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.

OPA2348: Microphone circuit

Part Number: OPA2348
Other Parts Discussed in Thread: TLV316, TLV2314

Hi Sir,

Microphone specifications:  -35db (Sensitivity )  ,  60db(SNR)

Sound distance: 70db @ 6m

After going through the amplifier circuit, we want to get an output of about 300mV.

How much magnification do we need to design?
Do you have any recommended parts?

  • Hi Jason,

    provided that it is a typical microphone specifiction, then 0dB = 1V / Pa = 94dB(SPL). 70db(SPL) means 0.063Pa. And a sensitivty of -35dB results in a signal of 0.018V x 0.063 = 1.1mV at 70db(SPL). So, for 300mV output level you would need a total gain of about factor 270.

    To recommed a suited OPAmp we need to know what kind of microphone we are speaking about. Is it a dynamic microphone or an electret microphone? So, please give us more details on the microphone. Do you have a datasheet of microphone?

    Kai

  • Former Member
    0 Former Member in reply to kai klaas69

    Thanks for the help Kai!

    Yes, Jason.  A part number and/or datasheet would definitely help here.  Let us know if there's any other information you may think is relative.

    Regards,

    Daniel

  • Hi Kai,Daniel,

    Please refer to the following documents

    Thanks!

    WM-61A(A mic 規格書).pdf

  • Hi Jason,

    you might want to have a look at this TI's reference design:

    tidu765.pdf

    Kai

  • Former Member
    0 Former Member in reply to Jason Li2

    Hi Jason,

    The resource Kai has suggested is quite good and may be exactly what you need.  I don't have the correct font available in adobe to read your pdf, but from the image I'm guessing the microphone is an electret.

    As to selecting an amplifier, I think you should consider your power concerns and available supply rails.

    I'm not an expert on this subject, but having quickly gone through this design method I'm fairly confident we'll be able to find something to fit your needs.

    Regards,

    Daniel

  • Hi Daniel,

    yes, it's an electret microphone:

    Kai

  • Former Member
    0 Former Member in reply to kai klaas69

    Thanks Kai!

  • Hi Kai,Daniel,

    <1>

    Is there a design using 3, .3v?

    If possible, we want to design low power consumption.

    <2>

    Which one do you recommend , inverting amplifier and  non-inverting amplifier?

    Thank you very much!

  • Hi Jason,

    how much supply current would be acceptable?

    Kai

  • Hi Kai,

    Because we need a total gain about factor 270. and bandwidth is 20KHz.
    I think we have two ways. Which one do you recommend? Or there is a better way?

    1. use 2 stage design: OPA2348(1MHz)
    iq=45uA x 2 =90uA
    2. use 1 stage design: TLV316(10MHz)
    iq=0.4mA


    <2>
    we saw that the document said:
    inverting amplifier is possible to achieve lower noise than with a non-inverting amplifier. But the R1 resistance must be higher than C3 impedance

    but if we use 3.3V power supply. R1 resistance can not be too large(about 2.2k)

    so should we use this design method? Or is there a better way?

  • Former Member
    0 Former Member in reply to Jason Li2

    Hi Jason,

    Using two stages will give you the advantages of being able to pick two different devices.  So, you can use the first stage as a low-noise, high-gain stage and the second to maximize your output signal.  You can also use lower GBW parts, as you have shown here.  However, this will complicate things and may require more current.  Your choice should be based on your top spec priorities.  Which is more important, noise or power?

    My suggestion is to make the design as simple as possible.  If you go through the design process and realize you'd rather have two stages, then you can switch.

    For your second point, I could not find where in the document this is being referred to.  Can you please point this out?

    The main downsides to the lower R1 value are the larger C3 that will be needed and the greater noise gain, as explained.

    Regards,

    Daniel

  • Hi Jason,

    I would do it with the TLV2314:

    Jason_mic_tlv2314.TSC

    Kai

  • Hi Daniel,

    Please refer to Page4, page6

    BR,

    Jason

  • Hi Jason,

    don't overlook the noise contribution of the electret microphone itself! A sensitivity of microphone of -35dB means a signal of 18mV. A signal-to-noise-ratio of 62dB gives a noise signal of 14.3µV. Multiplied with a gain of 270 results in a noise signal of 3.9mV at the output of amplifier. But the amplifier I suggested gives a noise level of only 500µV:

    So, the noise only increases to

    SQRT (3.9mV^2 + 500µV^2) = 3.93mV

    which is a difference of 20 x log (3.93mV / 3.9m) = 0.07dB which is not detectable by human ears :-)

    So, it makes no sense to look for an amplifier providing a lower noise.

    Kai

  • Former Member
    0 Former Member in reply to Jason Li2

    Hi Jason,

    Thanks for pointing that out.

    If you take the given configuration, you control the gain with R2.  So by increasing R2, you can increase the gain as you please.  Now obviously, if we make the resistors excessively large then the thermal noise can become an issue.  However, the noise gain is controlled by the ratio of R2 to R1.  So, we can increase the gain without having to grow the noise gain.

    I believe that the assumption here is that, given a non-inverting configuration your signal gain and noise gain equations would be the same: 1 + R2/R1.  Thus, increasing one would increase the other in the same proportion.

    I have not run Kai's simulation numbers, but he makes an excellent point.  If the microphone is not that great, then we don't there is a limit to how low the noise on our amplifier should be.  Beyond that, you'd just be paying for performance that you cannot use.

    Hopefully this helps.

    Regards,

    Daniel

  • Hi Kai,

    Thanks for your explanation~

    In addition, I would like to ask a few questions

    1.  Why not use the same design about R3,C3/R4,C2 , R7,C5/R8,C4. Is there any particular reason?

    2. Can I use the OPA2348 for your circuit diagram? Is the performance similar?

    Thanks!

    Jason

  • Hi Jason,

    a total gain of 270 means a gain of factor SQRT(270) = 16.4 per stage. For a bandwidth of 20kHz the OPAmp should have a minimum gain reserve (loop gain) of factor 10 at 20kHz. Combined with a gain of factor 16.4, an OPAmp with an open loop gain of 10 x 16.3 = 163 at 20kHz would be required which translates to a unity gain bandwidth of 3.26MHz. Because of that I have chosen the TLV2314 which is an inexpensive 3MHz OPAmp with low supply current. In contrast to the TLV2314, the OPA2348 is not fast enough and generates lots of distortion. Also, the OPA2348 is not capable to deliver the necessary gain at 20kHz:

    Jason_mic_opa2348.TSC

    The TLV2314, on the other hand, generates very little distortion at 10kHz and does not degrade the frequency response:

    Referring to your other question: The feedback resistors of the first stage are chosen low ohmic to keep the noise low. Here the signal is still small and the low ohmic feedback resistors will not contribute much to the supply current. The second stage, on the other hand, does not need to have low ohmic resistors because the signal is much higher (factor 16.3) and the feedback resistors will not contribute much to the noise. So, increasing the feedback resistors helps to keep the additional supply current low. Of course, if lowest noise isn't desired, you can take bigger feedback resistors also in the first stage. 

    Kai

  • Former Member
    0 Former Member in reply to kai klaas69

    Kai,

    Thanks for running the math and providing your simulations.

    I think it's clear from your GBW calculations that for the sake of not having to design with a high-speed amplifier, it's worth going the two stage route.

    Thanks again for your support.

    Regards,

    Daniel

  • Hi Kai,

    Thank you very much~

    I want to ask your last question.
    If I output directly to the MCU(ADC), can I remove C6,R9,R6 ?

    Thanks!

    Jason

  • Hi Jason,

    yes, C6 and R9 should be removed, unless you want a signal which also goes negative. R6, on the other hand, can remain in place and could be part of a RC charge kick back filter. The dimensioning of this filter mainly depends on the used ADC. R6 isolates the output of OPAmp from the storage cap then and assures stability.

    Kai

  • Former Member
    0 Former Member in reply to Jason Li2

    Hi Jason,

    If the purpose of C6 is to AC couple but you are on a single-supply system.  I would think it doesn't make much sense in this case and I think this is what Kai has alluded to.

    Connecting to the ADC is a bit more complicated and how you do that depends on which ADC and which type of ADC you use.  If you need help with this, I'd recommend a couple of resources.

    1. Watch a bit of the Texas Instruments Precision Labs series on ADCs.  They cover several topics in-depth, including a very detailed step-by-step process for designing the input driver for a SAR ADC.

    2.  Once you've picked out an ADC, you can post on e2e again to get an engineer from that team to further help.

    Regards,

    Daniel

  • Hi kai,

    Thanks!

    Jason

  • Hi Daniel,

    Thanks~

    Jason

  • You are welcome :-)

    Kai

  • Former Member
    0 Former Member in reply to Jason Li2

    Happy to help Jason.

    Regards,

    Daniel