Back in the day, we occasionally needed to measure the return loss of the antenna bays of our TV site, over their designed range of frequencies. I tend towards Kalvin's suggestion of interference from a strong external carrier. Compare the two- does it confirm the point made by radiolistener? Another test you can try, this time on the nanoVNA, is to use it to check the SWR with a 75 load, then a 150 load They should measure 1.5:1 & 3.0:1 respectively. Have you done the full check, which entails also checking the transmitter plus SWR meters into the dummy load? Also, try checking the SWR of your antenna at full power, if your SWR meters are rated for it.(do it fast, as other hams may look askance at a massive carrier up for too long ) Write that down, then check at the lowest power level you can reduce the radio to.

This is why VNA measurement is more correct and more precise than scalar SWR meters. Also, VNA meters don't have diode issue (at low power output). In opposite, VNA meters show you SWR value at specified frequency only. It show you average SWR value for all components in the signal. Another issue with scalar SWR meter is that it shows SWR value for entire bandwidth, it don't take into account that your TX has harmonics and out of band emissions. Almost all scalar SWR meters shows wrong SWR value at low power output. Just because not enough power output was used for measurement. But your SWR meter will show you SWR=1.0. You will see that your SWR meter will show you SWR=1.0 at low power. Just connect 75-100 Ohm resistor as a load on SWR meter and try to measure SWR at different power level. If you don't believe me, you can perform simple experiment to make sure that this is the truth. It is small, because output power is too small. Just because amplitude taken from a coupler is too small for diode. This is why it will show you SWR=1.1 while in reality it is SWR=2.9. The diode has limited working dynamic range and non-linear response at low voltage. Most of scalar SWR meters using diode as amplitude detector for a coupler output. Most of all because you're using too small power for SWR measurement with your SWR meter.

Nanovna swr