Post Go back to editing

SPICE MODEL - ASD Out Noise

Thread Summary

The user observed a discrepancy in output noise between the Analog Devices Signal Chain Designer Tool and SPICE simulations using the OP467G macro-model. The Signal Chain Designer Tool uses datasheet information for accurate noise simulation, while the SPICE model lacks necessary noise parameters. A support engineer suggested updating the NPN model with additional parameters (RB=1480, KF=6E-11, AF=1) to improve noise modeling accuracy.
AI Generated Content
Category: Datasheet/Specs
Product Number: OP467S

The following figures show the output noise inferred for the OP467G operational amplifier.

The first figure was obtained using the Signal Chain Design Online Tool provided by Analog Devices.

The second figure shows the same circuit configuration; however, in this case the simulation is performed using the SPICE model downloaded from the Analog Devices website, specifically from the resources section of the OP467G product page.

Could anyone provide an explanation for the discrepancy between the two results (acknowledging that the underlying models are different)? Additionally, is there a way to obtain or replicate the model used in the first tool so as to achieve consistent and correct results?

Thread Notes

Parents Reply Children
  • Hi David,

    Thanks for your feedback.

    I updated the model according to your proposal, but unfortunately the results are worse than before.

  • Sorry that the model update I received didn't work for you.  I took a look and was able to tweak the model to get the noise to match better over the frequency range provided in the datasheet.  Try this for the NPN model: .MODEL QN NPN(BF=33.333E3 RB=1325 KF=1E-15 AF=1.06 ).  Note that I did this for the OP467 model and not the OP467G version.

  • Dear David,

    Thank you for your effort.

    While the proposed modification to the model produces a correct noise spectral density within the bandwidth of interest for my application, it still does not align with the noise spectral density obtained from the well-constructed model used in the Analog Devices Signal Chain Designer.

    I will consider the case closed; however, future readers of this discussion should be aware that the modified model lacks accuracy for applications at both low and high frequencies.

    If the model used in the Signal Chain Designer is available, I would suggest that Analog Devices consider updating the models provided on the product page accordingly.

  • Thank You for the feedback.  Since Signal Chain Designer is modeled differently, the next best solution is the existing SPICE model is updated to match the noise across the entire frequency range and reverified.  I will pass this information to the appropriate team.