Post Go back to editing

AOP Instrumentation

Category: Software
Product Number: NO

Hello,

I would like to find an instrumentation AOP with a gain of around 1500. This AOP accepts low amplitude signals at the input around 1µV to 320µV. Do you know if there are comparator links to find this kind of component ? I use this component in the audio frequency band (10Hz-20kHz).

With a weak PSRR, LOw noise and the supply voltage is +3.3V and -3.3V. 

Thank for your helps

Regards.



Additional information
[edited by: 206_RC at 1:05 PM (GMT -4) on 12 Jul 2022]

Top Replies

    •  Analog Employees 
    •  Super User 
    Jul 12, 2022 in reply to 206_RC +1 verified

    Hi ,

    Yes, the input noise voltage AND the "0.1Hz to 10Hz" Vnoise (Vpp) are good parameters to choose for when trying to find a part that can extend the lower detection voltage range to 1uV, as in…

Parents
  • After looking on the Analog site, I found this link but there are parameters that I do not understand very well :

    https://www.analog.com/en/parametricsearch/11080#/

    Perhaps this parameter helps me to find the minimum input voltage level ?

         

    Thank for your helps.

    Regards

  • Hi ,

    Yes, the input noise voltage AND the "0.1Hz to 10Hz" Vnoise (Vpp) are good parameters to choose for when trying to find a part that can extend the lower detection voltage range to 1uV, as in your case.

    For example, you can use this list sorted for lowest noise:

    Excel file showing these parts:

    XLSX

    I took the lowest noise device in that list (AD8428) and simulated its output noise as shown below. Gain is 2,000V/V. I've filtered the output to 20kHz -3dB Bandwidth using C1:

    In your case, the output noise is a combination of 1/f noise (0.1Hz to 10Hz region) as well as the broadband / flatband noise voltage of 1.3nV/RtHz and simulation shows a RMS value of 591uV_RMS. With 1uV hydrophone input signal and a gain of 2,000V/V, the SNR will be:

    SNR = 20* log (1uV_RMS * 2,000V/V / 591uV_RMS) = 10.6dB

    So, you should be able to decipher that low level of input voltage and stay 10dB above the noise floor.

    Here is the simulation file you can use (Extract All before using):

    AD8428 Hydrophone amp with Filter EZ 7_12_22.zip

    Regards,

    Hooman

Reply
  • Hi ,

    Yes, the input noise voltage AND the "0.1Hz to 10Hz" Vnoise (Vpp) are good parameters to choose for when trying to find a part that can extend the lower detection voltage range to 1uV, as in your case.

    For example, you can use this list sorted for lowest noise:

    Excel file showing these parts:

    XLSX

    I took the lowest noise device in that list (AD8428) and simulated its output noise as shown below. Gain is 2,000V/V. I've filtered the output to 20kHz -3dB Bandwidth using C1:

    In your case, the output noise is a combination of 1/f noise (0.1Hz to 10Hz region) as well as the broadband / flatband noise voltage of 1.3nV/RtHz and simulation shows a RMS value of 591uV_RMS. With 1uV hydrophone input signal and a gain of 2,000V/V, the SNR will be:

    SNR = 20* log (1uV_RMS * 2,000V/V / 591uV_RMS) = 10.6dB

    So, you should be able to decipher that low level of input voltage and stay 10dB above the noise floor.

    Here is the simulation file you can use (Extract All before using):

    AD8428 Hydrophone amp with Filter EZ 7_12_22.zip

    Regards,

    Hooman

Children
  • Hi ,

    Looking at this a bit more, the low frequency (1/f noise region) does not matter much in your case. So, you should only look at Voltage Noise values. AD8428 is still your best choice.

    If your transducer is a high impedance one, then Current Noise values also matter.

    Regards,

    Hooman

  • For example, you can use this list sorted for lowest noise:

    Hi Hooman,

    Thank you for your helps, I try to understand your explanation but I've some difficulties. Your screen shot about the AOP choice is too small and I can't see the parameters you entered :

    This term '1/f noise (0.1Hz to 10Hz region)' what you mentioned I don't quite understand, after you said : It doesn't matter.

    This value :

    After the DS of the AD8428 I can see 1.5nV/sqrt(Hz) and not 1.3nV/sqrt(Hz). 

    In your simulation I can see :

    What does E1 mean and the value -1 ? And your power supply is not within the range recommended by the manufacturer

    As you know, I need this AOP because in another Topic my high pass filter does not detect the 1µV, that's why I wanted to choose an instrumentation AOP to amplify its level.

    If your transducer is a high impedance one,

    No, in fact in my Hydrophone there is an integrated preamplifier so I think  the output impedance of this preamp is very low.

    Thank you for your answer (explanation). 

    Regards

  • Hi ,

    Your screen shot about the AOP choice is too small and I can't see the parameters you entered :

    Response: Yes, sorry about the low resolution. You can use the Excel file I also attached earlier for the same exact information.

    This term '1/f noise (0.1Hz to 10Hz region)' what you mentioned I don't quite understand,

    Response: 1/f noise is the noise in the lower frequencies region. It's called 1/f noise as it typically increases when frequency is reduced. In your case, as you're going up to 20kHz, the 1/f noise region gets to be negligible compared to the flatband / broadband noise of 1.5nV/RtHz (sorry, my parametric table shows 1.3nV/RtHz but you found that the datasheet shows 1.5nV/RtHz which is the number to use).

    What does E1 mean and the value -1

    Response: I used E1 to generate a diff input for the In Amp. I wanted to create a voltage which is 180 deg. out of phase, so I used "-1".

    And your power supply is not within the range recommended by the manufacturer

    Response: Yes, you're correct. You need to select a device from the Excel file which can accept the +/-3V you're using.

    No, in fact in my Hydrophone there is an integrated preamplifier so I think  the output impedance of this preamp is very low.

    Response: Then selecting for the lowest Voltage Noise is the right approach.

    Regards,

    Hooman

  • Hello Hooman,

    In fact, probably this component may suit me, I'll have +5V and -5V also in my board. But whe did you use this assembly with 2 generators V3 and E1 ?

    Why can't you just use one sine wave generator to do the same thing ? Like this :

    There is certainly the reason but I do not see yet and also my electronic diagram does not work. 

    Another point, I would have liked to find this same component (same technical characteristics) but with a variable gain up to around 5000 (because I don't know exactly the desired gain).

    Thank for your helps.

    Regards 

  • Hi ,

    In fact, probably this component may suit me, I'll have +5V and -5V also in my board. But whe did you use this assembly with 2 generators V3 and E1 ?

    Response: The two generators used in simulation is just to apply a fully differential, ground-referenced input. In your application, your hydrophone plays that role. 

    Why can't you just use one sine wave generator to do the same thing ? Like this :

    Response: What you've shown is different in that V1 has no CM reference. Some simulations have issues with sources which don't have a ground reference. That's why I used "E" to generate an equal and opposite signal at the input which does have a CM reference to ground.

    Another point, I would have liked to find this same component (same technical characteristics) but with a variable gain up to around 5000 (because I don't know exactly the desired gain).

    Response: Other devices except AD8428 in the Excel file I shared with you earlier allow a range of gains, as shown here:

    So, all you need to do is to select those devices which show min / max gain range.

    Regards,

    Hooman

  • Response: The two generators used in simulation is just to apply a fully differential, ground-referenced input. In your application, your hydrophone plays that role. 

    I'm not sure, as you know in my previous answer I had said that there is a built-in preamp in the hydrophone. And I think at the output of a preamp I will only have 2 wires for signal and ground, that why I want to simulate with a sinusoidal generator (as I did with the LT6233 high pass filter). 

    Do you think if I can use this component AD8428 also as an Allen & Key type high pass filter ? So this AD8428 serves 2 roles: Amplify and filter the signal at the same time. 

    Otherwise I will have to amplify the signal (with AD8428) and add another AOP to filter (I don't know for this moment) before entering my ADC.

    Thank you for your helps

    PS : So in first I need to simulate the AD8428 with with a sine generator.

    Regards

  • Hi ,

    I don't know of a way to make AD8428 operate in Sallen Key low pass filter configuration. I think it's best if you stick with a gain stage and then add the active filter if you need it further in the signal chain.

    You can use your single-ended hydrophone transducer output on AD8428 applied to the +IN pin. You'd then ground the unused -IN pin.

    Regards,

    Hooman

  • You can use your single-ended hydrophone transducer output on AD8428 applied to the +IN pin. You'd then ground the unused -IN pin.

    Hello Hooman,

    Please (read my posts), I d'ont connect the hydrophone directly to the AD8428 because the Hydrophone has a built-in preamp. Thus I connect directely the preampli to the ampli AD8428. But you're right about the electronic connection : the output from the preamp to the amp AD8428 (+IN pin) and to the ground -IN pin (AD8428).

    Do you know Where (the link) I can find the AOP used for a high pass filter with a gain of 1 (Sallen & Key) ? Of course with a very small offset (Ibias).

    Thank you for your helps

    Regards 

  • Hi,

    This is what I get with the simulation with AD8428

    I have always a offset about -50 mV (which is a lot for a component with 50nA Ibias). Probably my simulation is not correct ? Do you have a idea for this simulation ?

    Thank you

    Please see in this attached the simulation

     AD8428_Y.asc   

  • Hi,

    I don't know if this is the right method but I find an output that has very little offset :

    I've just modify the input resistor and

    follow with another AOP to simulate a load. But in principle the gain is 2000 but the account is not there.