Dr. Pepper

You are a real master

then its necessary to analyze the structure and look which parameter is related to lamda and which is perhaps related to other factors. And yes you have to find a center frequency. For a single band antenna this is easy for such a multiband its much trial and error i think. For me its easier to re-optimize the design for another frequency range. For lower frequency's there should be better designs that are not so big, the only other range i can imagine is in the 5GHz region but there you not need such a wide bandwith. Attached you will find the cst file, feel free to recalculate it for another freq. range or to find the lambda depending parameters ;-) slot antenna 17-27-reflectors.cst

And additional a variant with directors. They harmonize good with the design but they give only a slight improvement.

An improved version of the slot dipole with wings. Good broadband behavior and good gain. 1.7 - 2.7 GHz with SWR <1.3, Gain from 2.1 - 2.5 GHz > 13 dBi

Another Option?:

Yes this was the picture i have in mind. Thanks for the hint

At the beginning CST looks very complicated, i first also needed some time to do something. At the beginning its useful to download a few projects and look into the history list how the antennas are modeled and what is all done. Later you will be able to make your own projects and play around. When you have an Nvidia card with cuda make sure you enable the use of this in the environment variables under windows, it will speed up the simulation.

I have a SAA2 NanoVNA which costs about 80 Euro but can only measure up to 3 GHz, if you want to measure up to 6GHz you have to spend about the double for it. I think such a VNA is the most basic tool to measure any antenna. It gives you the ability to correct the antenna that it works like it should.

If you realy want to try different designs than its worth to buy a nanovna, they are not so expensive anymore. 15 years ago when i try to make antennas i simply compared the signal i get from them from a external router. Better signal, better antenna. But with a vna you can measure if the antenna is good and you can compare different builds or even change some parts for adjustment. If you only want a cheap antenna that even works when the dimensions are not 100% perfect try to make a batwing antenna. They are so broadband that even a small mistake will not change that much that the antenna is useless. For the Haringtenna i think you have to find a can with exactly these dimensions to bring it to work.

Typically the antennas are designed to be like in the pictures. When you change something, even a small detail, it can be that the antenna is not working properly anymore. The resonant frequency will shift or the antenna will be not resonant at all.

Updated version of the Biquad in a pot for 2.4 GHz, this time with a better impedance match over a wider frequency range and vswr starting at 1 ;-). But as mentioned before there are better designs out there. Parameters

This is indeed much simpler and more stable over a wide frequency range. The idea with the directors is also not bad, perhaps with this the reflector can be a little bit smaller.

You are right, its not perfect. But the first goal was reach 16 dBi. I will try to tune the design for a more stable range.

Biquat + Horn for 2.4 - 2.5 GHz & optimized for gain Size Impedance + Gain Parameters: File: Biquad 2-4ghz-double-reflector.cst

If somebody is interested in this you can find a model on thingiverse but its important how you attach this yagi on the router antenna. It influences the swr strongly. I found it works best at the start or end of the antenna like in the pictures and worst in the middle. Best is to try some positions from top to down and look on the signal strength on the receiving end.

You not only have to consider conductivity also price, stability and weather resistivity. For performance copper or even silver would be the best for all parts of the antenna, for price and mechanical stability zinc coated steel would be the best. This should also withstand weather for a relative long period. Would be interesting to cover the steel sheets with copper in a galvanic process. Aluminium can also be used but is hard to solder on.