Hi James,
sorry, I have no reference for electronic components but this function
is well known for creep (relaxation) of (polycrystalline) piezoelectric
materials; it should be easy to find references for this. I guess that
the underlying physics is comparable in the two cases: A sum of
exponential relaxation processes with a many different time constants.
In your current case, the fit function won't describe the drift for
large times: Eventually, you should reach thermal equilibrium, and there
should be no drift any more, but the y = a + b*ln(x + c) function will
continue to change forever (though at a rate that decreases with 1/time).
So far my 2 cents,
Michael
____________________________________________________________________________
On 07/08/2025 03:18, James Ewing wrote:
> There’s a function suggested in Image->Stacks->Plot Z-axis Profile,
under the Data/Add_Fit button that I’ve found to be very useful for
fitting data in MRI studies that have a receiver gain drift due to
heating of some active components in the receiver coil. The function is
y = a +b*ln(x +c). I’ve been looking for a reference (citation) to this
function’s use, but haven’t found anything specific on the web. Can
someone tell me where this function was suggested as an heuristic
function for fitting heating of electronic components?
>
> - James Ewing
--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html
