Q: COM Objects and Unicode Problem

[Tim]

Here on our project, we have an ASP based application that works
perfectly with UTF-8 Unicode. The application calls several COM
objects (made in Delphi), but they are "known-good" as they've been
tested via ASP, and via internal files and database testing though
even Window's apps.

But when we call those COM objects from our new DOT NET apps, the
strings passed into the objects are apparently non-UTF8.

Multi-byte UTF-8 characters are received as single byte garbage
characters. We know this because the COM objects log all incoming
parameters to text-files (and we know that all other non-DOT-NET apps
have valid items in their Log files).

The STRANGEST part is if we send the strings to the COM object and get
it back (without processing really) to the calling App, the App can
show it on the screen as valid information.

What causes this behaviour and how can we send "real" UTF-8 data from
DOT NET?