Yet another highly scientific and thorough test from me.
I thought i do a quick height test, to see what we are dealing with.
Used 2 x RS, one as base with average base coordinates streaming correction LoRa data at 6dB (lowest power) and 1 Hz.
I think i used around 8-9 Kb/s bandwidth on the LoRa setting (sorry, forgot to take picture)
Another RS fittet to a DIY pole in kinematic mode with AR on both gps and glonas with LoRa correction input from base few meteres away.
I placed a A3 sheet with 1cm grid, leveled it with a laser and drilled 2 screws with 4cm height difference.
Moved the rover over the top on those two screws and measure for roughly 20 sec for each head. Area measure marked with a red circle
Result save from RTKlib as i saved output solution and can be downloaded below.
I reckon is about 5,2cm on the image, 12mm deviation. Satisfying when the accuracy is stated to 2 times horizontal value .
Her is output solution from this test. Have fun