Far from expert here, but some thoughts offered in deference to those more knowledgeable who will be along later.
At first thought it would seem to be a defect in the lock geometry - not much you can do about that and anyway seems unlikely .....
Second thought - does it make a difference which way round the flint is fitted (chamfer up or chamfer down). I've seen both orientations used and I'm thinking this will change the interception angle and might change the trajectory of the sparks?
Third thought - setting the flint a little lower or higher will alter the point at which it strikes the frizzen and would slightly change the timing of the lock, which might change where the sparks end up ?
Fourth thought - was the frizzen simply re-hardened or was the face first built up to compensate for wear (which would of course affect the lock geometry) ?
By the time you have tried all these the frizzen will need re-facing again.....
True freedom includes the freedom to make mistakes or do foolish things and bear the consequences.
TANSTAAFL