help on Agilent's 8560B/8560E spectum analyzer

@gilent's 8560 series has B and E types. The E series is suggested for VCO test and has a fine phase noise from 100Hz to 100KHz(-118dBc/Hz@100KHz) with a wonderful ocxo( 1e-7 aging/year), while the 8560B series is very old and has a worse phase noise (-100dBc/Hz@30KHz). Of cource the 8560E seies is much more expensive. My problem is whether the 8560B would have a similar phase noise if a wonderful external referece clk(such as 1e-7 aging/year) is used. Does someone answer my question? thanks.

well, i'll give a general opinion even if i've not experienced differences between 8560B and E. I've both SA.
8560 series is syntetized. If opt 001 is installed you have an expensive ovenized Xtal that increase both short time and long time freq. stability. Short time freq. stability is strongly related to close to carrier phase noise. So using an ovenized Xtal or rubidium or gps timing will reduce slightly the close to carrier phase noise.
But 30 or 100 Khz far, are too far frequencies. At that distance, the PLL circuitry and multiplication factor is dominant.
In conclusion:
for my opinion:
1) 118dBc/Hz(at)100KHz and -100dBc/Hz(at)30KHz are very similar
2) Replacing the Xtal with a more stable don't change the phase noise at 30 Khz far but reduce the phase noise in the close in region (i.e. 10..100 Hz far)