Why am I explaining this again? With only six, the transospark oscillations are too perfectly synced. Any bias would be the result of rounding errors, not input.