 766 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.

OPA548: Loop stability analysis of current-out bridge drive

Chen Changlin
Prodigy 30 points
Part Number: OPA548

Hi,  Team

I'd like to use OPA548 to build a bridge Howland current source, the schemaitc is posted as below. In the application the load is a coil,  the series R8 and L1 is its equivalent circuit model, R8 is 5Ohm, and L1 is 150mH

Here are a few of my questions, 

1.  How to analyze the circuit's loop stability with  TINA-SPICE?

    This circuit is composed of two typical sub-circuits, improved howland current source and inverting amplifier, right now I just analyze the loop stability of these two sub-circuits separately, is this reasonable? is there any other way to analyze the loop stability for this ciurcuit?

2.  Loop stability simulation of Howland circuit

    The simulation circuit and result are shown as below,   the frequency of input signal is 1Hz in the application, it seems the result looks good, the phase margin is enough.

            

3. Loop stability simulation of inverting circuit

    For this inverting amplifier, the load is a resistor in series with an inductor, the loop simulation result looks really bad, the phase margin is negative 342 deg?  Is there something wrong with my simulation circuit? How to compensate this circuit to make it stable when the load is inductive?   I've tried parallel some capacitors or RC to the feedback resistor, but it didn't work out.

   But then it is weird that the small signal transient response simulation result looks good, referring to the right bottom waveform, the input signal is a 1Hz/10mV pulse signal,   the signals in the screenshot are output current, output voltage and input voltage from top to bottom.  It complicts with the loop stability analysis, it seems the circuit is stable from the transient response simulation,   I can't figure out why, can you help with this?

            

And finally, the relevant TINA-SPICE schamatic are attached.

 loop stability simulaiton of OPA548 howland circuit.TSCCurrent-out Bridge drive with OPA548.TSCloop stability analysis with OPA548 inverting amplifier.TSC

Thanks for your help.

Best wishes.

Charlie Chen

  • 0
    TI__Guru* 93105 points

    Hello Charlie,

    Let me take a deeper look at your OPA548 inverting amplifier stage and the odd, highly negative phase margin you are getting. Let's see if I get the same results or something different. I am reasonably certain that the stability analysis needs to be accomplished with both amplifier stages in the circuit at the same time so more analysis is probably required.

    We are coming up on the holidays here in the US and we are short staffed right now. I am not exactly sure when we will be able to have more of a response for you.

    Regards, Thomas

    Precision Amplifiers 

  • 0 in reply to Thomas Kuehl
    Prodigy 30 points

    Hi, Thomas

    No rush, you can help me with this after the holiday.

    Merry Christmas and happy new year.

    Charlie

  • +1 in reply to Chen Changlin
    TI__Mastermind 49435 points

    HI Charlie,

    I performed the stability analysis.  When breaking the loop, I ensure to leave both amplifiers connected in the circuit and analyze the phase margin on each one of them, breaking the loop on each, one at at time.  The first stage (Howland circuit) is stable at 89 degrees of phase margin. The second inverting stage was marginal with a lower phase margin around ~30-degrees of phase margin.   Therefore, I added a feedback capacitor on the feedback of the second stage to improve stability to ~45-degrees of phase margin. 

    Please see below the simulation files. 

    The first stage of Howland Circuit, with 88-degrees of phase margin:

    The modified inverting stage, adding a small 39pF feedback capacitor, with 44-degrees of phase margin:

    TINA files:

    6354.check_OPA548_modified.zip

    Thank you and Regards,

    Luis

  • +1
    Guru 140200 points

    Hi Charlie,

    the phase margin of the second OPAmp is a bit too small. 26° phase margin can result in instability:

    chen_opa548_1.TSC

    I would suppose the following modifications:

    The first OPAmp, on the other hand, is stable and unproblematic:

    chen_opa548.TSC

    I also simulated the effect of interwinding capacitance of the coil. As the simulations show, the stability is not impaired by any interwinding capacitance Relaxed

    Kai

  • +1 in reply to kai klaas69
    TI__Mastermind 49435 points

    Thank you Kai,

    Charlie, please let us know if you have additional questions.

    Best Regards,

    Luis

  • +1 in reply to Luis Chioye
    Prodigy 30 points

    Hi, Luis and Kai

    Really appreciate your help. You help me resolve the instability issue and  I also learned how to break the ciruit to simulate the loop stability.

    But from your simulation, I have one more question about how to read the phase margin.

    From Luis's simuluation result, I noticed that Y-axis of phase if from zero to -180,  it might be a lagging phase, so the phase margin is calculated as "lagged phase - (-180).  But from Kai's simulation result, the Y-axis scale of phase is from 180 to 0 deg, it is the phase margin, so the phase margin can be read directly. 

    This is confusing, why there are two different simulation results for the same simulation method, what I mean is the Y-axis scale. Maybe it has something with TINA-SPICE itselt?

    Thanks.

  • +1 in reply to Chen Changlin
    Guru 140200 points

    Hi Chen,

    Luis has merely reversed the voltmeter Relaxed

    Kai

  • +1 in reply to kai klaas69
    TI__Mastermind 49435 points

    HI Chen,

    I don't believe the meter polarity is an issue.

    The general amplifier stability criteria, when looking at phase margin and breaking the loop, the key is to monitor for the phase shift change or delta in phase in the loop-gain phase plot, and ensure that this change in phase across frequency is less than < 180 degrees.  In other words, you need to plot the loop gain phase starting at low frequencies, and calculate the delta or phase shift change from low frequencies, to the frequency (fcl) where Aol*Beta (Loop-Gain) = 0-dB.   The reality is that depending on the circuit configuration, you can have a phase plot starting at +180 degrees or 0-degrees depending on the circuit. However, this is not an issue if you follow proper procedure and monitor the phase shift change across frequency.

    In my example, the loop-gain phase starts at 100milliHerz (low frequency, close to DC) at 0-degrees, and changes by negative - 91 degrees to the frequency (fcl) where Aol*Beta (Loop-Gain) = 0-dB.  The total change in phase is 91 degrees, and therefore the phase margin is calculated as 180 - 91-degrees = 89-degrees. Of course, the starting phase polarity changes if you reverse the meter. Nevertheless, if you monitor the delta change in phase from low frequencies to the fcl frequency, you can easily calculate the correct result for phase margin.

     See below the stability criteria definition on Tim Green's presentation, where they refer to change in phase or phase shift:

    Regards,

    Luis

  • 0 in reply to Luis Chioye
    Prodigy 30 points

    Hi, Luis and Kai

    I totally understand now, thanks for your help. My question is resovled successfully.

    Best wishes.

    Charlie