Not an answer to your question - however a great place to find PRJ-files for different coordinate systems is: http://spatialreference.org/ref/?search=wgs84&srtext=Search

---

These semiminor axis numbers are all basically the same except for level of precision or rounding. 

 

 

EPSG does not explicitly define semiminor axis for ellipsoids; instead, it provides an 'inverse flattening' that the semiminor axis is calculated from. The number of digits included in that value is fairly arbitrary.  I don't believe the added precision (tenths or hundredths of millimeters) will meaningfully impact transformations. 

 

 

Note that depending on the tool, sometimes the number of decimals displayed in a coordinate system description is less then the number of decimals actually used. For example, in the FME coordinate system description it looks like we are using a PRAD: P_RAD: 6356752.3142, which is only 11 significant figures. However, we also report an inverse flattening of 298.257223563, which is 12 significant figures. So I suspect we are using a P_RAD that is more precise but just not displaying the full precision.

 

 

For more on precision in FME coordinate system transformations see:  http://fmepedia.safe.com/articles/FAQ/Solution-Accuracy-of-Coordinate-System-Conversions

 

---

Thanks for the answer, this clarifies.

 

 

Indeed it shows that the difference between the values is not the problem. The only problem I had then was that ArgGIS gave some error messages when importing the files with different elipsoid parameters. As a solution I just redefined the coordinates systems of the files as being calculated with the ESRI values.   

 

---

If you problem is ESRI's recognition of your data's coordinate system, fmepedia has a good article on this: http://fmepedia.safe.com/articles/How_To/ESRI-Coordinate-System-Exceptions

 

 

Especially review the considerations section.

 

 

Exceptions are coordinate system name associations that help us link FME coordinate system names with names used by other tools such as ArcGIS. The exception can also define what the WKT string is that FME writes as part of the coordinate system description. So if if ESRI is expecting 6356752.314245179 for prad, then you can add an exception and use that value in the exception string instead of the default value used by FME. This would save you the effort of crafting a new coordinate system just to cover mismatches based on fine differences in precision.

---