LMH6702: Information about Noise versus temperature for LMH6702

Hello Quyen,

   We do not typically characterize noise over temperature, so we do not have the data readily available. There is however a way to estimate the noise at your desired temperature. The data recorded for noise is characterized at room temperature (25 degrees Celsius). In this case, the broadband noise will be the significant source of noise as well as being synonymous to thermal related noise. At each temperature, the equation with its respective values would be roughly the same (would be a minimal difference in transistor/bandwidth values between 25 and 80 degrees); the most significant value change will be T. Thus, you could estimate the respective noise components through taking the factor of sqrt(353.15K/298.15K). The estimated factor results to 1.088. Voltage noise, inverting noise current, and non-inverting noise current at 80 degrees Celsius would estimate to about: 2nV/sqrtHz, 20.13pA/sqrtHz, 3.27pA/sqrtHz at greater than 1MHz. 

Thank you,

Sima

Really? 

I suspect underlying this question is also a question of bias current shift over temp. The Current noise terms are RMS of the up down pnp/npn transistor currents which do shift over temp. For those terms, the DC shift over temp also gets into this question. In the simplest input class AB buffer stage at the V+ input, the input noise current is the RMS of the two base currents - those DC currents will shift over temp due to beta tempco also contributing to the over temp noise number. The VIP10 process used for the LMH6702 will show a beta that increases with temperature -hence the V+ current noise should have a negative tempco for input bias current due to that term, partially offsetting the abs. T effect, Here, the noise terms show a much lower V+ current noise than the inverting current noise - that indicates a base input buffer stage at the V+ input, 

Hello Michael,

    You are right; it is a very rough estimate due to not having the data readily available as well as the lack in ease of acquiring the transistor related value changes over temperature that you have listed. If it helps to get a sense of these values, I do have the input bias inverting/non inverting values over these two temperatures: IBN(uA): at 25C -8.22, at 85C -6.14; IBI(uA) at 25C -5.53, at 85C -6.02 (current flowing out of the device). 

Thank you,

Sima

It is likely the original design review will have plots of input noise terms over temp and process, if you can find it. 

Hello Michael,

   I did happen to find the original design review and various characterization plots. Unfortunately, none of these included information on input noise terms over temp and process. 

Thank you,

Sima

Those darn National guys, the BurrBrown equivalent is the OPA695, if that showed the spread over temp and process, at least you could get an idea from that device. 

Hello Michael,

    We did not find any noise over temperature/process plots for the OPA695 in the design review and characterization sheets. Seems to be a common issue for these devices.

Thank you,

Sima