Hector Martin

Circuit design for stuff pushing into super high frequencies is god damn witchcraft. Look at that. Trying to run the math in my head for the corner frequencies.https://twitter.com/wavedrom/status/997272423798980608 …

> Trying to run the math in my head for the corner frequencies. Wait you can *do* that?

Sort of. All of RF comes down to sinusoids, and sinusoids are the results of following circles at constant speeds. End result is that in geometric systems, frequencies are always a function of the radius. From there it comes down to figuring out where and how the resonances act.

You got field E and field B and they're sinusoidal periodic and tangential to each other, travelling at c/(relative permativity * relative permeability) if im remembering correctly, so frequency gives wavelength, and if wavelength lines up w/ geometry correctly you get a pole.

Basically, if you can model the geometry in your head, and superimpose the fields' path over it, you can suss out where and how geometry will create an effect in the wave as it travels through, and from there cutoff frequencies can be approximated with simple geometry.

Yeah so last time I built a cantenna I punched in the numbers into some online calculator. Seriously though, I still have no idea how you'd begin to geometrically analyze the frequency characteristics of a filter like that. My mind is utterly boggled by anything RF.

I haven't ever done any antenna design, that shit is black magic. All my RF understanding is in devices and pcb's. If you think this shit is nuts though you'd have a lot of fun with metamaterials. An RF columnator in metamaterials looks nothing like you'd expect.