• State
  • Locked Locked
  • Replies 5 replies
  • Subscribers 50 subscribers
  • Views 3195 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.

Multiple Feedback Bandpass with Op Amp on MSP430FR2311

George Fleming
Prodigy 60 points
Other Parts Discussed in Thread: MSP430FR2311

Hi there,


I am actually developing an application on a TI micro controller. However, the question pertains to the op amp integrated on the device. I am attempting to implement a highly selective op amp filter with a component selection given to me by the online filter wizard software:

   

The  ADA4661-2 in the simulation has a GBW of 4 MHz. The op amp on the MSP430FR2311 has the same 4 MHz bandwidth. I also utilized the buffered bandgap, VREF+ on the MSP430 as the level shifting reference voltage. The circuit functions fine when the input is a function generator. As I sweep the frequency from 10 Hz to 25 KHz, the op amp filter has the expected selective response. Likewise, when I input random white noise the filter seems to reject a majority of it:

       

My goal is to detect a high frequency audio pulse with a MEMS microphone (INMP404). However, when I try to feed the input of the microphone into the op amp filter, the filter becomes unstable and defaults to this:

And even if I remove the input source, the same waveform remains until I remove power to the circuit and reapply power. Is there some form of capacitance or other circuit phenomenon that forces the Multiple Feedback Topology to fall into an unsteady state? If I implement a regular inverting amplifier with the microphone and MSP430 op amp, I have no issues. I am only experiencing this issue with the highly selective BPF.

  • 0
    TI__Guru* 93105 points

    Hi George,

    I am not familiar with this adaptation of the multiple-feedback band-pass filter. It is interesting in that it uses what appears to be significant  positive feedback. I've seen some two op-amp topologies that apply a small amount of positive feedback to enhance the Q in a band-pass, but not a single stage design. It may be that the circuit is borderline stable and adding a scope probe to the output adds enough capacitance to push it into oscillation. 

    Most active filter texts recommend limiting the Q (fo / BW) to a maximum of 10 and to keep the voltage gain low for the MFB topology. It appears from the software screenshot that a gain of 100 V/V (40 dB) is being attempted.There may be some issue with the circuit's when it is being pressed for high gain and Q. 

    You may want to take a look at TI's Filter Designer and see what solutions it provides for your band-pass circuit. You can find the software here:

    Regards, Thomas

    PA - Linear Applications Engineering

  • 0 in reply to Thomas Kuehl
    TI__Prodigy 200 points
    Hi Thomas,

    I think it is probably my Q is too high; I have it set to 50. Your right, I also looked into MFB documents and nobody recommends going above 25. I hooked up the function generator again and set the frequency to 17.6khz (the oscillation frequency) and slowly ramped the input until I reached 6 mV. At that point, amplitude peaks on the output rippled and shot to up until it looked like the screenshot above. I think too high of a Q causes instability. I am going to try a lesser Q of 20 and see if I can maintain stability. If not, then maybe 10 is a more appropriate limit.

    Regards,
    George
  • 0 in reply to George Fleming II
    TI__Guru* 93105 points

    Hi George,

    The active filter synthesis tools assume ideal op amps which isn't a bad assumption for most op amp characteristics. However, rail-to-rail output op amps exhibit a complex, open-loop output impedance (Zo) that can be very non-deal at different points along the amplifier's usable frequency range. Often there is a reactive component to Zo that can be capacitive, or inductive, and the resistive portion can be quite high. The complex Zo modifies the filter transfer function sometimes reducing the phase margin to the point where the filter is marginally stable. Unless the op amp's simulation model accurately exhibits the true Zo characteristics a simulation may not reveal a marginally stable circuit. We have striven to incorporate the correct Zo characteristics in our Precision Amplifier models for those brought to market over the past five years, or so.

    I took a look at the MSP430FR2311 data sheet and User's Guide looking to find out more about the internal op amp. The only mention I find about an amplifier is the transimpedance amplifier (TIA). If that is the only amplifier on the microcontroller, it has a committed connection that is specifically for a TIA application. I don't think that it lends itself to the multiple-feedback (MFB) filter configuration you had shown in the schematic. If you end up having to use an external op amp for the filter, you may be able to take advantage of a dual op amp which would more likely be able to meet high-Q, high-gain requirements.

    Regards, Thomas

    PA - Linear Applications Engineering

  • 0 in reply to Thomas Kuehl
    TI__Prodigy 200 points

    Hi Thomas,

    The uC does support general Op Amp features which makes it really cool for many sensor applications.

    I believe your right that the filter software simplifies the model. I know it only takes into account component tolerances and GBW product. In that regard, I did my best to match the op amp characteristics in the simulation:

    Still waiting on shipment of new component values for Q = 20. Will let you know how it goes.

    Regards,

    George

  • 0 in reply to George Fleming II
    TI__Guru* 93105 points

    Hi George,

    Okay, now I see the OA op amp function within the SAC block diagram. I am not super familiar with the MSP430 microcontrollers.

    Thanks, Thomas

    PA - Linear Applications Engineering