• State Resolved
  • Locked Locked
  • Replies 12 replies
  • Subscribers 50 subscribers
  • Views 2578 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

INA199A1 Unexpected gain change

297925
Prodigy 80 points
Other Parts Discussed in Thread: INA210, INA199, INA199EVM


Dear Sir,

I am using an INA199A1DCK (sampled) in a standard shunt monitoring application.
Vvcc = 12V, Vref = 5V, Bidirectional.  Rshunt = 50mOhm. The circuit is protected with a R/C filter and a dual zener diode (AZ23C24), as shown in the datasheet.
The Vcm is within the specified range.

The resistors used for the filter (two resistors both in series with IN+ and IN-) are 2k2. Concluding from the datasheet this would give a gain of 45.
It does not though, the gain becomes 17. I verified the marking of the INA199A1 to be correct.

I removed the zener diode and capacitor to eliminate influence, and so there is only the shunt, two 2k2 resistors and then the INA199A1 inputs.
Still the gain stays 17.

Any idea how this can happen?

 

 

 

  • 0
    Prodigy 80 points

     

    It seems to me that this device does not meet common mode voltage specifications. The bias currents start increasing rapidly at 23V already.
    At a certain point (below 26V) the device is damaged. 100E Resistors and clamping diodes do not protect sufficiently against this.

    Also the gain deviation formulas do not seem to apply. Adding 100E...2k2 gives other gains then the formula's say.

     

     

     

  • 0 in reply to 297925
    TI__Mastermind 25815 points

    Hello Tim,

    Here are a couple observations:

    a)  If your resistors in series with the inputs are 2.2kohms (I assume this is what you mean by 2k2), this value is too large and will cause significant error.  The datasheet recommends 10ohms.  There is a better discussion of this in the INA210 datasheet.

    b)  If you're seeing common-mode rejection errors, I would look at how you supply 5V to the ref pin.  The source should be of low impedance.  If you're using a resistor divider, please buffer it with an op-amp.

     

  • 0 in reply to Pete%20Semig
    Prodigy 80 points

     

    Hello Pete,

    a) I already assumed the formulae only are valid for small series resistors. However the datasheet does not mention this.
    The error which the formulae specified would have been acceptable; but the calculated error is not correct. (Expected gain 45, Real gain 28 at 910E)

    Furthermore, the increase of these resistors has been done to prevent the device from being damaged .
    The device can not handle the 26V max. rating specified in the datasheet, by my observations.

    The common mode impedance is low (7805 + cap). I am also not having common mode rejection issues (even not with 910E series resistors)
    My problem is the devices get damaged by CMV below 26V.

    Regards Tim

     

     

     

  • 0 in reply to 297925
    TI__Guru 63225 points

    Hi Tim,

    Pete is tending to a family matter, I will attempt to fill in for him here.

    The INA199's ability to withstand the +26V common-mode voltage is verified in the final test solution for the INA199 before it is released from manufacturing. 

    That said, a transient or series of transient events on the bus that exceed the +26V rating will damage the part.  Are you monitoring the common-mode voltage to verify that there aren't any transient events that violate the +26V rating? 

    If there aren't any transient events, will you please send me a rough schematic of your test setup including your methodology for testing the +26V common-mode input for the part?  Please include how you are measuring the voltages/currents and what information you use to determine that the part will not survive the +26V common-mode voltage?  You've mentioned above that the "Bias Current" begins to increase.  Is this the Input Bias (Ib) current into the input terminals of the amplifier (IN+, IN-), or is this the quiescent current (Iq) of the INA199 that begins to increase?

    Thanks and Regards,
    Collin Wells
    Precision Analog

     

     

     

     

  • 0 in reply to Collin Wells
    TI__Guru 63225 points

    Hello Tim,

    Do you have any updates on this issue?  Has it been resolved?

    Thanks,
    Collin Wells
    Precision Analog

  • 0 in reply to Collin Wells
    Prodigy 80 points

     

    Hi Collin,

    We are no longer using this component for now.

    But to help you with this problem:

    This setup is the exactly the same as Figure 25 of the datasheet.  I tested the common mode withstand voltage just by having a seperate supply on one of the shunt resistor's connections.The other connection is open. So there is just a CM voltage and (almost) no DM voltage. These voltage were verified with a DMM.
    So i am not testing with transients, just DC.

    Now i dont have exact numbers, but the voltages over the protection resistors start to increase, so the Ibias increases before CM=26V.
    If i remember correctly the bias current starts increasing at 23/24V. Both Ibias currents also start to differ leading to an output change.

    At some point between 23...26V the INA breaks down and only outputs the reference voltage.

    To work around this we attempted increasing the Rprotection to 910E. This ensures the INA no longer being damaged, but the gain drops significantly, and is not conform with the stated formulae.

    Regards Tim.

  • 0 in reply to 297925
    TI__Mastermind 25815 points

    Hello Tim,

    In an attempt to re-create your observations, we have measured the input bias current of a known good device using the INA199EVM.  We do not see an unreasonable increase in input bias current at common-mode voltages above 23V.  Our results are consistent with Figure 8 in the PDS.

     

  • 0 in reply to Pete%20Semig
    Prodigy 80 points

     

     Hi Pete,

    Just to give you an idea of the schematic we use:

    Could the used Vref or supply be of influence?

    Did you see a apperent differential voltage starting to form; due to the bias currents starting to differ a little bit at Vcm > 24V?

  • 0 in reply to 297925
    TI__Guru 63225 points

    Hi Tim,

    Pete and I discussed the tests he ran last week for you on the INA199.  There were not any increases in the input bias current or the input (Icc) current of the INA199 as the input common-mode was increased that didn't align with the information in the datasheet.  The datasheet and Vref values shouldn't matter as long as they are within the datasheet limits.

    After looking at your schematic, the "D2" diodes appear to be 24V Zener diodes.  These diodes would begin to conduct before 24V and will be in full reverse-breakdown around 24V.  If these diodes are installed in the system during your testing they are the reason for the increase in current.  I looked at the datasheet for the diodes and they can begin to conduct as low as 18V (see image below).

     

    Please let me know if removing the zener diodes fixes the bias current issues you're experiencing.

    Thanks,
    Collin Wells
    Precision Analog

  • 0 in reply to Collin Wells
    Prodigy 80 points

     

    Hi Collin, you must be right regarding the bias currents. I thought this was causing the damage, but i am no longer sure if i removed the zeners, since i dont have the hardwar anymore. However it does not explain the damaged components. They were damaged with Vcm being in range.


    But since we are no longer using this setup and i must continue on other subjects; i will no longer continue this discussion.
    Many thanks for the support.

     

     

  • 0 in reply to Collin Wells
    Prodigy 60 points
    Collin. Are you available to answer a question?
  • 0 in reply to COURTLAND THOMAS
    TI__Mastermind 25815 points
    Hello Courtland,

    Please create a new forum thread for your question. That way it will be routed to the proper engineer for support.