• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 1 answer
  • Subscribers 47 subscribers
  • Views 1158 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

OPA2330: 20kHz Noise riding on output signal in current measure circuit

Paul Moore
Prodigy 115 points
Part Number: OPA2330
Other Parts Discussed in Thread: OPA2325

Please explain why the OPA2330 is generating so much output noise in this current feedback circuit.
Is there a way to modify the circuit to eliminate the noise?
Why does Microchip MCP6232 work so much better (noise-wise)?
(please see the attachment for detailed information)Tech Support Request - Current Feeback Circuit Noise Problem.docx

  • 0
    TI__Guru* 93105 points

    Hi Paul,

    First, I would like you to confirm that the OPA2330 full circuit application looks as shown here so we can make sure we have a correct understanding of it.

    I used a load current of 5 Amperes as an example to produce an output voltage within the linear range. Then, the issue comes down to the noise difference you are observing between the OPA2330 and the other op amp you are comparing it to.

    Fundamentally, the two op amps appear to be very different from each other by virtue of their voltage offset and offset drift over temperature. The OPA2330 maximum voltage offset is 1/100th of the other, and that is attributed to it being an auto-zero op amp. It uses internal switching and offset storage to nearly zero the offset. It is highly likely that the higher offset op amp does not employ such techniques, because it doesn't need to when high voltage offset is acceptable in a circuit. 

    There is a possibility that your application circuit is sensitive to the input switching that takes place in the OPA2330 input stage. Sometimes, the charge transfer that takes place in the input switches results in output noise due to an input current noise to voltage noise conversion. That can occur in circuits where the op amp's input sees a high impedance looking back into the source and feedback paths. I don't think we have seen that occur in a circuit where the source and feedback impedance are low as they are in your circuit, but there may be something different here because of the current feedback. If that is happening it could be difficult to remedy. 

    Would it be possible for you to evaluate one of our non auto-zero op amps in the circuit to see if the noise improves? If so, can you tell me what op amp characteristics are important in this application? 

    Regards, Thomas

    Precision Amplifiers Applications Engineering

  • 0 in reply to Thomas Kuehl
    Prodigy 115 points

    Hi Thomas,

    Thank you for the quick reply.

    Yes, the circuit you show is the one I am using.

    ILoad is 0A to 10A, VF1 is 0V to 3.13V.

    The opamp inputs sit at the rail at 0A.

    I am open to trying another opamp (the MCP has terrible vos).

    Design specs:

    1. CMR for the inputs should be a more than the rails.
    2. GBW <1MHz is adequate
    3. Input offset in sub mV, over temperature, is greatly desired but not necessary.
    4. Has to be SOIC8 package.

     

    Thank you for your help.

    -Paul

  • 0 in reply to Paul Moore
    TI__Guru* 93105 points

    Hi Paul,

    Thank you for the additional information about the op amp requirements for the high-side current application. Internally, we are in discussion about your OPA2330 circuit attempting to understand the output noise. Close examination of the spike occurrences within the noise image you provided indicates that it is probable that it is being generated by the offset auto-zero switching.

    Moving away from an auto-zero op amp our e-trim op amps offer the next lowest voltage offset and drift. They do not employ any auto-zero or chopping to attain their very low offset so there is no possibility for switching noise.

    An op amp that easily meets or exceeds the list of requirements is the OPA2325. You can view the OPA2325 datasheet here:

    https://www.ti.com/lit/ds/symlink/opa325.pdf

    The OPA2325 has much wider bandwidth than the OPA2330 and that was something I wanted to check. I did a stability analysis on your circuit and it looks like the phase margin is high with both op amps. Instability should not be an issue with the OPA2325 in the circuit.

    Let me know if you have any questions.

    Regards, Thomas

    Precision Amplifiers Applications Engineering

  • +1 in reply to Thomas Kuehl
    TI__Guru* 93105 points

    Hi Paul,

    One of my applications colleagues who covers the precision low voltage op amps such as the OPA2330 suggested adding a 200 pF filter capacitor from its output to ground. You might try this if the PC board will accommodate the capacitor, and then see if that reduces the noise to an acceptable level. If the PC board doesn't accommodated the capacitor the alternate op amp approach may be easiest. 

    Regards, Thomas

    Precision Amplifiers Applications Engineering

  • 0 in reply to Thomas Kuehl
    Prodigy 115 points

    Thomas,

    I'm purchasing some of the OPA2325 opamps to try out.  Spec's look very good.

    I will also try the output capacitor.

    I will let you know what happens.

    -Paul

  • 0 in reply to Paul Moore
    TI__Guru* 93105 points

    Hi Paul,

    Sounds good. We are interested in hearing the results from the two proposed options.

    Thanks, Thomas

    Precision Amplifiers Applications Engineering

  • 0
    Prodigy 115 points

    Thomas,

    Results of testing.

    OPA2325:  Noise similar to MCP6232, maybe 10mV greater, but very good.  Offset measures 160uV, right at edge of spec.  100°C drift will add another 750uV (spec)

    OPA2330 with 100pF ouput cap:   Noise is dramatically better than without cap. 200mVpp down to 50mVpp.  Offset is still amazing.  I measured 47uV and 27uV in two devices.  Adding a cap across the output resistor, R9 in your drawing, was another significant improvement that would work well with any of the opamps.  The OPA2330 does still require the output cap.

    The lower cost TSZ182IDT performs similarly to the OPA2330.  Noise is still a problem but maybe the output cap trick will help this too.

    I have some pictures but I don't see a way to upload them at this moment.

    I will claim that the output cap is the solution to my original post.  Solved!

  • 0 in reply to Paul Moore
    TI__Guru* 93105 points

    Hi Paul,

    Glad to hear that adding the cap at the op amp output much improves the noise performance in your application. It could be that the much increased bandwidth of the OPA2325 over the OPA2330 results in higher integrated noise at its output - maybe comparable to the OPA2330 charge transfer noise. Charge transfer noise for the OPA2330 would be higher in frequency than the gain-bandwidth of the op amp and adding the cap at the output should limit the bandwidth and reduce the higher frequency noise.

    Yes, the OPA2330 input voltage offset should be very low. Certainly a good reason for using an auto-zero op amp!

    Regards, Thomas

    Precision Amplifiers Applications Engineering