swept sine seems to have incorrect amplitude and phase
When doing a simple swept sine against itself, I'd expect a transfer function with an amplitude of 0 db and 0 phase. Instead, amplitude is proportional to excitation amplitude. 1000 on the excitation is worth 60 dB in the transfer function.
Phase will show a strong trend (though not strictly linear), but whether the trend is positive or negative varies from one run to the next.