Bob, I know what happened, it was a segment problem.
When I later added a matching network at the center of the 2 x 1/2 waves model to see if I could correct the mismatch...the segments were set at 27 for wire #1, roughly 16" per segment. This was the only wire in that model and that is pretty low for that a full wavelength at 27.205 MHz.
I increased the segments by close to 3 times. So, I think your point was the segment length was at play with the model, and you were correct.
I tried talking to DB about this same issue a while back but instead, we got into a discussion about his use of auto-segmentation usin4Nec2 and my idea never got to 1st base. IMO, a proper wire segment count selection is vital to an accurate model and to a large degree it is possible the Average Gain Test in Free Space results will predict for the modeler when a model is accurate or very close.
I don't know exactly how auto-segmenting works with 4Nec2, but DB has told me he uses the feature.
In the past when I've talked to Roy Lewallen about the importance of currents in a model...he told me that he only discusses potential bug type problems with his clients. I understand maybe why that is the way it is.
However, he did warn me to always keep and eye on the currents in the Currents Log Report. I've posted about this before. He also told me NOT to depend on the "Conservative" or "Minimum Recommended" selections using the automatic segment selection feature...he said these two selections were intended as relative indicators to be used in quickly checking segmentation only. He further said sometimes the feature works and sometimes it doesn't work positively for the results. At that moment I had no idea what it all meant, but over time it seems to make sense to me because now I watch for the problems.
This is why I recently started modeling in Free Space and then converting the model to real Earth, so I could use the AG test as intended. It is remarkable how much difference a model can develop with changes in the segmentation. And without the AG Test, I don't know how one could ever tell the accuracy for sure within the scope of the software. Although, in the real world only being able to duplicate result in a real world test should count as positive, and that should go without saying.
Question is, have my models changed since I started this new way of building a model? Yes!
Change in a few models was a bit more dramatic than others, but I did see a difference and my old way of making the wires with more than one segment close to = 3" per segment or = 6" per segment. In the scheme of things, it worked out I was using too many segments for my models.
Bob, you have a surprising grasp of the segments and/or currents ideas, but you would really have to be there. Too bad you didn't take my offer to join me in trying to get a working grasp on modeling some years ago...you would be good at it...better than me I'm sure.
...runna pile-up on 6m SSB