This method seems a little dangerous as I can envision where the string "0a" could occur in your hex string by accident, for example the two byte hex "b0 ac" represented this way would be an invalid match.
Here's a function implementation of your approach, not using hex codes. You could paste this function into the ArcMap python window and use it there. Note I'm using chr(10) and chr(13) for "\n" and "\r" so this function can also be used inside the Calculate Value tool in modelbuilder... as the usual use of "\n" breaks the geoprocessing messaging string representation in the code...
I'm also using arcpy.da.UpdateCursor because it is very fast compared to the 10.0 flavor.... and the "with" construct helps you out by closing the cursor even if the code fails -- avoiding the possibility of a nasty hanging file lock!
import arcpy
def strip_newlines(tbl, field, eolchar=""):
with arcpy.da.UpdateCursor(tbl, [field]) as rows:
for row in rows:
row[0] = row[0].replace(chr(10), eolchar).replace(chr(13), eolchar)
rows.updateRow(row)