ALl right, figured out the problem right after I sent the question.  It's an
ASCII encoding issue.  ASCII encoding behaves differently in 2.0 and 1.1,
not the hashing itself.

The problem is that your code is broken - it's converting from
arbitrary binary data to a string using an ASCII encoding. What do you
expect it to do when it comes across a byte outside the ASCII range
(i.e. anything over 127)?

Here's a program which demonstrates the problem:

using System;
using System.Text;

class Test
{
   static void Main()
   {
       byte[] data = new byte[]{140};
       string text = Encoding.ASCII.GetString(data);
       Console.WriteLine ((int)text[0]);
   }
}

Basically, you were relying on unspecified behaviour, and it's changed.
Now as to what you can do about that - the easiest thing would probably
be to emulate the previous behaviour. The simplest way of doing that is
something like:

   static string OldBytesToAscii (byte[] data)
   {
       char[] c = new char[data.Length];
       for (int i=0; i < data.Length; i++)
       {
           c[i] = (char)(data[i]&0x7f);
       }        
       return new string (c);
   }

A better solution for moving forward in the future is to base64 binary
data when you need it in a reliable text form.