• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 50 subscribers
  • Views 803 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.

OPA2188 weird overshoot curve

Thiago Finotti
Intellectual 580 points
Other Parts Discussed in Thread: OPA2188, INA330

Hello,

I am running a simple transient test for checking overshoot and stability with OPA2188 driving a capacitive load.


The reason for that is OPA2188 will be buffering a 1.25 V as input reference for INA330, and datasheet of INA330 asks for a 0.1uF cap.


So here is the circuit:

Now, when I use a load (R1) greater than 250Mohms, curve goes ok, like this:

But if I use a load lower than 200Mohms, down to 50Mohms (give or take) a weird dip happens:

At 1Mohms curve goes back to "normal" overshoot.

Other input voltages also made output curve switch like this.

What am I missing??

  • 0
    Expert 1220 points

    I am not a SPICE expert, but in both instances your Y-axis scale records the same number. This tells me that the results the SPICE engine is calculating is simply the limit of the linear matrix.

    Have you tried a different integration method? I think you can choose Gear or Trapezoid.

    You can also try selecting 'Zero Initial Values'.

    In either case, I think there is nothing to worry about and it is not a real over or undershoot.

    -Ken

  • 0 in reply to Ken Dillinger
    Intellectual 580 points

    Hello Ken,

    Thank you for the reply!

    Great suggestion, I've tried Gear and Trapezoid, up to 5th integration order.

    Also have tried zero initial values and use initial conditions.

    What really changes is the sharpness of the curve, but the curve stays nearly the same.

    It agree with you it is probably nothing to worry, no real world, but I could only know that because it is a very simple circuit and its response is very predictable.

    But I couldn't find where this oddity came from... and I'd like to, it might be useful in a more complicated circuit!

  • 0 in reply to Thiago Finotti
    Expert 1220 points

    For future circuits, you can always break the loop and check the rate of closure of the closed loop response to that of the open loop response. 20dB/dec is what you would want to strive for in a design.

    It certainly has "a lot" of peaking, but it settles well with no ringing. If anything, maybe a complex pole-zero, zero-pole pair in the loop response. That is just a guess on my part. :-(

    -Ken

  • 0 in reply to Ken Dillinger
    Intellectual 580 points

    Agree, Ken!


    Thank you!