• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 50 subscribers
  • Views 257 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.

Original question:

LMV324A: What is the normal distribution of the LMV324A Input Offset Voltage?

OPA837: Is it appropriate to assume offset voltage and bias/offset currents as zero mean random variables?

TAPABRATA SEN
Prodigy 30 points
Part Number: OPA837
Other Parts Discussed in Thread: OPA810

Hi,

I have a rather general query regarding offset voltage and bias/offset current specifications, given in datasheets. But for reference I've mentioned OPA837.

I was doing some error analysis on a circuit, containing couple of op-amps and few other stuffs, using propagation of uncertainty principle. However, I'm somewhat confused about the nominal value consideration of the above parameters (offset voltage and bias/offset current of the op-amps). I understand that the typical values, given in datasheet, correspond to the 1 sigma value. However, is it appropriate to assume these parameters as zero mean random variables? 

The question I ask because typical schematics often show these variables have zero mean and normal distribution. However, in some recent datasheets (like OPA837), the offset voltage or offset current distributions (histograms) do not have zero mean. Intuitively, considering the mean as the most-likely value of a random variable, is it not true that zero offset voltage or current is unlikely (due to device mismatch)? Thus, choosing zero-mean could be contradictory, I feel. As a result, what should be the correct assumption of the the mean (and standard deviation from the mean) of these variables?

Looking forward to hearing from you. 

  • 0
    TI__Genius 15555 points

    Hello Tapabrata,

    You can use the distribution charts for greater insight into the mean and deviation of these parameters. Some datasheets (such as OPA810) offer you these values along with the figure. Note that devices are tested to ensure that these parameters fall within the maximum and minimum values given.

    Best,

    Hasan Babiker

  • 0 in reply to Hasan Babiker31
    Guru 47485 points

    I did the OPA837 plots, 

    Yes - the offset voltage and current are mean zero

    the bias current mean is not of course

    More interesting is the drift histograms - those were screened in test as well (maybe not still) where the Ios drift had some really odd outliers - so the histogram limits on the drift plots are relatively hard limits. 

  • 0 in reply to Michael Steffes
    Prodigy 30 points

    Thank you very much for your replies Mr. Hasan and Mr. Steffes. I would like to conclude this discussion asking another general question. This question arises as not all datasheets have the distribution charts.

    I understand that if distribution charts are not available, the offset voltage and offset currents can be assumed to be zero mean and the specified typical value of these parameters as the standard deviation. However, parameters which are not zero mean, like bias current spec., will it be reasonably correct to assume the typical value as the mean and also same amount of standard deviation. For example, the bias current of the popular OP07 op-amp is specified as +-1.8nA (at 25 degC) and no distribution chart is provided. In this case, can I assume the mean value to be 1.8 nA with a standard deviation of 1.8 nA? 

    To clarify further, I would consider another different parameter. The drain-to-source on-resistance of a multiplexer IC (CD4053)is specified as 470 ohm (for 5 V supply). In this case, it is correct to assume the mean value of the on-resistance as 470 ohm, with a 1-sigma deviation of 470 ohm, as well. 

    I understand my question deviated from the initial product consideration (OPA837). However, I did so to completely clarify my doubt. Looking forward to your replies. 

  • 0 in reply to TAPABRATA SEN
    Guru 47485 points

    Well it is difficult to comment for all parts across the different groups and timeframes in which they were done - 

    For specs like the input bias current on the OPA837 for instance, that is just the mean it is unipolar - I am often suspicious of +/- specs on input bias current - correct sometimes, not always. 

  • 0 in reply to Michael Steffes
    Prodigy 30 points

    Thank you very much for the clarifications.