Narrowband versus wideband communication

Many of the wireless protocols that we will cover are known as wideband. We will see that the converse (narrowband) also has a place, especially for LPWAN. The differences between narrowband and wideband are as follows:

  • Narrowband: A radio channel whose operational bandwidth does not exceed the channel's coherence bandwidth. Generally, when we talk about narrowband, we speak to signals that are 100 kHz or smaller in bandwidth. In narrowband, multipath causes amplitude and phase changes. A narrowband signal will fade uniformly, so adding more frequencies does not benefit the signal. Narrowband channels are also called flat fading channels because they usually will pass all spectral components with equal gain and phase to one another.
  • Wideband: A radio channel whose operation bandwidth may significantly exceed its coherence bandwidth. These are generally greater than 1 MHz in bandwidth. Here, multipath causes the "self-interference" problem. Wideband channels are also called frequency-selective since different parts of the overall signal will be affected by the different frequencies in wideband. This is why wideband signals use a diverse range of frequencies to distribute power across many coherence bands to reduce fading effects.

Coherence time is a measure of the minimum time required for an amplitude or phase change to become uncorrelated from the previous value.

We have covered some forms of fading effects; however, there are many more. Path loss is a typical case where the loss is proportional to distance. Shadowing is where terrain, buildings, and hills create signal obstructions versus free space, and multipath fading occurs with the recombined scattering and wave interference of a radio signal on objects (due to diffraction and reflections). Other loss includes doppler shifts if the RF signal is in a moving vehicle. There are two categories of the fading phenomenon:

  • Fast fading: This is characteristic of multipath fading when the coherence time is small. The channel will change every few symbols; therefore, coherence time will be low. This type of fading is also known as Rayleigh fading, which is the probability of random variances that cause an RF signal due to atmospheric particles or heavily built-up metropolitan areas.
  • Slow fading: This occurs when the coherence time is long and there is movement over long distances usually due to doppler spreading or shadowing. Here, the coherence time is long enough to successfully transmit significantly more symbols than a fast fade path.

The following figure illustrates the difference between fast and slow fading paths.

Different RF Signal Fading Effects. From left to right: General path loss across a line of sight. Middle: Effects of slow fading due to large structures or terrains. Right: Combined effects of distance, slow fading, and fast fading.  
We will see that technologies employing narrowband signals use what is known as  time diversity to overcome issues with fast fading. Time diversity simply means the signal and payload are transmitted multiple times in the hope that one of the messages gets through.

In a multipath scenario, the delay spread is the time between pulses from various multipath signals. Specifically, it is the delay between the first arrival of a signal and the earliest arrival of a multipath component of the signal.  

The coherence bandwidth is defined as the statistical range of frequencies in which a channel is considered flat. This is a period of time when two frequencies are likely to have comparable fading. The coherence bandwidth Bc is inversely proportional to the delay spread D:

The time at which a symbol can be transmitted without intersymbol interference is 1/D. The following figure illustrates coherence bandwidth for narrowband and wideband communication. Since wideband is larger than the coherence bandwidth Bc it is more likely to have independent fading attributes. That implies different frequency components will experience uncorrelated fading. Whereas, narrowband frequency components all fit within Bc and will experience uniform fading.

Coherence bandwidth and effects on narrowband and wideband. Frequencies f 1  and f 2  will fade independently if |f 1  - f 2|  B c. Here narrowband is clearly shown to reside within B c  and wideband clearly exceeds the range of B c  by some margin. 

One must ensure that the time between sending multiple signals from a multipath scenario. is spread far enough as not to interfere with symbols. This is called Inter-Symbol Interference (ISI). The following figure illustrates a delay spread being too short as to cause ISI. Given that the overall bandwidth B>> 1/T (where T is the pulse width time) and B1/D is implied, we can then generally state that bandwidth must be much larger than the coherence bandwidth:  B>>Bc. 

Delay spread example. Here 2 signals from a form a multipath event. If the delay spread   D is less than the pulse width  T  then the signals may not be spread far enough to override another multipath component. Whereas if the delay spread is large enough there may be no multipath conflict. 
Generally, lower frequencies have a greater penetrating ability and less interference, but require larger antennas and have  less  available bandwidth for transmission. Higher frequencies have greater path loss but smaller antennas and more bandwidth.

The overall bitrate will be impacted by the delay spread. For example, say we use QPSK modulation and the BER is 10-4, then for various delay spreads (D) we have: 

  • D= 256μS: 8 KBps
  • D=2.5 μS: 80 KBps
  • D=100 ns: 2 Mbps
上一章目录下一章