CSP IPH Model: Performance varying Collector Azimuth Angle

Hi everyone,
currently I'm performing some parametric analysis on my parabolic trough solar field varying the collector azimuth angle from +90 to -90° and trying to understand not only how the capacity factor varies but also how the generation profile varies during the year. In general going from 0° to 90° or from 0° to -90° the trend is clear: when the azimuth is 0° there is maximum annual production but concentrated in summer, viceversa for azimuth 90°, and something in between for azimuths between 0° and 90°.
Also the different daily profiles you get considering +45°(from South to West) or -45°(from South to East)  are clear, with a generation concentrated in the morning in the first case and in the afternoon in the second case.

What is not clear to me is the following thing: when considering a "real" weather file, like 1 year or a TMY, I understand that the case in which I have +45° Azimuth will perform typically slightly worse compared to -45° since during the morning we have more fog, clouds, temperature is lower and so on. However, when I use a Clear Sky weather file (e.g: I substitute the real DNI with the Clear Sky DNI computed from Meinel Model and I update the GHI accordingly), SAM results computed at +45° and -45° Azimuth still differ in terms of annual production.

This makes poor sense to me since with a Clear Sky DNI we should have perfect symmetry between morning and afternoon and the annual productions considering 45° or -45° of collector azimuth should be the same. Or at least, the yearly sum of the product between Clear Sky DNI and Cosine Collector Efficiency provided by SAM should be equal among +45° and -45° cases, which is not the case. And this applies for each couple +15° -15°, +30° -30°, and so on.

So I tried to compute using the formulas in the SAM Manual my own azimuth and zenith angles for each hour and the cosine efficiency considering 45° and -45° Azimuth and I tried again to do the yearly sum of the product between DNI Clear Sky and cosine efficiency in the two cases: the two sums were equal, showing that the two configurations are equal in terms of annual production as it should be (but I would be glad if you confirm it).

So I tried to compare my Azimuth, Zenith and Cosine Efficiency with those provided in SAM results (see attached excel file if you are interested, plus clear sky weather file and SAM file) and I noticed that while azimuth and zenith were equal (with differences up to 0.5° in each hour), the cosine efficiency was slightly different during central hours and with lots of differences at the start or end of the morning (pointed out in the excel), thus generating the above-mentioned difference in annual productions among 45° and -45° case.

So the final questions are:
- From a theoretical point of view, considering Clear Sky Conditions, having a plant operating at +45° or -45° should be the same in terms of annual productions, or at least yearly sum of product between DNI and Cosine Efficiency?
- If that's so, why SAM displays different annual productions among the two cases? (we are talking about a difference in capacity factor of about 1-2% in some cases using clear sky conditions)
- Why formulas on SAM manual give a different cosine efficiencies compared to SAM results?

Thank you in advance for your answer,
Marco