Hello everyone, today's article will focus on some basic knowledge of optical communication.
As we all know, our entire communication network now has a great dependence on optical communication technology. Our backbone network, optical fiber broadband and 5G are inseparable from the support of optical communication technology.
The so-called optical communication is a technology that uses optical signals to carry information and transmit data in optical fibers.
Light waves are a kind of electromagnetic waves, so the optical signal also conforms to the physical characteristics of electromagnetic waves.
If you want to improve the information transmission volume of optical communication, it is basically divided into the following three ideas:
The first idea: increase the baud rate of the signal.
Baud rate( Baud), to be precise, called Baud, called baud rate is just a colloquial habit. It is defined as the number of symbols transmitted per unit of time.
Baud rate is easy to understand, the more symbols I transmit per second, of course the amount of information.
At present, as the chip processing technology increases from 16nm to 7nm and 5nm, the baud rate of optics and optoelectronic converter devices has also increased from 30+Gbaud to 64+Gbaud, 90+Gbaud, and even 120+Gbaud.
However, the baud rate is not infinite. The higher you go, the more difficult it is to implement the technology. High baud rate devices pose a range of system performance impairment issues that require more advanced algorithms and hardware to compensate.
It should be noted that the baud rate is not a bit rate (transmission rate).
For binary signals, 0 and 1, 1 symbol is 1 bit. Then, the number of symbols per second (baud rate) is equal to the number of bits per second (bit rate, bit/s). For quaternary signals, 1 symbol can express 2 bits, the number of symbols per second ×2 = the number of bits per second.
Quaternary, same baud rate, bitrate double (double amount of information)
So, in order to increase the number of bits per second (information transfer rate), we need a symbol that can express as many bits as possible. How? We'll talk about that later.
The second idea: adopt more fiber or channel count.
With more fiber, this idea is easy to be rude. The more optical fibers there are, the more one lanes are equivalent to two lanes, four lanes, eight lanes, and of course, the amount of information transmitted will double.
However, this approach involves investment costs. Moreover, the number of fibers is too large, and the installation will be troublesome.
In a single fiber, it is better to build multiple channels.
The number of channels can be spatial channels or frequency channels.
Spatial channels include patterns (single-mode/multimode), fiber cores (multi-core fibers), and polarization (to be discussed later).
In the case of frequency channels, this brings us to WDM (Wavelength Division Multiplexing). It puts different business data in different wavelengths of optical carrier signals and transmits them in a single optical fiber.
WDM wavelength division multiplexing
Wavelength × frequency = speed of light (constant value), so WDM is actually frequency division multiplexing
WDM is also not infinite wave number. Each wavelength must be within the specified wavelength range, and there must be a protective interval between each other, otherwise it is easy to "crash".
At present, the industry is working hard to expand the frequency band of optical communication to the "C +L" band, which can achieve 192 wavelengths and a spectral bandwidth of nearly 9.6THz. If the single wave is 400G, that is the transmission rate of 192×400G = 76.8Tbps.
The third idea, which is also the one we want to focus on today, is high-order modulation.
That is, using more advanced modulation techniques, the bits that can be represented by a single symbol are increased (corresponding to the first idea), and the bit rate is increased.
For modulation, everyone must not be unfamiliar. The PAM4, BPSK, QPSK, 16QAM, 64QAM, we often hear about are all modulation techniques.
When I talked about electrical communication and mobile communication before, I mentioned that if you want electromagnetic wave symbols to express different information, it is nothing more than adjusting several physical dimensions of electromagnetic waves.
The physical dimensions that everyone is more familiar with are amplitude, frequency, and phase.
Light waves are also electromagnetic waves, so the idea of modulating light waves is basically the same.
Optical fiber communication system, there are mainly 6 physical dimensions for multiplexing, namely: frequency (wavelength), amplitude, phase, time (OTDM), space (air separation multiplexing), polarization (PDM).
█ Amplitude modulation
Frequency multiplexing is actually WDM wavelength division multiplexing, which has just been introduced. Next, let's look at amplitude modulation.
In the early optical communication systems, we used direct modulation (DML, Direct Modulation Laser). It is an intensity (amplitude) modulation.
In direct modulation, the electrical signal is directly modulated by the intensity (amplitude) of the laser using the on-off keying (OOK) method.
This one is a bit like our nautical traffic lights. When it is bright, it is 1, when it is dark, it is 0, and one symbol is one bit, which is simple and clear.
The advantage of direct modulation is the use of a single device, low cost, and small accessory loss. However, its drawbacks are also many. Its modulation frequency is limited (related to the oscillation of the laser), which produces a strong frequency chirping, limiting the transmission distance. Direct modulation of the laser may occur in linear frequency modulation, so that the output line width increases, the dispersion introduces pulse broadening, the channel energy loss, and the generation of crosstalk to the adjacent channel (do not understand it to skip it).
So, then came the External Modulation Laser (EML).
In external modulation, the modulator acts on the modulator outside the laser, and with the help of physical effects such as electro-optical, thermal light or acousto-optic, the optical parameter of the laser beam emitted by the laser is changed, so as to achieve modulation.
As shown in the following figure:
There are two commonly used methods of external modulation.
One is EA electrical absorption modulation. The modulator is integrated with the laser, and the laser has a constant light intensity of light, which is sent to the EA modulator, which is equivalent to a gate, and the size of the door opening is controlled by the voltage. By changing the size of the electric field, the absorption rate of the optical signal can be adjusted, and modulation can be realized.
There is also the MZ modulator, which is the Mach-Zehnder Mach-Zehnder modulator.
In the MZ modulator, the input laser is split into two channels. By changing the bias voltage applied to the MZ modulator, the phase difference between the two lights changes and is superimposed on the modulator output.
How does voltage produce phase differences?
Based on the electro-optical effect – the refractive index n of some crystals , such as lithium niobate , varies with the intensity of the local electric field.
As shown in the figure below, the arms are dual paths, one is the Modulated path and the other is the Unmodulated path.
When the voltage acting on the modulation path changes, the refractive index n on this arm changes. The propagation rate of light in the medium v = c / n (the rate of light in the vacuum divided by the refractive index), so the rate of light propagation v changes.
The length of the two paths is the same, some people arrive first, some people arrive later, so there is a difference in phase.
If the phase difference between the two lights is 0 degrees, then after adding up, the amplitude is 1 + 1 = 2.
If the phase difference between the two lights is 90 degrees, then after addition, the amplitude is the square root of 2.
If the phase difference between the two lights is 180 degrees, then after adding up, the amplitude is 1-1=0.
Everyone should also think of it, in fact, the MZ modulator is based on double-slit interference experiments, and the principle of water wave interference is the same.
Peaks and peaks are superimposed, peaks and valleys are offset
█ Light phase modulation
Next, let's talk about light phase modulation. (Knock on the chalkboard, this part is the point!) ）
In fact, we have just talked about phase, but that is the amplitude difference generated by the phase difference, which is still amplitude modulation.
First of all, let's recall high school (junior high school?). ) mathematical knowledge - imaginary numbers and trigonometric functions.
In mathematics , an imaginary number is a number of shapes such as a + b * i. The real a can correspond to the horizontal axis on the plane, and the imaginary b corresponds to the vertical axis on the plane, so that the imaginary number a + b * i can correspond to the point (a, b) in the plane.
You should also remember that the coordinate axis can actually correspond to the waveform, as follows:
Waveforms, in fact, can be represented by trigonometric functions, for example:
How beautiful, how enchanting
X = A * sin(ωt+φ)= A * sinθ
Y = A * cos(ωt+φ)= A * cosθ
ω is the angular velocity, ω = 2πf, and f is the frequency.
Φ is the initial phase, 0° in the figure above.
Remember that? Look at A as amplitude and θ as phase, which is the waveform of electromagnetic waves.
θ=0°,sinθ=0
θ=90°,sinθ=1
θ=180°,sinθ=0
θ=270°,sinθ=-1
Okay, the basics review is complete, now into the main text.
First, let's introduce the constellation chart.
In fact, when I introduced the phase change of the MZ modulator just now, I have already seen the shadow of the constellation chart. The following charts belong to the constellation chart. The small black dot in the figure is the constellation point.
You will find that the constellation chart is very similar to the vertical and horizontal coordinate system that we are very familiar with. Yes, the constellation point in the constellation chart is actually a pair of amplitude E and phase Ф.
I/Q modulation (not IQ modulation) should be proposed.
I, is an in-phase, homogeneous or real part. Q, is a quadrature phase, orthogonal phase or imaginary part. The so-called quadrature is a carrier wave with a −90 degree difference in phase relative to the reference signal.
Let's move on.
On the constellation chart, if the amplitude is unchanged, with two different phases 0 and 180 °, representing 1 and 0, 2 symbols can be passed, that is, BPSK (Binary Phase Shift Keying).
BPSK
BPSK is the simplest and most basic PSK, very stable, not easy to error, strong anti-interference ability. However, it can only transmit 1 bit per symbol, which is too inefficient.
So, let's upgrade it and do a QPSK (Quadrature PSK, quadrature phase shift keying).
QPSK is a quad-level phase shift keying (PSK) modulation with 4 level values. Its frequency band utilization is twice that of BPSK.
Image courtesy of Keysight
With the increase of the base system, although the frequency band utilization increases, it also brings disadvantages - the distance between the elements is reduced, which is not conducive to the recovery of the signal. Especially when subjected to noise and interference, the bit error rate will increase.
To solve this problem, we have to increase the signal power (that is, increase the signal-to-noise ratio of the signal to avoid an increase in the bit error rate), which reduces the power utilization.
Is there a way to balance the frequency band utilization and the distance between the various symbols?
Yes, this introduces QAM (Quadrature Amplitude Modulation).
QAM is characterized by not only different phases but also different amplitudes between each symbol. It is a modulation method that combines phase and amplitude.
Look at the following GIF, you will understand:
Amp, amplitude. Phase, phase.
In fact, QPSK is the QAM with a level of 4. The image above is 16QAM, 16 symbols, each symbol 4 bits (0000, 0001, 0010, etc.).
64QAM words, 64 symbols (2 to the nth power, n = 6), each symbol 6 bit (000000, 000001, 000010, etc.).
How did QPSK make this kind of modulation?
We can look at a picture of mashing QPSK through the MZ modulator:
Image courtesy of Keysight
In a transmitter, the electrical bitstream is divided into I and Q parts of the signal by a multiplexer. Each of these two parts directly modulates the phase of the laser signal on one arm of the MZ modulator. Another MZ modulator shifts the lower branch phase π 2. After the two branches are recombined, the result is a QPSK signal.
High-level QAM is more difficult to modulate. Limited by space, I will explain it to you next time.
When introducing wireless communication modulation earlier, it was said that 5G and Wi-Fi 6 are rushing 1024QAM. So, can optical communication be so high-level QAM?
Don't hide from you, someone really did this.
A few years ago, a company demonstrated 1024QAM modulation based on advanced galaxy shaping algorithms and Nyquist subcarrier technology, based on 66Gbaud baud rate, to achieve 400 km transmission at 1.32Tbps, with a spectral efficiency of 9.35bit/s/Hz.
However, this high-order modulation is still in the laboratory stage and is not commercially available (and it is not known whether it is possible to commercialize). At present, the actual application does not seem to exceed 256QAM.
Although high-order QAM brings a significant increase in transmission rate, it has high requirements for component performance and high requirements for chip computing power. Moreover, if the channel noise or interference is too large, there will still be the problem of the high bit error rate just mentioned.
1024QAM, the rhythm of dense phobia
At the same 30G+ baud rate, the optical signal-to-noise ratio (OSNR) of 16QAM is about 5dB higher than that of QPSK. As the number of constellation points in the constellation increases, the OSNR of 16QAM will grow exponentially.
As a result, the transmission distance of 16QAM or higher QAM will be further limited.
In order to further squeeze out the bandwidth potential of optical fiber communication, manufacturers have sacrificed a new big killer, that is, coherent optical communication. Interested readers can learn more.
█ PAM4 and polarization multiplexing
At the end of the article, let's talk about two "doubling" technologies - PAM4 and PDM polarization multiplexing.
Let's start with PAM4.
Before PAM4, we traditionally used NRZ.
NRZ, which is the abbreviation of Non-Return-to-Zero, literally means "non-zero", that is, non-zero encoding.
Signals encoded in NRZ are digital logic signals that use high and low signal levels to represent transmitted information.
NRZ has unipolar non-zero codes and bipolar non-zero codes.
Unipolar non-zero code, "1" and "0" correspond to positive and zero levels, or negative and zero levels, respectively.
Unipolar non-zero code
Bipolar non-zero code, "1" and "0" correspond to positive and equivalent negative levels, respectively.
Bipolar non-zero code
The so-called "non-zero" does not mean that there is no "0", but that every time a bit of data is transmitted, the signal does not need to return to the zero level. (Obviously, NRZ saves bandwidth compared to RZ.) ）
In optical module modulation, we use the power of the laser to control 0 and 1.
Simply put, it is luminous, and the actual emitted optical power is greater than a certain threshold, which is 1. Less than a threshold value, which is 0.
The transmission 011011 is like this:
NRZ modulation
Later, as mentioned earlier, in order to increase the logical information transmitted per unit of time, PAM4 was developed.
PAM4, or 4-Level Pulse Amplitude Modulation, Chinese called four-level pulse amplitude modulation. It is an advanced modulation technique that uses 4 different signal levels for signal transmission.
Or the transmission 011011, it becomes like this:
PAM4 modulation
In this way, the logical information represented by a single symbolic period has doubled from 1 bit of NRZ to 2 bit.
NRZ vs PAM4 (Eye diagram on the right)
So the question is, if 4 levels can be doubled, why don't we make 8 levels, 16 levels, 32 levels? Isn't it cool to double the speed at will?
The answer is no.
The main reason lies in the technical process of the laser. Achieving PAM4 requires a laser capable of precise control of power.
If the process is not OK, engaging in higher digit levels will cause a high bit error rate and will not work properly. Even PAM4, if the channel noise is too large, it will not work properly.
What is PDM polarization multiplexing?
PDM polarization multiplexing is polarization Division Multiplexing
I don't know if you've read an article I wrote about antennas before. Inside the antenna, there is a concept of dual polarization, in space, the electromagnetic wave "rotates" 90 degrees, you can achieve two independent electromagnetic wave transmission.
Dual polarization of the antenna
The reason for polarization multiplexing is actually similar. It uses the polarization dimension of light to transmit two independent data information through two mutual quadrature polarization states of light in the same wavelength channel, so as to achieve the purpose of improving the total capacity of the system.
It is equivalent to achieving dual-channel transmission, which, like PAM4, doubles.
PDM polarization multiplexing, X polarization and Y polarization, independent of each other
Image courtesy of Keysight
Well, that's all for today's article. Thank you for your patience to watch, we will introduce coherent optical communication in the next issue, not to see or scatter yo!
—— End of text ——
bibliography:
1, know it or not, know it or not, what is coherent optical communication, Keysight Technology
2) David takes you to light communication, Finissa David
3, the words of large-capacity optical fiber communication, Fiber, zhihu
4. Understanding Optical Communication, Yuan Rong, China Machine Press
The reproduced content represents the views of the author only
Does not represent the position of the Institute of Physics, Chinese Academy of Sciences
If you need to reprint, please contact the original public account
Source: Fresh Dates Classroom
Edit: Cloud open leaves fall