In the past I have primarily used real/average ground for the models I have posted unless I have stated otherwise. More recently, since the last models I have posted here or anywhere, I have played with various different ground types and, just as you have above, noted some differences. Which one is more accurate would depend on the ground conditions where you are transmitting/receiving from, and honestly, most people really don't have the ability to measure such things themselves...
Her is another example to add to yours of a 0.66 wavelength vertical over four 0.25 wavelength radials. Yes, this antenna was intentionally modeled to be longer than the mythical 0.64 magic maximum electrical wavelength... A note, the grounds used here are real/good, real/average, and real/poor.
Something to notice when it comes to the lobes of these three identical models aside from the grounds they are over, average ground has the lowest gain, and strangely, average ground is the only model here that has a higher angle lobe that is more dominant that the lowest angle lobe... I guess there goes the 0.64 wavelength antenna being the absolute best low angle gains in all circumstances as the model shows it is possible to exceed that length in some cases depending on the environment the antenna is in. I think it is worth noting that for some ground conditions that you could make the antenna even longer than this while still having a dominant low angle lobe, but that is another discussion...
It is also worth noting that the good ground model above has two lobes, while the average and poor ground models of the same antenna the same height above ground have three.
Anyway, what I would recommend is a de facto standard that we modelers can use for said ground for posted models, I would recommend the real/average set as that is what I tend to use unless stated otherwise. Or we can opt to just note it every time we post a model. I'm pretty sure Marconi also uses average ground of some type for the most part as well with his EZnec models.
I do have to agree that using average ground does not necessarily produce models with an average pattern compared to all of the other possibilities, but I don't think that was the point. I was going to say here that average ground has an approximate average of conductivity and dielectric constant for ground, but that is only half right. Only the conductivity is near average, the dielectric constant variable used is more than twice as large as the good and poor ground types... I wonder why that is.
I, personally, have no problem with using whatever standard or method we settle on. Otherwise I will simply note the ground used for whatever model I happen to make in the post with the model...
The DB
