The only part on my Sigma4 model that shows to be out of phase with the rest of the radiator...is the first 4 x 7.15" inch segments at the bottom around the area where the gamma would be.
For some reason however, my radials show a little more that half of the radial, from the bottom up, are in phase and the rest, up to the hoop, are out of phase and I don't understand that.
Well, I took 106" for the radials and figured the radiator at 3.2 times 106": 106" x 3.2 = 339.2" x 12 = 28.2666' for the radiator being the longest radiator according to the rule. I think I actually noticed a tip-over point, where the pattern skewed badly, at a tad less than 28.26' in the model however. Real antennas are different and modeling can do everything.
I posted the other day that I see a similar tip-over point around .625 wl for the 5/8 wave where the current maxima changes from its low angle of about 8-9* degrees to a higher 40* degree angle with the .64 wl model. Like I told 007 however, in that case the gain in the 10* degree lobe was only a fraction behind the primary lobe, so it was still good, but diminished rapidly beyond .64.
If the model of the Sigma4 don't work in this regard, I wouldn't waste my time trying to model that antenna in ARRL Antenna Handbook.
None of my models have had any kind of verification as to accuracy.