Discussion of precise time and frequency measurement
I asked this very question a few years back... the consensus was to lower power dissipation into the typical 50 ohm load.
