The aim of the null test branch was to fix the large error bars in the null test. It turned out that these resulted from the global channel_map not being accurate. The sats seems to switch channels which made the null test script put a lot of noise into the maps. I fixed this by completely changing my method of identifying sat channels. Now, each 30 min obs had its own channel map. This was also implemented when making tile maps, which resulted in a huge improvement.
The aim of the null test branch was to fix the large error bars in the null test. It turned out that these resulted from the global channel_map not being accurate. The sats seems to switch channels which made the null test script put a lot of noise into the maps. I fixed this by completely changing my method of identifying sat channels. Now, each 30 min obs had its own channel map. This was also implemented when making tile maps, which resulted in a huge improvement.