After rewriting my code to be less sensitive to blocking everything is working. I broke up the processing in signal acquisition, number crunching and display which is much simpler than keeping all that going at the same time. For signal acquisition I am experimenting with sample size less than a full revolution of the antenna to keep update times reasonable.
Now that I have a compass sensor working I wonder if I can improve on the +/- 1 ° or so from the compass by using GPS when in motion. The GPS "course over ground" which when in motion gives direction based on position differences could be more accurate. Ideally I could switch navigation source.
I have been using "RDFMapper" which displays a google map and take GPS and RDF input several ways based on settings. It draws a bearing line from the location of the direction finder which can be relative to the direction of the vehicle or a fixed specified angle from north.