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.

OPA189: Measure Input Offset Voltage

Part Number: OPA189
Other Parts Discussed in Thread: LM7705

Hello together,

I would like to measure the Input Offset Voltage of the OPA189:

The OPA189 has a typical Input Offset Voltage of ±0.4 μV and max. ±3 μV.

I have a HP3458A Multimeter, which can measure DC Voltage in Microvolt range. So I bought the Evaluation Modul DIYAMP-SOIC-EVM and I solder two resistance with 10 kOhm that the gain of the OPA is two.

In the case that the OPA is near at the specific Offset Voltage I should measure a DC Voltage at the Output of about ±6 μV max.

But therefore I measure a Output Voltage of about 140 μV.

That is the circuit:

With:

OPA189

R1 = R2 = 10 kOhm

C1 = 0,1 μF

I hope you can help me to improve my circuit,

Marv

  • Hi Marvin,

    make no short circuit between the inputs of OPAmp! Use a bipolar supply voltage and connect the +input of OPAmp (non-inverting input) to signal ground. Decrease the resistors to 2k each. Connect the HP3458A not directly to the output of OPA189 but insert an isolation resistor of 50...100R. This isolation resistor will isolate the output of OPA189 from capacitive loads and will by this stabilize the OPAmp.

    Kai
  • All good points Kai,

    Why not also increase your op amp gain to 100V/V - the ATE test circuits all make these measures with a lot of gain. There are other errors that can get into your output offset measurement but with V+ at ground and balanced supplies and lower R values, Vos should be dominant.
  • Hello Kai,

    thanks for your ideas. Before I build the circuit in real life I download the TI simulation programm tina to improve my circuit.

    I create a circuit which measure the Offset voltage of 400 nV (as specified in the datasheet).

    The circuit 1) use dual supply and function as good and I measure almost 800 nV at the 0utput (Gain = 2).

    My problem is that I have to use in reality a single supply to measure the offset voltage.

    So if I build the exact same circuit with single suppy I do not measure 800 nV, instead of that 4,03 mV.

    Can you tell what I do wrong? Is that a problem of the software tina, or it is a problem of the circuit?

    (By the way, can you tell me why the bias current IIN+ is negativ?)

    Thank you for help, kind regards,

    Marvin

    1) circuit with dual supply:

    2) circuit with single supply

  • Usually, the output cannot swing completely to the supplies, try that 2nd single supply circuit in sim with a -0.23V on the negative supply (like the LM7705 would give you) and see if input offset might not go back down?
  • Hi Marvin,

    as Michael already mentioned the output of OPA189 cannot fully go down to 0V. See the "voltage output swing from rail" specification on page 9 of datasheet. Using the LM7705 negative bias generator is a good remedy.

    There's another issue with your simulation: You should drastically increase the feedback resistances. 1R is no good choice. Take something in the kOhm range.

    Marvin, I have no clue about what you are trying to do. Measuring the offset voltage with such a low gain makes no sense at all. Michael has already suggested to choose a higher gain, too. So, please elaborate in detail what your application is.

    Kai

  • Oh and Marvin (et al) keep in mind what is in the model is usually one (random) polarity of 1sigma of the bipolar offset voltage distribution. When you get to estimating full output offset and drift range, you need to use the full distribution range - discussed this a little in this recent article -

    www.planetanalog.com/author.asp

    The more interesting aspect here was the polarity of the CMRR in the model - coming from a CFA background, the CMRR in those is always a contracting gain error - for VFA, what should it be? Apparently random and sometimes changing polarities as models are updated? Picked that issue up in the article that followed the one above,

    www.planetanalog.com/author.asp
  • Thank you Michael for the fast answer,

    that solves my problem.
    Perhaps you can you tell me why the bias current IIN+ is negativ?

    In the normal way the bias current come from the OPamp and don't flows in the Input.

    Kind regards,
    Marvin
  • So the input bias current specs are bipolar - these are probably random leakages from various paths and what they put into the model is somewhat random with a magnitude less than the max. Also, you might try looking at those values at different input Vcm points - might change.
  • Hi Kai,

    in the first way my application is about measuring a very low voltage from an ASIC which gives at the output voltage in mircovolt range. My application should be able to measure this low voltage. So I took an OPamp with a very low offset voltage and a gain of 2 to generate at the output a low voltage. This circuit should verify my multimeter and the cable interface among real conditions.

    You are right if, I want to measure only the offset voltage I have to yous a higher gain and so much higher resistance.

    Sorry, if I have cofuse you.

    Marvin