Polar Diagram 1
This plot shows the polar diagram broadside on to the elements. The many nulls are typical of turnstiles and cannot be avoided. I tried to optimise for highest overall gain and least low angle null.
Polar Diagram 2
This plot shows the polar diagram end on to the elements.
Graph of VSWR.
As can be seen the reflector pulls down the dipole impedance at resonance from around 70ohms to nearer 50ohms. This would be good news were it not for the need to stack two of these to produce the turnstile. The reason I haven’t modelled the turnstile is simple – I don’t know how but I did measure the SWR of the completed antenna with an MFJ analyzer – around 26 ohms with an SWR of 1.9:1 across the band. As I’ve no guarantee that the receiver input really is precisely 50ohms I’ve decided to not let this trouble me overmuch.
I did model a unit with a higher impedance by using a variation on a folded dipole (a proper folded dipole being too tricky for me with manual bending of the 8mm tube) – this would have been a straight 8mm rod together with 1mm copper wire running from the ends to the centre.
This approach allowed a design with an impedance of around 120ohms to be produced which when stacked would have been a better match to 50ohms. The construction would have been easier too. I gave up on this style though because I couldn’t get the overall gain much over 4dBi and the low elevation gain and polar diagram suffered even more. One possible reason for the change in gain is that I modelled the folded dipole as I would have needed to construct it with the plain 8mm rod 6cm lower than the 1mm wire which would have been where the coax connected at the end of the tube. Normally both sides of a folded dipole will be the same distance from other elements as it mounts at 90 degrees to the axis rather than being in line.
A conclusion is simply the place where you got tired of thinking.