Aloha and thanks so much for the prompt response, Paul! Your plots helped me figure out what I was doing wrong with DISC input, and now the output seems more reasonable -- see initial sanity check attached.
One remaining issue is the time-step used for creating a TMY3 data file. I set up my site weather station to output 15-minute averages of GHI, and in a spreadsheet I average the four periods to get hourly data, then I use the ending time of the hour as the time-step (i.e., the average of 1215, 1230, 1245 and 1300 data is logged against hour 1300). Solar noon here runs from around 1238 to 1246, so I thought that particular time-step convention would be a good approximation, but every now and then DISC calculates a negative value for DNI. I suspect the issue arises because the time-step of my GHI database is not completely synchronized with the sun position calculations in DISC. I'll fool around with my database averaging times (or maybe just use the xx45 data) to see if I can get the negative numbers to go away, but if you have any thoughts on that please let me know.
The next question (re SAM required inputs) is how to derive DHI from GHI and DNI? Is DHI = GHI - DNI? Or is DHI = GHI - (DNI*cos sun angle)? Or something more complicated?
Thanks again for all your help!
Denning :-)