300 Hz (corresponding to alternator speed of 6000 rpm, or engine speed of 2000 rpm)
As you can see, there is some clipping going on on the upswing of the waveform, as the on-board charge pump takes voltage from the diode drop of the MOSFET. Once the controller has built up its charge, it will apply Vgs to the gate of the MOSFET, causing it to conduct. However, this clipping really doesn't affect the waveform that much, which is why I suspect TI used that 300 Hz figure.
Precisely when the MOSFET starts to conduct, is another matter entirely. Some MOSFETs have V(gs-threshold) of 2 VDC, 3 VDC, and so on. The MOSFET I'll be using has a V(gs-threshold) of about 1.6 VDC at room temperature, and will drop to about 1.25 VDC once it heats up to 80 C inside the engine compartment.
This is what is likely to occur at 1200 Hz (corresponding to alternator speed of 24000 rpm, or engine speed of 8000 rpm)
Like I said, the waveform itself is going to start being distorted. Might be good to introduce a capacitor to filter out the added high-frequency noise associated with this distortion.