• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 50 subscribers
  • Views 1333 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.

OPA365 Vos(max) Ta=-40℃ to + 125℃

alpha
Expert 2530 points
Other Parts Discussed in Thread: OPA365

To whom it may concern

Our customer is considering to replace from MAX4488 to OPA365.
One concern is MAX4488 is specified Vos(max) in over specified temperature range , Ta = -40℃ to +125℃ but OPA365 is not.

Could you provide the value like MAX4488?
If you don't have , how can we assume Vos(max)?

Your reply would be appreciated.

  • 0
    TI__Genius 12631 points

    Hello,

     The max. Vos @ 25C is 200uV. The typical Vos drift is shown as 1uV/C. The typical usually indicates a 1sigma distribution over process corners and mismatch. The worst case is going to be a 5-6sigma so we are looking at 6uV/C.

    From 25C to 125C that is 600uV of Vos change, so the max is 600+200 =  800uV.

    From -40C to 25C that is 390uV of Vos change, so from a minimum of -200uV Vos @ 25C the minimum @ -40C = -590C.

    So we are looking at a worst case of -600C to 800C of absolute Vos. This is going to be a very pessimistic view and you will typically see much better performance.

    I looked at the Max4488 datasheet and its quite possible they used a 3 sigma distribution to calculate the worst case offset across temperature, since they show

    Vos max @ 25C = 350uV

    Vos drift max = 6uV/C

    but Vos max across temperature = 750uV.... if we use the same principles as above then the worst case Vos should be closer to 950uV. Hope this makes sense. This product does not belong to my team but I will forward it to the correct team as well so that they can put in their own inputs.

    -Samir

  • 0 in reply to Samir Cherian
    TI__Guru 60610 points

    Bell,

    Actually, Samir would be correct in calculating the maximum Vos over temperature if OPA365 used bipolar input transistors.  However, since OPA365 utilizes CMOS input, Vos and drift are NOT correlated and thus must be added like vectors – as sum of squares.  Also, since a typical drift value represents one standard deviation of Normal Gaussian distribution, it is up to customer to decide what failure rate he/she is willing to accept (OPA365 is not tested in production over temperature and thus no absolute maximum value can be assured).

    Therefore, in calculating the max Vos over temperature, customer first needs to decide what failure rate is acceptable to him.  If 1 out of 370 units could fail, 3 sigma (+/-3uV/C) should be used in the calculation resulting in: 

    Max Vos (125C) = +/-√[200² +(100C*3uV/C)²] = +/-360uV  where ΔT=125C-25C=100C

    Max Vos (-40C) = +/-√[200² + (65C*3uV/C)²] = +/-279uV  where ΔT=25C-(-40C) =65V

     

    If 1 out of 15,000 units could fail, 4 sigma (+/-4uV/C) should be used in the calculation resulting in:

    Max Vos (125C) = +/-√[200² +(100C*4uV/C)²] = +/-447uV  where ΔT=125C-25C=100C

    Max Vos (-40C) = +/-√[200² + (65C*4uV/C)²] = +/-328uV  where ΔT=25C-(-40C) =65V

    Below please see Normal Gaussian distribution probabilities:

  • 0 in reply to Marek Lis
    Expert 2530 points
    Hello Marek

    Could you please tell me why customer can can calculate as sum of squares in CMOS input case?

    Our customer would like to confirm why there is difference between bipolar and CMOS input case.

    They have additional question
    Could you provide how much sigma value is defined for Vos max value?

    Your reply would be appreciated.

    Thanks and best regards.
  • 0 in reply to alpha
    Expert 2530 points

    To whom it may concern

    Could you please answer about the previous post ?

    Your answer is important for our customer to decide whether OPA365 could be use.

    Your reply would be appreciated.
    Thanks and best regards.

  • 0 in reply to alpha
    TI__Genius 12631 points
    Hello,
    In the case of a bipolar amplifier, there is a strong correlation between the offset voltage and the amplifiers vos drift. So we can use a linear equation to predict the amplifiers offset over temperature. In the case of a CMOS amplifier, the Vos @ 25C is uncorrelated from the Vos drift. When combining 2 uncorrelated effects, for mathematical correctness we should take the square root of the sum of squares of the individual effects. This is similar to calculating the effective noise of a system due to multiple, uncorrelated random noise source.
    Looking at the Vos distribution graph on Page 4 of the datasheet, it looks like the Vos is around +/-40uV (maybe a little less) so we are looking at a 5-sigma variation to max. value.
    -Samir