Predicting parts performance in space depends on laboratory measurements of heavy-ion device cross sections. Analysis of this cross section data has been a major source of error until now because everyone does it differently. Due to a breakthrough in SEU analysis techniques by Sandia Laboratories, Space Radiation can now give you the best set of standard model parameters that can be derived from your Brookhaven or other heavy-ion data.

You create a file containing the particle LET, angle of incidence, measured cross section, and cross section error. Space Radiation analyzes the file and determines the best values of sensitive volume dimensions, LET threshold, funnel length, and Weibull parameters according to the standard SEU model in common use today. It even tells you if the standard model doesn't apply to your device.

As extended output you get a listing of the normal- incidence cross section most commonly used in SEU data presentation, as well as an omnidirectional cross section completely compatible with the "effective flux" approach used by NASA.

What are the advantages of HITFIT?

1. It provides the best representation of the available data in terms of the standard SEU parameters and model.
2. An overall goodness-of-fit measure is available. If the fit is poor, various problems could be indicated. Some possibilities are typographical errors; data analysis errors; failure of the test setup; failure to correct for important factors such as a passivation overlayer, multiple event upsets, and shadowing by the package; or a breakdown of the standard model.
3. No guesswork is required to determine the parameters, especially the depth, often taken arbitrarily to be 1 µm.
4. The technique returns estimates of the uncertainties in each of the SEU parameters.
5. Complete self-consistency between the determination of SEU parameters from test data and the prediction of SEUs in space environments can be guaranteed because the same approach is used in both cases. This is of particular importance with the LET threshold and funnel length which have various interpretations in the literature.
6. No need for the approximate "effective LET" concept frequently used in test data analysis.