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/OPA695: How does the simulation result of the OPA695 differ from the parameters given in the datasheet?

Part Number: OPA695
Other Parts Discussed in Thread: TINA-TI,

Tool/software: TINA-TI or Spice Models

OPA695.TSC

1. Simulation software result: bandwidth is195MHz

2. The parameters given in the datasheet: bandwidth 350MHz

so why ?

  • Hello user5346877,

    I've contacted the modeling team and will let you know once we have more details.

    Best,
    Hasan Babiker
  • Hi,

    the bandwidth of such OPAmps extremely depends on load capacitance. Even small stray capacitance at the output of OPAmp can have a huge impact on bandwidth. 20pF load capacitance will extend the bandwidth from 208MHz to 514MHz!

    695.TSC

    So, I guess the 350MHz bandwidth specification of datasheet takes into account some pF of load capacitance.

    Kai

  • Thanks for your help。

  • HI Kai:
    I have simulated for 350MHZ bandwidth, it need about 15pF of load capacitance. Bandwidth testing under the conditions of 15PF load does not seem to have any meaning. 

    thanks!

  • Hi Kai,
    I noticed the bandwidth at G=16 is not 100% tested, it's a typical value only for information.

    When we add a capacitive load to a voltage feedback amplifier. We could modify the Aol curve to analyse the circuit. The capacitive load added a pole to Aol curve, it will reduce the bandwidth of the whole circuit, and it will cause overshoot and reduce the stability.
    However, I'm not sure whether it will add a zero to current feedback amplifier. Or may be it will modify the Z(s) curve. The bandwidth depended the Z(s)/(Rf+Ri*NG).

    Waiting for Hasan's response.

    Regards,
    Ricardo
  • Hi Ricardo,

    a capacitive load will definitely destabilize the OPA695:

    The ringing of the step respsonse has to do with the resonance at arround 430MHz which is formed by the 20pF load capacitance:

    Adding a load capacitance should be avoided, of course. But sometimes, the unavoidable load capacitance resulting from the layout or something else can unwantedly increase the bandwidth a bit.

    Kai

  • Hi Kai,

    Though adding a capacitive load increased the bandwidth. However, it sacrificed the phase margin.

    When we analyse a voltage feedback amplifier.Vout/Vin=Aol/(1+Aolβ), the loop gain is Aolβ. When the Aol(db)+(1/β)(db)=0(db),and phase shift =180°, the amplifier instability. Desired phase margin >=45°.

    We can also using this method to analyse current feedback amplifier. Vout/Vin=α*NG/(1+(Rf+Ri*NG)/Z(s)), the loop gain equal to Z(s)/(Rf+Ri*NG). You can find it in chapter 9.1.5.1 in datasheet. We recommend setting Rf+Ri*NG=663, it's a constant without phase shift. So the phase margin depended on  Z(s).

    In this situation. We still have more than 45° phase margin. The amplifier is stable.

    However, it seems that we couldn't calculate the phase margin from Gain and Phase plot. In this example, there is no phase margin left.

    I suspect the capacitive load modified the Z(s) curve. But I have no exact evidence.

    Do you have any idea about how to analyse the stability of current feedback amplifier?

    Regards,

    Ricardo

  • Hi Ricardo,

    let's wait for Hasan's response...

    Kai
  • Hello,

    Sorry about the delayed response, the OPA695 has an older model and it seems the bandwidth was fitted to a closed-loop gain of 8 and does not give an accurate depiction of the GBP. Thank you for bringing up this issue, and we'll look into resolving this as soon as we can. For now please use the graph provided in figure 1 of the datasheet for an accurate representation of the frequency response of the device.

    Best,
    Hasan Babiker

  • Hi Hasan,

    Thanks for your help. Hope you will update the model soon.

    Regards,
    Ricardo