OK here are screen shots using a V 3 RTL-SDR and Spektrum. The antenna is not extended the same as when I did the test with the 8560A. So the dip is at a slightly different frequency.
Procedure:
Connect RF bridge as shown in the 8560A test with a 50 ohm termination on the DUT port
Set relative mode per the Spektrum UI
Remove 50 ohm terminator and attach antenna
The ripple is reflections at the connection to the RTL-SDR and the RF bridge. That is proved by the period of the ripples. The length derived from the 2-way travel time in the cable from the RTL-SDR to the RF bridge is 2.8 m which is the measured length of the cable. The RTL-SDR is not very close to the 50 ohms of the cable which is a good quality cable. As demonstrated in the 8560A case, the RF bridge is pretty good.
The ripples were 2-3 times larger before I tightened the SMA connectors with a torque wrench. I also put a 4 dB MiniCircuits SMA attenuator on the RTL-SDR to reduce the mismatch.
It's important to note that the amplitude response of the RTL-SDR is no where close to flat. So normalization (aka relative mode) is critical to getting a usable result.
Have Fun!
Reg
Here are the same tests using the V2 RTL-SDR
I don't know why the V 3 dongle is a poorer match, but I'd guess it's the bias tee.