Here's a solution to J. Penry's Geodetic Math Problem that does not require state plane calculations nor ellipsoidal calculations. Instead, it assumes that lat/lon are plane coordinates and accounts for convergence by multiplying angular (arc) differences along parallels by the cosine of the latitude.
Sort of a poor man's LDP, I guess. Perhaps it's not the most sophisticated image, but sometimes it does one good to do something by hand.
[sarcasm]the quality of your north arrow indicates your calcs are probably questionable....[/sarcasm]
It is puny, isn't it? It wasn't used in the calculation, so perhaps it didn't distort the results too much.
There is a typo (??) in one of the descriptions. Angle DAB is Arctan(DB/DA), not Arctan(DB/AB).
Sometimes it's good to think outside the black box.