• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 48 subscribers
  • Views 3403 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

EQUAL OFFSET VOLTAGE IN A DUAL OPAMP?

marc windels
Prodigy 50 points

Other Parts Discussed in Thread: OPA2209, OPA209

I measured the offset voltage (referred to input) of 2 pieces of dual opamp OPA2209.

piece1 : opamp A : +25 uV   opamp B  :+41 uV

piece 2 : opamp A +31 uV  opamp B : +21 uV

In my design, the offset voltage of the total circuit is proportional to the difference of the offset voltages of both opamps A and B.

May I assume that for every opamp OPA2209 the offset voltage of opamp A and opamp B will have the same polarity?

May I assume that for every opamp OPA2209 the offset voltages of opamp  A and opamp B  will change in the same direction with increasing ambiant temperature?

Thank You

  • 0
    TI__Guru 74331 points

    Marc,

    Both of the OPA209 channels in a dual package are trimmed at the wafer level to zero but subsequently shift during molding process due to the packaging stress.  Since these shifts are completely random, there is NO CORRELATION between the input voltage offset of each of the OPA2209 channels - this is true for all op amps.

    In other words, the input voltage offset of one op amp gives you no information regarding the offset of the other op amp even if they are in the same package; therefore, this is exactly the same as comparing the offset of one single op amp to the offset of another single op amp.  In the case of OPA2209 each channel's Vos can be anywhere between -150uV and +150uV - see histogram below.  The same is true for the input voltage offset drift.

  • 0
    Prodigy 50 points

    Thank You Marek for your very clear answer.

    I have 2 other questions

    1) Is this true for every dual opamp? Even for dual opamps that are not trimmed to zero at the wafer level?

    2) For the dual opamp OPA2209 I suppose that I may assume that the difference of unity gain frequency of opamp A and opamp B will be small.

    This is relevant information in a design that employs active feedback for phase compensation.

    Thank You

  • 0 in reply to marc windels
    TI__Guru 74331 points

    ,

    Marc,

     

    Please see my answers below:

     

    1) Is this true for every dual opamp? Even for dual opamps that are not trimmed to zero at the wafer level?

    Yes, there is no correlation between the input voltage offsets of two channels of any dual op amp as long as their normal distribution is centered around zero (no gross systematic mismatches).

     

    2) For the dual op amp OPA2209 I suppose that I may assume that the difference of unity gain frequency of opamp A and opamp B will be small.

    Yes, there is a strong correlation between two channels of the dual op amp when it comes to the specs centered around a fixed value like of OPA2209 bandwidth (typical BW=18MHz), quiescent current (typical IQ=2.2mA),  etc.  Even though the absolute value of the BW may be changing with the process as much as +/-30% over a period of time measured in years, the difference in the bandwidth of two channels of a dual package OPA2209 should always be less than few percent since it is only a function of resistor, capacitor and device matching.

  • 0
    Intellectual 410 points
    The problem is that any precision (low offset) op-amp is already trimmed to cancel out all the first-order effects that you can expect to be matched. For cheap crappy high-bias op-amps, you will get approximate matching, but for the parts you're talking about, what you're seeing is the random residual left over after all that's been compensated out. So sorry, your total offset is going to be 3 dB higher (1.414x) than the per-op-amp figure.
  • 0 in reply to Marek Lis
    Prodigy 50 points

    Thank You Marek again,

    I think I understand it now.

    My confusion was about the word "typical" and "maximum" in the specs of the offset voltage.

    I thougt  "typical" was the same as "average" and I thougt maximum was the random variation from the average.

    In the distribution I clearly see the average is zero. The difference between the actual offset voltage and the average offset voltage is due to a random phenomenon.

    So typical means that I have a certain great probability that for every opa2209 the offset voltage will be < 35 uv (in absolute value)  and maximum means that the probabilty that for every opa2209 the offset voltage will be < 150uv equals 1. (at given testconditions)

    Thank you Marek for your time and effort.