Post Go back to editing

Excess noise in OpAmp Simulations

Category: Software
Product Number: AD8655
Software Version: 17.1.15

I am seeing higher noise in LTSpice than expected and want to know if it is simulation inaccuracy or representative of real noise I could expect from a part.

I am performing noise analysis for a simple OpAmp circuits in LTSpice. e.g.: simple non-inverting with signal gain of 2. If I use the ideal OpAmp model I get expected results (only noise from resistor Johnson noise). However when using models of actual OpAmps I get excess noise which is above what I expect from the datasheet.

Example:

In the circuit below expect noise of ~ 8 nV/rt-Hz at 10 kHz (given the Johnson noise of the two resistors plus the voltage noise, ~3 nV/rt-Hz from the datasheet) but instead get ~14 nv/rt-Hz. This amp should have de-minimus current noise, and in any case the observed excess noise does not depend on the resistor values.

I suspect it is something with the simulation but want to be sure.

Thank you.

Noise_check_AD8655.asc

  • Hi  ,

    I've tried testing the Voltage Noise on my end. (The way it was tested when the model is being built)

    I attached it here for you to see and simulate and relate to the datasheet.

    I noticed that it is not as expected from the datasheet, but, this is probably the way it was asked to be modeled.

    If you do have questions on the results of your model simulation, you can always communicate with the FAE in charge of the part for them to verify if this is really a model issue. If it is, then, we will be expecting a request from them to assess this issue.

    Nonetheless, models are created and verified by the FAE to assess if it really represents the actual part performance.AD8655_Input_Voltage_Noise.zip

    Best Regards,

    Earl

  • Hi Earl,

    Thanks for the reply. If i understand you correctly, you agree the sim doesn't match what you would expect from the datasheet?

    I really am not familiar with the relationships between the various parties you mention. How can I find the FAE for a product?

    NB: I did test this with a second OpAmp, and neither matched the expectations form the datasheet, though some agree better than others.

  • Hi  ,

    I can't really say on my part if the model behavior matches or not, the voltage noise could have been tested differently than the way I was testing it. (Which is a straightforward approach).

    You can always communicate with our ADI contact center or if you do have someone already in communication with ADI then you can ask an FAE to assist you.

    For now, what I can do to help you here is verify our noise simulation if it really does not follow the datasheet. With that said, please give me time to do so. Slight smile

    Best Regards,

    Earl

  • Hello  and ,

    Jumping in really quick here. It really looks like the noise behavior of the model does not match the actual behavior of the part. I have done some simulations on my own and it looks like the broadband noise of the model is higher compared to the actual broadband noise of the part:




    I think the noise behavior of the model should be fixed. I can initiate a request for it to be fixed.

    Best regards, 
    Paul

  • Hi  ,

    Thanks for verifying this concern, please let me know if you've already made a rquest for this, I'd gladly take it for me to fix it.

    Best Regards,

    Earl

  • Hello ,

    I have already initiated the request through the LTSpice Service Desk. Thank you very much.

    Best regards,
    Paul

  • Hi  ,

    I assigned it to myself. I'll give you an update outside EZone on the progress of updating this model.

    Best Regards,

    Earl