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.

  • Resolved

INA826: what is the minimum input signal of INA826?

Prodigy 70 points

Replies: 15

Views: 722

Part Number: INA826

Dear E2E team

This is Besson, a electronic engineer from ABB. 

we are planning to use INA826 in our Level Measurement project as a pre-amp to face a piezoelectric sensor which will generate a 40us narrow pulse with only 200uVpp amplitude. The equivalent C and R of piezoelectric sensor are 900pF and 30Mohm. The power supply what we used for this project is 3V. 

Previously we used opa345 as a pre-amp to achieve 26dB Gain because of low power consumption and low ib, and acceptable slew rate (we want bigger than 1V/us). 

When we looked into the details of comparison between opa345 and INA826, we were aware of that the ib is 35nA of INA826 is much bigger than opa345, at the same time, take the 200uVpp signal from sensor into consideration, we worry about if INA826 can handle such lower signal or not, also doubt that if INA826 can provide enough bandwidth to achieve 26dB Gain or not, please help us to comment, many thanks. 

By the way, we also looked opa314 as a substitute for better performance than opa345, if there is much better alternatives, please tell us, thanks.

Best regards

Besson

  • More info: we will use 1.5V as Vref for INA826.
  • In reply to Besson:

    There is not enough information to answer your questions:  what is the input common-mode voltage?  Is the input signal at 200uV for  40us and 0uV for 40us around Vcm?  If so, the bandwidth of INA826 in the gain of 20 (26dB) is more than enough (BW~250kHz - see below.)

    The input signal depends on supply voltage, Gain, reference voltage and Vcm, thus to answer your queestion you must specified Vcm.

    For 50mV<Vcm<1.35V, 0.1<Vout<2.85 (see below), which allows input differential signal from -70mV to +67.5mV around Vcm.

    Please download the Vout vs Vcm calculator using following link: 

    Marek Lis, MGTS
    Sr Application Engineer
    Precision Analog - TI Tucson

  • In reply to Marek Lis:

    Dear

    The piezoelectric sensor is differential, and we used 1.5V as Vcm for input common mode voltage for INA826 in Tina, and we can get correct simulation result from Tina.

    for the schematics and simulation result, please check the attahced figures.

    But we still have concerns about the 35nA ib of INA826, it is much bigger than the current design - opa345.

    Based on presented information, could you please tell us any other consideration we should take into account if we want to use INA826 as a pre-amp for this application? thanks.

    BR

    Besson

     

  • Hi Besson,

    there are several ways to amplify the signal of a piezoelectric sensor. The important question is: Which topology is the manufacturer of sensor recommending? Can you give us a link to the sensor?

    Kai
  • In reply to kai klaas69:

    Dear

    The piezoelectric sensor which we used was designed by internal team, not a standard product from market. I think we can understand it as a common piezoelectric sensor.
    Previously, we use opa345 as a difference amp to amplify this signal, and amplified very well, but in some cases with harsh enviroment, the circuits was interfered by 50Hz noise, so we plan to use INA826 to improve this perfermance.
    could you please answer my question about the ib-35nA presented previously? thanks.

    BTW, do you know why Tina can't implement a signal analysis for frequency response when there is two signal sourse (VG1 and VG2 used as a differential sensor in attached figure, Feb 12), and how to get it?

    BR
    Besson
  • In reply to Besson:

    Hi Besson,

    has your sensor actually three terminals and produces a true differential signal?

    Kai
  • In reply to kai klaas69:

    Hi Kai

    Our sensor has two terminals without the common one but will produce a true differential signal with my consideration.

    BR
    Besson
  • In reply to Besson:

    Hello Besson,     

    From your schematic above, the 50Mohm of your piezo does not come into play for two reasons. 

    1. It is AC coupled to the amplifier stage so the DC blocking caps C11/C12 decouple the sensor from a DC perspective. 

    2. Even if the DC blocking caps C11/C12 weren't there, the 50Mohm is in parallel with the 100k so 100k dominates. 

    To calculate the Input offset current contribution:

    VosIOS= 100koh*0.7nA = 70uV

    G = 1 + (49.4 k/2.61K)  = 19.9272

    So, you'll see 1.395mV of offset referred to the output of the INA826 with a gain of ~20.  

    However, once again, it doesn't matter since  the DC blocking cap C2 decouples this DC error.  

    Please let us know if you still have questions.

    Best regards,

    Raphael Puzio

    Raphael Puzio

    Precision Amplifiers Marketing

     

  • In reply to Raphael Puzio:

    Hi Raphael

    Thanks a lot for your answer, it's clear now for me to understand the Ib and Ios to face our piezoelectric sensor application, and I think INA286 is adequate.

    I also get support from FAE Cera Wei from TI, I will test our sensor when the INA286 EVM board is available.

    Best regards,

    Besson    

  • In reply to Besson:

    Hi Besson,

    the edges of signal are a bit rounded but the INA826 seems to be fast enough:

    besson.TSC

    For low noise operation it's crucial to keep the distance between sensor and INA826 as minimal as possible! Avoid any cabling here. If you need cabling, do it at the output of INA826, but not at the input. It's better to increase the signal amplitude before doing the cabling, though. So I would add a second amplifier to increase the signal amplitude of INA826's output signal. Also, use proper shielding of sensor and amplifiers.

    Kai

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.