I've chosen for nchar(1) because it only says that the value won't be NULL or empty, but there's no limitation on the characters used. Also depending on your collation sometimes a character is covered by char but in another collation it's a nchar.
Try this:
CREATE TABLE #Chars(
c1 char(1) COLLATE SQL_Latin1_General_Cp850_BIN,
c2 nchar(1)COLLATE SQL_Latin1_General_Cp850_BIN,
c3 char(1)COLLATE SQL_Latin1_General_CP1_CI_AS,
c4 nchar(1)COLLATE SQL_Latin1_General_CP1_CI_AS)
INSERT INTO #chars
VALUES(CHAR(0128),CHAR(0128),CHAR(0128),CHAR(0128))
SELECT * FROM #chars
DROP TABLE #chars
As you can see the nchar value is consistent, while the char value depends on the collation.
[font="Verdana"]Markus Bohse[/font]