Why doesn't wireless transmission and reception require a common reference voltage

Question

so I have been struggling with this for a while now. I have tried to search for the answer but haven't succeeded yet. Here is how it goes:

If we take USART as an example for wired communication between two devices, the TX and RX lines are referred to a common signal(GND) so that a "1" transmitted is also received as a "1", which makes complete sense. But when we go to wireless communication, how does a receiver interpret the incoming data correctly? I know that electrical signals differ from EM communication. But why and how is a "1" sent is also received as "1" since there is no reference signal. Any reference to reading resources would be highly appreciated!

Answer 1

Wired systems rely on current or voltage to carry signals. These can use a common reference, like early telegraph systems that literally used the earth as a signal return. That’s not strictly necessary though: wired systems can use different means to detect the presence of signal, like detecting edges or sensing a carrier signal.

Radios use electromagnetic waves to do that. Waves are dynamic changes in the electromagnetic field, and the receiver only need to be able to discriminate between these changes and background noise to recover the information. RF waves require no physical connection between the transmitter and receiver: they propagate through free space with no physical medium required.

It’s much the same as using light to communicate. You can sense the light being ‘on’, ‘off’, or at some intensity if the difference between it and the background light is large enough. It's why you can see stars at night but not during daytime: sunlight reflecting / mixing with air drowns out starlight. (You can always see stars in space however, even if we can't hear you scream.)

Nevertheless, light waves, being electromagnetic energy like RF, travel through free space too.

Light: Particle or Wave? Yes.

Here's a timeline of the evolution of light wave / particle theory, with a side of Maxwell. http://global.canon/en/technology/s_labo/light/001/11.html

Answer 2

Maxwell's equations are chock full of derivatives, which is to say that we only care about change not absolutes.

A changing electric field becomes a magnetic one; a changing magnetic field becomes an electric one. An unchanging field does neither of these things.

If one of the changing fields hits a wire or coil, we get a signal we can process.

Beyond the physics, in practical terms we also tend to avoid absolutes in how we encode data for transmission. For example, we don't just turn an RF carrier on and leave it on and have that mean something, we turn it on and off, and we use an Automatic Gain Control to figure out that we're seeing a variation around an average channel energy which we can use to distinguish on from off. Or better yet, we don't transmit at one frequency, but rather toggle between two, and the receiver decides which is stronger. Or still fancier encoding techniques. And then we go and wrap all of this with redundancy, for example verifying the message with a checksum.

Answer 3

But when we go to wireless communication, how does a receiver interpret the incoming data correctly? I know that electrical signals differ from EM communication. But why and how is a "1" sent is also received as "1" since there is no reference signal.

In many RF systems that transmit data there is a reference. For instance, the modulation can be FM (frequency modulation) with digits 1 and 0 being transmitted at two different frequencies. This is also mirrored in non RF applications by a system called FSK (frequency shift keying). The two different frequencies are fairly easily distinguished and can be demodulated back into 1s and 0s.

There are no propagation phenomena that can significantly "abuse" the steady frequency of an FM transmission other than doppler shift (not applicable to static systems). So, there is a known reference and the transmission itself isn't relied upon for that reference.

However, a modulation system that uses AM (amplitude modulation) is more prone to ambiguity because it's the amplitude of the RF carrier wave that dictates whether a digit 1 or 0 is transmitted and, given the vagaries of transmission distance and "other factors", you would benefit from an "in-built" reference.

But, if the data you transmit is (say) Manchester encoded or (say) scrambled, the now plentiful data-driven carrier-amplitude changes can be referenced by the receiver and the modulated carrier converted into the encoded data whereupon, it can be decoded by logic circuits to produce the original data.

Hopefully there are enough keywords in this answer that can allow you find the reading resources you need.

Answer 4

Just to give one common example. Many digital radio systems use a system known as frequency modulation, in which a 0 and a 1 are encoded as slightly different frequencies (the technique is sometimes called shift keying). The broadcast signal is thus a continuous stream modulating regularly (at roughly the data bit rate) between these two frequencies. The antennas at each end are broadband enough to handle both frequencies, with the receiver having more frequency-sensitive downstream processing to distinguish them and turn each back into its 0s or 1s as appropriate.

Answer 5

I'll offer a bit of frame challenge here...

The answer to "what is the difference between wired and wireless transmission?" is: there is no significant difference.

But that's not the question you're asking 😁 Your question actually is: what's the difference between wired/wireless transmission, and optical transmission.

1. There is no big difference between wired and wireless transmission

In both cases, information is carried by electromagnetic fields that propagate as electromagnetic waves. Usually what matters to the receiver is the local electric field, for example a voltage on the gate of a FET, that will turn it ON or OFF.

If you use waveguides, for example wires or PCB traces, then you can make those EM waves go where you want. Then you can make a simplifying approximation that your cable is a lumped element which has stuff like "current" and "voltage" and "ground is zero volts". But if your cable gets long enough relative to the wavelength, then the approximation breaks down, you have to remember the signal actually propagates along your cable, which in this case is a transmission line, in other words a waveguide. Then, there is no concept of "current" or "voltage" in the whole cable, rather each point along the cable has local EM fields which wiggle around electrons, creating time-varying local currents in the conductors, which will all be different along the length of the cable.

Likewise, a transmitting antenna does not "create" electromagnetic waves from an "electric signal". The incoming electric signal is already an EM wave traveling along a waveguide (like a coax cable, but not necessarily). The antenna is also a waveguide that takes in the EM waves coming from the feed cable, and it is shaped just right to throw these EM waves into the air. It's basically an impedance transformer, which is the same as a lever or a funnel, but I digress. And a receiving antenna is also a waveguide, which has just the right shape to capture EM waves from the air and funnel them into another waveguide, like a cable or a PCB stripline.

Say you have a radio transmitter and a radio receiver.

If you move them close together until the antennas touch and make electrical contact, is it wired or wireless? The same physical phenomena are involved. Now if you disconnect the antennas (say they're mounted on SMA ports) and connect the ports together with a coax cable, pretty much nothing has changed except free wave propagation in the air was replaced with propagation inside a waveguide. The receiver will get a lot more power, but if it handles it, it'll work fine.

The difference is how the signals are encoded, and what frequencies are used. Different encodings and frequencies work best is you want your EM signal to propagate in different types of cable or in the air. For example simple logic levels aren't modulated on a carrier frequency, so they're really not suited to radio transmission.

If we take USART as an example for wired communication between two devices, the TX and RX lines are referred to a common signal(GND)

Well no. First, "GND" is not the same potential at the receiver and the transmitter. If your cable is long enough and/or some current flows into the "GND" wire then both ends of this wire will be at different potentials. There is no "GND".

This works because all receivers are differential. They care about voltage between two pins. Sometimes they are explicitly differential, for example a RS-485 receiver has "INPUT+" and "INPUT-" pins. Sometimes that's hidden, for example you have pins labeled "INPUT" and "GND", but what counts is the voltage between them. So if "GND" is at different potentials on both ends of your cable, the receiver won't care because it only knows about its local "GND" potential. It has no idea what "GND" means at the other end of the cable. As long as its "INPUT" pin is enough volts above its "GND" pin, that's a logic 1.

The notion of "voltage reference" is a convenient simplification to make human's job easier. But chips don't care. For example a FET has a gate. But what turns it on is the voltage between gate and source, so it has two input pins, one of which is low impedance (the source). And, since a 74HC logic inverter has two FETs in its input stage, it has three input pins: the one labeled "INPUT" connects to the gate of both input FETs, and the other two input pins are "VCC" and "GND" which connect to the sources of these FETs.

So, what your receiver cares about is voltage between two pins, ie electric field applied to the gate of the input FET, which determines if it turns ON or OFF.

When a "1" is transmitted, the transmitter launches an EM wave onto the wires, then this wave propagates, and its local voltage rides on top of what is locally labeled "GND". At the other end, the EM wave transfers its electric field into the receiving transistor, which turns on or off.

This happens at all frequencies. At low frequencies it is not noticeable, but it happens nonetheless.

2. About optical...

DC, radio and light are all EM waves. The difference between wired/wireless and optical is the wave-particle duality.

Wired/wireless use the same physical mechanism, ie they receive EM waves as waves, using the EM fields, usually voltage.

Optical receives EM waves as particles, ie photons that kick an electron or molecule into an excited state. That is, photons trigger a mechanism that causes electrons to move, which then is current.

Answer 6

Requiring a common voltage reference is not a property of all wired communication standards, just those which use voltage levels to encode the information being transmitted.

As a counter-example, Ethernet BASE-T is a wired communication standard in which the data is transmitted by currents flowing though the cable, which in turn produce voltage differences on each side of the cable. As such, there is no need for common voltage reference, and Ethernet indeed works fine with devices which don't share common ground.

Answer 7

It's a matter of the frequency components used in the transmission.

A "baseband" transmission that has a DC component will need a common reference voltage, because the voltage above the reference carries information.

Wireless systems encode data so the DC part of the signal does not carry information, which allows for data to be transmitted across AC coupled links. The receiver just substitutes whatever DC level is easiest for them to handle by biasing the receiver, for an analog system this is typically ground, for a digital system this is the middle of the ADC range.

This is also done in wired systems. Ethernet for example is magnetically coupled on both sides, so the cable is floating. Power over Ethernet uses this and uses low frequencies that do not pass the transformer to transport power.

UART requires the common reference because filtering out the DC part would cause the "idle" signal to reduce to zero over time, so the information "link up but idle" would fade.

It would likely be possible to recover that signal, but the receiver would have to be built differently and that would be a different standard (IRDA comes closest to that).

Answer 8

There is no absolute voltage. Like altitude, you can only measure it relative to some reference level.

But there is such a thing as absolute amplitude of a wave. You can tell the difference between noisy and quiet, and no reference level is required.

Answer 9

The answer lies in how we convert a wireless signal. The wired signal uses some components to measure if the voltage is above some value to label the signal as HIGH or LOW, and voltage is defined from a common ground. The wireless signal comes out as a wave with a variety of frequencies, but if you've ever seen Galloping Gertie, the Tacoma Narrows bridge, then you've seen resonance.

Resonance is a system's large response to smaller (lower energy) oscillations and it occurs whenever the actuations are at a frequency near one of the natural frequencies of the system, which depend on some choices of the physical system and it is the reason tuning your radio connects to a different station. By doing this and reflecting them, antennae "catch" the right waves of a frequency in some short range of frequencies inside them.

I wanted to do a precursor before answering the question. How do we convert a wave into 1 or 0? Well we don't have an amplitude reference because varying distance would reduce the signal power. We also don't have a timing reference, so we can't use the phase of the signal. The only other characteristics are relative change of amplitude over time (we can pretend receivers are stationary since these signals are light and much faster than the motion of vehicles, see Doppler Effect), relative phase changing, or the frequency. We'll discuss the last one, which is how FM radio works. But the names you can look up for future reading are these three options, amplitude shift keying, frequency shift keying, phase shift keying, there are others as well, but a text or lecture notes on wireless encoding should kelp you out a lot.

By sending a frequency to the higher part of the accepted window vs the lower part we have a way to distinguish HIGH and LOW just as we did before. The last part is a bit on timing, mostly because it is just a cool thing that we get for free is that when do we check if the signal is HIGH or LOW?

Well, the resonant frequency is the one that sets how often, so we can check once per period, but it turns out it is not super important because we can actually check twice per period, but if you are infinitely unlucky, you could get exactly halfway where the frequencies are changing and those won't look HIGH or LOW. We could do three times, but turns out four is better. If you draw a wave with 3 evenly spaced points and move the phase back and forth, you might be able to see why 4 is better. But if you want background on that, look at the Nyquist sampling theorem.