Analogtechnik

Michael Steer - Microwave and RF Design, Volume 1- 5

Since I have built a Radio & learn currently for High Frequency & EM, I thought I share my thoughts.
After the 19th century, more and more measurements have been done, conducting Radio Waves and Electron Beams.
Thus by enhancing more 'Parallel' circuits on the Load Side of the circuits, people thought about connecting 'Boxes'.
These Boxes describe the transmitting and receiving parts of circuits, which have been developed earlier by mathematicians, applied by electric engineers during service or war, and optimized in the 1960ties. (in matrix)

So, instead of using Maxwell Equation all the time, you use impedances in matrixes in a shorter form to calculate with these systems. (Y-Matrix, Z-Matrix- >S-Matrix)  If the length of the source and the load are not 'balanced', (same), then reflection occurs on the power transmission line.

People use these techniques on the load (for example a sensor), which changes due to vibration their length, and thus the voltages. (However still a different Measuring method like H-Bridge ) This can be applied to small projects, but even the biggest (Magnetotron). Thus, today Vacuum Tubes (former in computers in 1945) are applied on high-frequency transmitters. This is crucial:

a) You can use an Impulse in the Input towards an unknown system. Thus you receive a Sinc-Function as an Output. ('Impulse Response', on the output if the system is f.e. a filter) You can calculate the time difference from Input to Output. Thus, the biggest X000km Transatlantic Cables can be measured. Speed of light * time difference = Length of the Cable.

b) You can use a harmonic impulse to test, whether your device does hold the specification of the EMC (Electro Magnetic Compatibility). For example of medical devices. There could be any metal in your room/lab, that creates via an Open Loop an EM Wave, which disturbs your desired medical device. Pulse your device at that frequency to test, whether your circuit misbehaves at this frequency. (Sinusoid or Pulse) .. Heard from Lab assistant the procedure.

c) You can use a Signal to test your Filter. Your filter can be used to specify the heartbeat of a baby (EEG) or could be used to detect voices. You can configure this filter in the digital domain or you can design it analog to maximize energy efficiency. Guess, how 100mW for each Solar Cell, could actually lead to such high Power for Satellites. (Higher Frequency ->  Smaller distance).
Some receivers are built from many 'subharmonics' (each detects a part).

d) You can use between two parallel lines (in the middle) the voltage difference to tune your circuit from the input signal. (impedance tuned towards Load due to the impedance of the long transmission line (f.e. cable or air) -> capacitance + resistance. ) 

e) You can measure from a Signal generator, the exact substance due to the spectrum (precise the spin of the atom) from this Radio signal. (NMR by Applied Science)
Some Audio/Sound Musicians used spectroscopy in the Los Alamos Labs(1945+-).  (Determine the properties of Crystalline Samples.)

Hint for Digital: Digital Signal Processor (DSP) used to be called a 'Numeric Oscillation Device'. You use the internal clock and a PLL, thus a later frequency mixer, and output the Digital Signal as an Analog Signal. (From Data to Analog. Especially for Ethernet/HDMI.)
Music Synthesizers are obviously on lower frequencies, but the same principles apply there as well. (Be careful, when you design a Mixer. And get to know 'Modulation')

Thus Antenna Design and Filter Design become a major part, that is inscribed by these matrices. Noise becomes the most crucial part. (Signal to Noise Ratio), by Propagating through the atmosphere. For example: by the damping of the noise, you would use an amplifier to balance the losses before the signal gets to the receiver's filter.

Remember: The Noise of linear devices, thermic devices, diodes (create harmonic distortion), etc. is majorly affected by temperature differences. So high resistance creates on an higher scale a lot of Noise. This is only due to the vibration of Atoms, instead of the Electrons. However Electrons can be stimulated at Kilovolts (kV) (at room temperature and atmospheric pressure) for plasma at high voltage without creating much heat in the atom. One thing, that I find fascinating is, that on the higher frequency, a lot of Power is emitted. Thus the resistance migrates normally (removed). You obviously need in CERN or the LHC, High-Frequency Transmitters. (also review 'Corona Effect'). From two colliding Impulses, you basically 'receive' / detect ('detect' comes from comparing two phases), the difference at the Output and create a statistical investigation. (yey. It is a 'Higgs Boson'). Huge High-Frequency Devices from WW2 (1945) are still used as Transmitters for Testing the Device under Test (DUT).

One way to work with these matrices is for example to transform these lengths of impedances, such that it fits your application. So don't remove small metals on your tinkering cellphone. The answer is most of the time: Symmetry. In Impedance from Source to Load, the Imaginary Part (Capacitance + Inductance) or in Matrices (Matrix* Inverse Matrix). Even in the placement of the board. (Btw. If they say 'Solid State- Device, it only means 'with transistor'. So the transition from vacuum tubes to 'Solid State Chips'. Solid State-Memory (transl. with transistors')

A Contrast is maybe also important. Just because High frequencies sound cool, are light, and require careful design, the opposite is the same. The brain works on the lowest frequencies (Theta Waves ~8 Hz vs Alpha Waves ~120Hz). The lowest frequencies create the most distorted Phase fluctuation and this is obviously due to the Chaotic Behaviour of Electrons and Atoms. At the highest frequencies, Shot Noise is majorly the cause, that affects your Signal.
Fast CPUs(5GHz) aren't better than slower frequencies of Brains from the entire invention of human history. 

Something I also like to share: (not relevant)

There are studies about deep seas measurement in the pacific ocean, in the deepest layer of ocean level to discover and measure the noise. (oceanographic-seismic measurement). These random distributions are different from quiet telegraph lines and measure frequencies between 0Hz .. 8 Hz. (some  < 0.2 Hz), which contributes to the energy transmission line with their database to optimize the energy leakage. At some point, even temperature is neglectable until you reach Shot Noise. (measuring Photon Phase Shifts)
Nevertheless, it is like measuring the heartbeat of the Earth.

There is for me something underlying in frequencies near the Direct Current (0 Hz), which links all parts of nature in a series+parallel. Contained unique, but different filters (passing Signal, repressing Noise), Redundancy (neglecting Transmission Rates below it, because the signal can be restored by the redundant bits f(x)), and a Link of Epsilons.  (Smallest Number in math. I reference here an Unknown Area. 'Epsilon Subsampling') After Neurons interconnect with each other after a cascade of Neural Spikes, CFD (Cerebrum Spinal Fluid) is pumped through the brain. They pass the 'Pons', which are interlinked by all senses.  Dreams are associated with this region and during sleep, the lowest frequencies are stabilized.
Even If you have a heart transplantation, then something is 'still missing'. The right synchrony of your brain and the heart. (That is magic for me^^)
P.S. If you follow the Shannon-Nyquist Criterium (Or Whittaker-Kotelnikov- Shannon), then strangely, the spoken language is above >2x the frequency of the Alpha Waves (Active State). Math should be true. Right? (Weirdly Inverse)

Random Matrix Theory(RMT) tries to describe these processes and this is the current edge of research. (Start: Dyson (70ties), Even regularly in the University of Saarbrücken. F.e. Epsilon Subsampling)
RMT tries to describe these processes. (Elements in eigenvalues. because the spectral density is of the most importance) It isn't necessary to know them. 

> Using the save practices and implementing them reliably is better. (Gauß Noise) >Thomas Knight built his own EEG (1965).
Once you know a noise source, you can configure your PLL (feedback loop + ["regulator" or better "filter"/Kernel]) or even your 'predictor' towards it for finer precision.
Firstly the Wiener ChinChin provided a possibility for describing noise in the frequency spectrum (Power Law Distribution), while secondly, the Cramer-Rao-Bound provided a maximal theoretical lower bound for the predictor. It cannot be better, or else your model is not optimal.
(RF Contents are Open Source and explained the best by w2aew. If you knew 100%, then you are lying.)