MPPT reduces energy production?

I developed a simulation that used a single inverter with a single MPPT tracker input.  I had four strings of seven panels, all in the same subarray (so all on the same MPPT).  I used the 3D shade calculator to simulate shading effects.

The results of this simulation produced 9968 kwH in year 1.

I expected that if I added more MPPT trackers and split them up among the strings, it should improve energy production.  However, when I do that, the simulation actually has WORSE performance.  This seems counter intuitive.  I've tried it with one, two, three, and four MPPT trackers.  Each time I add an MPPT tracker, the energy production goes DOWN.

What am I doing wrong?

One thing that I did notice is that the losses graph shows that the "AC Inverter Power Consumption" goes up each time I add an inverter.  Does it really take that much extra power to run an MPPT tracker?

I also noticed that the shading loss goes UP when adding MPPT trackers. 

Not sure if it matters, but the inverter that I'm using for simulations is the <>.  I haven't necessarily decided that this is the right inverter - I'm just using something for now to get rough numbers so that I can decide how many MPPT trackers I need in my inverter.

MPPT    |   Annual Prod (KwH)  |  AC Inverter Power Consumption Loss (%)  |  Shading Loss (%)
     1       |              9828                 |                                  0.67                                      |           7.461
     2       |              9794                 |                                  2.64                                      |           8.052
     3       |              9587                 |                                  4.26                                      |           8.608
     4       |              9415                 |                                  5.88                                      |           8.81

Hi Ron,

I can't tell exactly what is happening without seeing your file. However, I suspect the problem may be that SAM is applying the shade losses calculated by the 3D Shade Calculator only to Subarray 1. When you changed from a single MPPT to four, did you adjust the properties of the active surfaces in the shade calculator scene to indicate which active surface belongs to which subarray?

I'm happy to review your file if you don't mind attaching it to your post.

As for the inverter consumption, I see that difference too, and am investigating and will follow up when I learn more.

Best regards,
Paul.