...under perpetual construction.
The RF output of the HP 8657B signal generators I use is not particularly clean; apart from the close-in spurs and phase noise there are of course also output harmonics that sometimes can spoil the measurements results. Here below you will find some measurement of the 8657B output harmonics and how to clean it up with a simple low-pass filter.
The following graph shows the output power at 10 MHz for the fundamental and its first harmonics for different output levels (note that the output went thru a combiner so the actual output power is about 3.6 dB lower than the programmed power): or, viewed as relative amplitude of the harmonic with respect to the fundamental: note the characteristic shape of the harmonics relative level vs. the output power; as explained in the instrument manual, the internal ALC loop is used to change the output level in a 5 dB range, while larger level variations are done with the internal step attenuator. This implies that the output power amplifier operates always with an output level that does not change more than 5 dB, so the output harmonics levels changes only in the narrow range when the ALC loop is used to control the output amplitude.
The effect of the generator harmonics on a wideband power amplifier characterization can be seen in the following graph; at low output power the measured harmonics are actually these coming from the generator and not from the PA itself!