I'm trying to write a function that strips out all non-ASCII characters in a field when I select it, but I'm having trouble with the pattern matching. I've tried this (using the collation statement to remove the diacritics):

Code:
Create Function dbo.udfRemoveTroublesomeCharacters   (@IncomingText  VarChar(1000))
Returns VarChar(1000)
As
Begin

    Select  @IncomingText   =   @IncomingText  Collate SQL_Latin1_General_CP1253_CI_AI
    Declare @KeepValues     As  VarChar(50)
    Set     @KeepValues =   '%[' + CHAR(32) + '-' + CHAR(126) + ']%'
    
    While PatIndex(@KeepValues, @IncomingText) > 0
        Set @IncomingText = Stuff(@IncomingText, PatIndex(@KeepValues, @IncomingText), 1, '')
    
    Return @IncomingText
End

Go
However, when I test it using this:
Code:
Select dbo.udfRemoveTroublesomeCharacters('ars;tlkjDSFGHSDFGÉ hq4é59’87nq34[5vui$"£%^"$%^"(*345v')
It returns this: arstlkjDSFGHSDFGEhq4e59’87nq345vui£345v

As you can see, that still allows the character ’ to come through despite the fact that its ASCII code is 146. Conversely, it removes something like ; which has an ASCII code of 59.

Short of individually listing each character that's allowed, how can I get it to stick to the range I've asked for?

Thanks...