• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 48 subscribers
  • Views 1629 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.

TINA-TI: Question about amplifier noise gain in TINA-TI noise simulations.

Wesley Pot
Prodigy 20 points
Part Number: TINA-TI
Other Parts Discussed in Thread: OPA2189, OPA1632, , THS4541

Hello,

I am trying to make some noise simulations in TINA-TI of my designed instrumentation amplifier circuit made with the OPA2189 and the OPA1632.

According to the book: Operational amplifier noise by Art Kay, inverting and non-inverting amplifiers have the same noise gain of 1 + (Rf/Ri), while the signal gain is different.

An inverting amplfier with a signal gain of 1 has a noise gain of 2.

When i calculate the noise of only the OPA1632 fully differential amplifier, the noise reffered to the output (RTO) is twice the noise refferend to the input (RTI). The noise simulation in TINA-TI gives the correct output noise and total noise indicating the correct use of the noise gain but the input noise is not 2 times smaller than the output noise, it is the same value assuming the use of the noise gain of 1.

When the OPA2189 and OPA1632 as instrumentation amplifier is simulated i expect that the output noise of the first stage (OPA2189 with signalgain = 100 and noise gain = 100) is multiplied with the noise gain (=2) of the second stage (OPA1632 with signalgain = 1 and noise gain = 2), but the simulation gives the same noise at the output of stage 1 and stage 2.

So the OPA1632 noise simulations used the noise gain of 2 to get the correct output and total noise, but the input noise en the output noise of the previous stage is not correctly using the noise gain of stage 2 the get the correct output noise of the compleet circuit.

Also a test with the OPA2189 as inverting amplifier (signal gain = 1 and noise gain = 2) gives the correct output and total noise, but the incorrect input noise.

Am i doing something wrong with the noise simulations or is TINA-TI not correctly using the noise gain?

Kind regards,

Wesley Pot

  • 0
    TI__Genius 15565 points

    Hello Wesley,

    Could you attach a TINA file and possibly provide some screenshots highlighting the issues you are encountering?

    It is important to understand how TINA is generating this input noise plot. To run an AC & Noise simulation in TINA you need one input source and at least one output node. So what TINA does is assume the relationship of output noise = input noise * AC gain. Basically TINA back calculates the input noise by dividing your output noise plot from you AC response. Now this assumption may cause confusion when trying to interpret your input noise plot, for example if the ac response provided attenuation rather than gain, the input noise would actually end up being larger than the output noise of your circuit. Hopefully this explanation helps clarify what you might be seeing.

    Best,

    Hasan Babiker

  • 0 in reply to Hasan Babiker31
    Guru 48395 points

    And Hasan,

    You need to be careful throwing around that "noise figure" phrase - that has a specific meaning described in section 7 of this early summary document on noise, 

    https://www.ti.com/lit/an/sboa066a/sboa066a.pdf?ts=1595429991933&ref_url=https%253A%252F%252Fwww.google.com%252F

  • 0 in reply to Michael Steffes
    TI__Genius 15565 points

    Thanks Michael,

    That's a good point. I've edited the phrase to "noise plot" instead to avoid confusion.

    Best,

    Hasan Babiker 

  • 0 in reply to Hasan Babiker31
    Prodigy 20 points

    Hello Hasan,

    If i understand it correctly, the OPA1632 fully differential amplifier is an inverting amplifier, so when i calculate the noise, i calculate first the noise reffered to the input (RTI), witch is at the non-inverting input of the opamp where the voltage noise source is. Then i use the noise gain to get the output noise, and to get the input noise reffered to the signal source witch is at the inverting input, i devide the output noise with the signal gain. And that is than also what TINA-TI shows in the input noise analysis instead of the RTI noise at the non-inverted input?

    Attached is the TINA file of the OPA1632, where the noise analysis shows that the input noise is equal to the output noise. I expeccted that the input noise, the RTI noise at the non-inverting input was, but if i understand it correctly (see above) it shows the input noise at the inverting input where the signal source is.

    OPA1632 - Noise Analysis.TSC

    Kind regards,

    Wesley Pot

  • 0 in reply to Wesley Pot
    TI__Genius 15565 points

    Hello Wesley,

    It is better to think of the FDA as two inverting amplifiers, each with its own noise source. See picture below:

    In our FDA datasheets, the "Input voltage noise" spec that is listed is actually the differential noise between IN+ and IN- of the part. Similar to a single-ended opamp, the differential output is two times the input voltage noise because the noise gain is 2 for each inverting amp. 

    Now in regard to your TINA file, your signal gain is -1 from your single ended source to the differential output. Because of this, your measured differential output noise will be the same as the input noise at the source in TINA. Attached are two TINA files showing you how to measure the noise of the amplifier itself. You will see in the first that the output noise is greater than twice the amplifier voltage noise, this is because current noise is also contributing to your overall output noise. Bringing the resistors down as shown in the second TINA circuit, you can see the effects of just the voltage noise since the current noise is reduced significantly. 

    OPA1632 - Differential Noise Analysis.TSC

    OPA1632 - Differential Voltage Noise Analysis.TSC

    Best,

    Hasan Babiker

  • 0 in reply to Hasan Babiker31
    Guru 48395 points

    Good points Hasan, 

    Here is that 10kohm file run for output noise on each meter. the one across the inputs is just the modeled input voltage noise showing the expected 1.3nV

    And yes dropping the R's to 10ohms (which the amplifier cannot drive as a feedback load) pushes this very close to just 2X the input noise at the output.

    The OPA1632 does not have a noise analysis section, but here is the noise analysis I put into the more recent FDA's I was working on, This applies to all FDA's, This is from the THS4541 datasheet,