Here's those SA screenshots I promised a couple of weeks back.
Hendorog was kind enough to run my SSA3032X back and even grabbed them for us while we were having a good chat. Anything else you need please ask.
These are are pretty good results. There are 1 MHz offset spurs on the 40 MHz plot. Where could these be coming from? Perhaps they would go away with an external 10 MHz reference?
Based on my understanding of the device errata, I don't think that the measurements would tell us anything. It says that there should be spurs at f_data +/- f_out. f_data is 1.2 GHz, I think? f_out is your carrier frequency, so we really should be looking for signals around 1 GHz. However, these would be filtered out in the analog front-end. We might be able to see mixing products ((f_data + f_out) mixing (f_data - f_out)) which would result in a spur at 2*f_out, and more generally the even harmonics. Though, my feeling is that non-linearities in the PA would add to the THD much more than these spurs. So, I don't think that we should be able to see any effect.
To measure the above, we'd probably need to probe before the filters/PA. The spurs might be independent of the signal power, so it should be measured in the bottom of an amplitude range (to reduce the relative power of the desired output signal). The measurement probably should be done with a 800 MHz or so high-pass filter, or a notch filter to block the programmed output frequency. There may also be an effect in the phase noise of the output signal (or the phase noise of the 1.2 GHz spur).
All in all, I don't think that it will have any real-world effect on instrument performance.
I'm much more interested in UI improvements/firmware updates (I noticed that the SDG6000X has separate 10 MHz ref in and out ports... I like that. (Another aside, why don't instrument manufacturers label the ports with their impedance and signal levels? I don't like having to dig that information out of the datasheet)).