I'm designing a differential driver for 50 ohm loads  in order to characterize phase noise/jitter. (20ps rms jitter 100Mhz - 1Ghz output)
I've chosen a differential topology to make sure noise in ground loop/common-mode noise doesn't affect measurements.
While differential probes are often used for scopes (Lecroy, Tektronix etc.) for jitter measurements,
the same is not true for spectrum analyzers.

1. Spectrum analyzer probes are generally single-ended. Why? Does this not affect measurement accuracy?

2. Is it advisable to use a balun to obtain single-ended output for phase noise measurements?
Would it be better to have a separate single-ended output from the chip for this purpose?

Thanks.