Tom wrote:

I selected a
westerly heading since there would be no turning error on the compass.

The idea of setting the course 180 *is* exactly the turning error! Every
tendency to bank will immediately be displayed by the compass, even
before the plane begins to actually turn. It actually works very much
like the "turn coordinator". (BTW, this works as well on course 0, of
course. But because the compass will react in the wrong direction, a
correct reaction would require a very high level of abstraction.)

Also, I read the accident report sited by Stephan and found no possible
connection to the T&B mentioned earlier.

The conection is that you have an instrument in front of you which
reacts differently than you expect. When you're trained with and used to
an artificial horizon, then that trutrak thing looks ecactly like one.
So I expect it to display pitch information which it doesn't. And when
you've learnt to rely on a horizon as your primary reference for
attitude, then the issue is no more minor, but may become that last
straw which breaks the camel's back.

I stay with my statement: Something which looks exactly like a well
known and widespread instrument but works differently is a possible
source of confusion and a very bad idea, especially when it comes to
such an unforgiving thing as flying in clouds.

Stefan