Wireless Made Simple (Part1)

The whole idea of 6G is to overcome the current barriers by leveraging new inventions and innovations.

The whole idea of writing this note is to explain to a layman how wireless technologies have evolved over the years and what bottlenecks in this evolutionary process that 6G will try to solve. I plan to do this in two parts. The current note will explain the legacy technologies and the challenges associated with them, and the 2nd part will talk about how 5G is solving some of these challenges and how 6G plans to solve them using new upcoming capabilities in software and silicon technologies. So let’s begin with the 1st part.

1st Generation (1G)

The following picture explains.

Here we had a tall antenna of height like those used currently for TV broadcasting, transmitting with a range covering the whole city. The receptors were located in cars, so the signal to transmit back to the tower (typically 10s of km) needed a high-power transmission capability. This was only possible using the car battery as the source of power coupled with an external antenna.

The modulation scheme used was simple - either AM or FM using a single frequency (f1) capable of driving a couple of channels - serving the full city, thereby creating a huge bottleneck for the consumers in making a connection.

This was one of the worst kinds of design for a mobile network as it wasted a huge quantum of spectrum and prevented comsumer’s from carrying portable devices.

2nd Generation (2G)

The typical 2G (GSM) network looks something like this.

Note that there are two important changes here (compared to 1G).

Firstly, the city has now been divided into cells of much smaller sizes. This is done to improve the capacity of the overall network (to serve more customers) and to empower these customers to carry handheld devices as they need less power to reach their signals to reach the towers.

Secondly, the frequency of these individual cells is different from those placed adjacent. This managed interference by preventing the signals with the same frequency from colliding with each other.

The modulation scheme used in a 2G system was TDMA (Time Division Multiple Access). The way it worked is explained in the figure below.

This is how it works. Let’s say we have a spectrum of BW 1 MHz. This will be divided into 5 channels each of 200 KHz. Now these channels will be further divided into 8 timeslots (a certain fixed duration of time). These repeating timeslots of the Frame made up of many such channels will end up creating logical channels for each individual user.

The basic problem with this approach is that it is highly inefficient, as these individual nailed-up connections are dedicated to a customer which he/she continues to hold even when the timeslot is not being used (pauses between words).

3rd Generation (3G)

A typical 3G (CDMA/WCDMA) network looks something like this.

In the case of CDMA (Code Division Multiple Access), the whole network is converted into a single-frequency network. The transmitted signals are modulated using codes that see other users as background noise. The network’s transmission capability is limited by the power of this background noise. The math works as shown below.

The key difference between TDMA and CDMA is that in the case of the former the whole BW is sliced into chunks of 200 KHz, whereas for the latter all the channels (defined by separate codes) leverage the full BW of 1.25 MHz for carriage.

The other significant difference is that CDMA/WCDMA was the first cognitive system with power control capabilities, i.e. the UE has a mechanism to inform the BTS to transmit at a reduced power in case it is close to the BTS. This is done to control the overall noise levels of the combined system to a manageable level to drive spectrum efficiency.

Hence, the channel’s capacity is not wasted as it is limited by ONLY the overall background noise and nothing else - making CDMA significantly more efficient than GSM.

However, there are some problems too.

The channel size of CDMA is limited to 1.23 MHz (guard band of 200 KHz between channels) and 4.2 MHz for WCDMA (with a guard band of 400 KHz on both sides) - making the data aggregation capabilities limited.

The reason for this small channel size is to prevent the filters from getting too complex with the purpose of limiting out-of-band transmissions. The higher the width of the carrier more difficult it becomes for the filter to prevent transmission out of the band. The following figure explains.

As you can see it is extremely difficult to design a wideband filter with sharp cut-off to limit out-of-band transmission. Hence, 3G CDMA/WCDMA-based systems never scaled to large bandwidths, thereby limiting their overall data throughput.

4th Generation (4G)

In the case of 4G, the layout of the network follows the same single-frequency configuration as we described above. However, the physical layer of 4G networks uses frequency modulation with a twist. This is called OFDMA (Orthogonal Frequency Division Multiple Access).

Here instead of transmitting the user’s signals over the whole bandwidth (as was done in the case of 3G), the BW is broken down into small subcarriers, each of 15 KHz. The user’s signals are packed into these carriers (both in the time and frequency domain) for the purpose of transmission.

OFDMA - BTS to Handset

The real magic comes from the fact that these individual carriers of 15 KHz are overlayed on top of each other orthogonally without causing any interference with each other. This is explained in the following figure.

Laying out the carriers orthogonally serves two important purposes. Firstly, it enhances the ability of the system to aggregate a large quantum of spectrum (20 MHz vs. just 5 MHz in the case of 3G), thereby significantly increasing data speeds. Secondly, it makes the filter design simpler - preventing the need for sharp cut-off at the edge to filter out adjacent channel interference.

SC-FDMA - Handset to BTS

But the handsets while transmitting back to the BTS follow a modulation scheme called SC-FDMA (Single Carrier Frequency Division Multiple Access). But why? But before we answer this question let’s understand what is SC-FDMA. Now SC-FDMA distributes the user signals between the OFDMA carriers using DFT (Discrete Fourier Transform) such that the Peak Average Power Ratio (PAPR) is minimized for the uplink signal.

The reason for doing that is to make the amplifier design of the handset simple and to save cost. Since the BTS is not that price-sensitive, and power control regulations are relaxed, hence the focus was on the handset side so as NOT to make the design complex - resulting in increased cost.

5th Generation (5G)

Now 5G in the physical layer uses the same OFDMA strategy for multiplexing but on the user side (uplink) it has been able to do away with SC-FDMA and replaced it with the same OFDMA as it used in the downlink. Apart from this, it has been able to enhance spectral efficiency by using advanced modulation schemes like 256 -QAM (Quadrature Amplitude Modulation) and advanced Error Correction Coding techniques. In order to appreciate this let’s understand how a wireless channel works.

But before we dive down into the details of 5G and follow it up with the possible advancement in 6G let’s try to understand the challenges in any communication system.

Key Challenges

The key challenges faced during wireless communication are described in the following figure.

Noise

5G has deployed various mechanisms to overcome these challenges. In order to contain noise better and improved modulation schemes like 256 - QAM has been deployed. These modulation schemes are dynamically altered to take care of various specific use cases in order to improve the signal quality.

Multipath fading

Multipath fading is explained in the following diagram below.

The signal transmits via two paths to reach the receiver totally out of phase, thereby canceling each other out - resulting in total fading. The impact of this fading increases significantly at higher frequencies when signals follow straight lines (at higher frequencies) and get reflected from multiple directions.

Shannon’s Capacity

It is the theoretical maximum throughput a physical channel (wireless/wireline) is capable of supporting given its bandwidth and signal-to-noise ratio. The channel capacity is given by C = Wxlog2(1+S/N). W is the bandwidth, and S/N is the signal-to-noise ratio.

To be Continued in Part 2…..