So I did some tuning accuracy tests on this 8569B and I'm impressed; it actually sits inside the specifications in the manual. I unfortunately don't have a microwave source yet, so I could only test the 0.1-1.8 GHz band with my RF generator, which is a limitation, but this is the band I'm using most at the moment anyway!
The manual states that tuning accuracy for the instrument's entire band (0.1-115 GHz; external mixing is required over 22 GHz) is +/- 5 MHz or 0.2% of center frequency (whichever is larger) +20% of frequency span per division.
My results are as follows. All measurements were taken with 500 KHz/div coupled 100 KHz RBW and no internal attenuation.
Internal 100 MHz calibration signal:- Center Freq: 100.0 MHz
- Front Panel Readout: 0.100 GHz
- Actual: 100.0 MHz
100 MHz signal from my 8657A @ -10 dBm, no modulation (which is bang on accurate according to my refurbed Tek 2465B):- Center Freq: 100.0 MHz
- Front Panel Readout: 0.100 GHz
- Actual: 100.0 MHz
Down in the low part of the band the SA is bang on.
1000 MHz signal from the 8657A @ -10 dBm, no modulation:- Center Freq: 1.0000 GHz
- Front Panel Readout: 1.000 GHz
- Actual: 1.0015 GHz
So up at 1 GHz, it's off by almost exactly +1.5 MHz. 0.2% of 1000 MHz is 2 MHz < 5 MHz, so it's easily in spec according to the manual, adding the 20% of [edit]
RBW freq span/div is irrelevant in this case.
Now I realize there are probably way more precise ways to do this (and there are probably variables I'm not accounting for and it's likely farther off in higher bands...), but working with what I have, I'm quite impressed for the price and 34 years old.