Thanks for the clarification on the OBC file. Makes sense. Also, I eventually found the SeaDAS/System Info in the OCSSW tools, thanks.
Yes,I must have had a typo when I transcribed the pixels/lines from the lonlat2pixline output on the L1B to the extractor gui the other day. When I did it again it was the correct region. Incidentally, is there a lonlat2pixline that works on L1As? Something like what the L1A extractor runs for the gui, but standalone for command line?
Nevertheless, the extractor bug (i.e., not propagating to Pixels from Lon/Lat) was reproducible when Bing and I worked on it offline on Friday. This morning I tried running the extractor >before< opening the L1A in SeaDAS, and it worked. Then I opened the L1A in SeaDAS and tried the extractor and it did not work. Therein lies the bug.
Dirk