With the R&S RTB2002, RTB2004 I needed to check the Frequency counter and Generator accuracy.
I purchased a Leo Bodnar Electronics MiniGPS 10 Mhz Reference clock. I used the output to connect to one channel 50ohm feedthrough termination and used this as the reference for the RTB trigger and counter measurement.
The second input was connected to the Generator output of similar amplitude and set the frequency to minimal difference. The specification of the RTB is +/-2.5PPM which should give at 10Mhz +/- 25Hz range. The frequency readout of the Trigger Counter was within 5digits +/-20Hz.
After a bit of trial and error, I determined the RTB was approximately +15Hz, and stable within 1Hz over a couple of daily temperature and power cycles. This was very similar to "Peter's" +12MHz. So it would seem adjusting the RTB timebase should be possible to achieve +/ 2Hz short-term calibration checking against the GPS reference.
I also found that the Generator Frequency can be adjusted and displayed (when below 10MHz) to within 0.1Hz i.e. 9.99999855Mhz, above 10Mhz seems as if this is not so easy as while you are able to input the figures and Enter, the display is rounded to nearest 1Hz i.e. 10.000001. As the RTB reference was measuring high the generator needed to be lower to compensate to give zero or slow movements between traces, so it was an easy task to accurately measure the frequency error to with 0.1Hz.
Using the R&S remote LAN interface and sending "DEV:MENU ADJ" on the SCPI device control, I was presented with the Frequency Adjust option menu on the RTB screen. The default on mine was also "128". It appears the sensitivity is about 1.33Hz per unit, so I set the TCXO offset for 117 and pressed SAVE, a lower number decreases the reference frequency. I was then presented with an RTB timebase frequency error <1Hz using my 10Mhz GPS-derived reference.
Thanks for the previous poster Peter for the "undocumented" feature allowing the "simple" procedure to fine-trim the RTB timebase (and generator output frequency) within <0.1PPM. It would appear from the factory they are within the +/-2.5PPM spec so are not adjusted from default to give their best.
The remote frequency trigger reading "TCO:RES:FREQ?" can also give an extra digit to 6 (a single Hz below 10Mhz or 6 digits above).
I'll monitor the long-term frequency drift with the GPS 10Mhz reference and adjust the RTB as necessary, hopefully, easily maintaining the <0.1PPM frequency accuracy.