Faraday Consultancy Limited
Faraday Consultancy Limited (FCL) is a private limited company registered in England and Wales, Company Registration Number 2938426, managed and owned by Chris Angove.

An old HF (short wave) transceiver using valves (vacuum tubes). Setting this up for transmission would have been a long process, having to refer to graphs, select crystal resonators and manually tune circuits to appropriate harmonics.

Blog: What Really are LOS, NLOS and BLOS?

Wireless networks, in particular wireless fidelity or WiFi® based on IEEE 802.11, have become very popular over the last few years. Perhaps this is why there is widespread adoption of the expressions line-of-sight (LOS), non-line-of-sight (NLOS) and beyond line-of-sight (BLOS) when referring to a radio channel as 'seen' from each of the antennas.

Consider the 'free space' part of a radio channel, the part between the transmitting antenna (A) at one end and the receiving antenna (B) at the other end. If A can 'see' B then the link is LOS. For example, to communicate efficiently between the earth and a deep space vehicle near Pluto, say a distance of perhaps 5x109 km, a LOS link would be set up using suitable high gain antennas, one at each end, and very accurately aligned. Of course, there would not be much room on the space vehicle but on Earth we should be able to build a big tracking antenna, perhaps with a 70  m reflector diameter. Once set up and pointing at the space vehicle, it would be LOS. That is a fascinating example of LOS and, incidentally, with a propagation time of 4 to 5 hours. By reciprocity, the same free space path loss (FSPL) will apply in either direction, assuming the same frequency. At that distance the received signal would be very weak and only able to carry relatively small amount of information in bits per second. In theory, the actual frequencies used in a deep space LOS link like this are relatively unimportant, so why do we not use a low frequency that we can easily handle like 1 MHz?. After doing the calculations at this frequency, for the free space path loss (FSPL) between the Earth antenna and the space vehicle we get a horrendous value of 226 dB. In fact, the only place I can remember coming across such large dB values is in FSPL calculations. A useful property of antennas is that, the higher the operating frequency, the smaller the antenna needs to be for the same gain. For a reasonable gain the antenna dimensions need to be of the order of at least a few wavelengths. So at 1 MHz, the wavelength is 300 m and it would be challenging to deploy an antenna of any reasonable gain. In such a link, we also need a reliable backup communications path for initial deployment and to re-gain control of the space vehicle, should it's (high-gain) antenna go off-track. Usually this is achieved with several relatively low gain but broad beamwidth antennas orientated such that, no matter what the attitude of the vehicle, there will always be sufficient antenna gain performance orientated towards the Earth.

So why do we not instead use a very high frequency, say something like 20 GHz? We should be able to design an antenna with plenty of gain at this frequency but some of the electronics might be challenging with the current technology, especially the noise performance of the receiver. Another problem is likely to be atmospheric absorption at the Earth's because there is no atmosphere at the space vehicle. So taking all of these factors into consideration, we come up with a sort of 'goldilocks' frequency around a couple of gigahertz.

A very useful property of free space (by definition, a vacuum) is that absorption is zero. So where does that huge FSPL value come from? The FPSL, again by definition, applies to isotropic radiator 'antennas': antennas that radiate equally in every direction, so they have a spherical radiation pattern. So they are: (a) impossible to build and (b) totally hypothetical. Even if we could design them they would not be very useful in space communications because so much power would be wasted, radiating it directions we do not require. An exception is with the backup communications links which we discussed earlier where, actually, we do want something as close to an isotropic radiator that we can achieve. So the large FSPL value is because lots of the power radiated is lost as it is propagated in the wrong directions! Another way of describing this is to think about it as spreading the power into directions where it is not required, and therefore wasting it. This, of course is spatial spreading and not temporal spreading such as we might observe in a frequency spectrum. The largest possible amount of spreading occurs with the isotropic radiator. In fact, this is why it is frequently used as a reference when specifying the gain of antennas: antenna gains are expressed in the logarithmic unit dBi meaning decibels (dB) relative to an isotropic radiator.

Hence we design antennas which are (a) steerable and (b) of the highest possible gain so that we can point it at the Earth. The antenna is normally fixed to the vehicle structure and tracking is achieved by moving the whole body.

One very useful property about free space is that the absorption, again by definition, is zero. But part of the propagation path is not actually free space, the Earth's atmosphere, and of course there will be no atmosphere at the space vehicle.

In the early days of radio propagation at VHF and UHF with, for example, analog terrestrial television broadcasting, it used to be said that the path from the transmit antenna to the receive antenna had to be line of sight (LOS) for a reasonable quality of service. As the television broadcast service became more popular, in many cases it was found that the television audience could often receive an adequate service even when the propagation paths were actually 'not quite LOS'. In these cases there would have been one or more of the mechanisms of diffraction, refraction or reflection contributing to the propagation channel. Each of these in some way affects the apparent propagation path of the signal and therefore our assessment of whether NLOS is possible. In those early days sometimes we might indeed have concluded that NLOS was possible when in fact the phenomenon might have been the exception for a particular location rather than the rule. Conversely, but in a very few cases, the path might have actually been LOS but the signal failed to propagate reliably, if at all. These were usually accounted for by signals from one or more strong reflections being received as well as those via the main direct path from the transmit antenna, adding in anti-phase and cancelling. The television transmission companies came up with the LOS requirement as a reliable 'rule of thumb': something that was readily understood without needing to go too much into technical details.

Radio communications today, 50 odd years later, are very different. The majority of services are substantially designed to be mobile or nomadic. Mobile services require to provide adequate service provision, even if there is relative movement between the transmit and receive antennas. Nomadic services must allow for movement of the transceiver to new positions, but it is not necessary to maintain communication whilst it is actually moving. In trying to maintain such communications in today's typical urban and suburban environments, we cannot expect the LOS conditions to exist substantially so the mobile networks are primarily designed to use NLOS propagation. NLOS services do not just 'tolerate' diffraction, refraction or reflection, they actually exploit them for their normal operation.

Going back to the example of  WiFi, the IEEE 802.11 standard and its various updates was designed to accommodate these NLOS mechanisms. The general term 'multipath' is used to include all the indirect wave paths that the signal takes in getting from the transmitter to the receiver. (It is normally understood that when we talk about transmitter and receiver in the context of radio propagation, we really mean the transmit antenna and the receive antenna respectively. With WiFi they are virtually co-located anyway.) Of course if there actually was a LOS path that would also be accommodated, but here is a third possibility: a wave which takes a direct (straight line) route including transmission actually through one or more media. This direct route will only be successful if the absorptive properties of the media concerned and the transmission distances through the same media are within certain limits.

We are left with the spreading loss or free space path loss (FSPL) which applies to the direct LOS path from the transmit to the receive antenna. This assumes that the atmosphere is a perfect, homogeneous, loss free insulator with no absorption and that there is no influence from waves other than the direct wave originating from the antenna. From plane wave propagation theory applied to the far field, it is found that the power flux density (PFD) of a wave originating from an antenna  will obey an inverse square law. In the direction of radiation, the wave radiates as if the antenna was at the centre of a sphere. The PFD will therefore reduce in proportion to the square of the distance from the source. A receive antenna will have a fixed aperture for a given frequency, so the received signal power level will experience the same inverse square law relationship. This is very interesting but how often do we encounter a situation like this? We might find something similar with ground satellite communications but that is about all. In almost every other case there will be at least some levels of non-direct or NLOS involved.

Let us think about each of these four mechanisms (reflection, refraction, diffraction and absorption) in the context of NLOS.

Reflection or, more precisely, dielectric reflection occurs when a propagating wave in a medium of a particular intrinsic (wave) impedance meets another medium which has a differing intrinsic impedance. The extent of both media and the boundary between them must be in the order of 'many wavelengths'.  (A rule of thumb for 'many wavelengths' is usually taken as about 10 or more.) The angle of reflection will be equal to the angle of incidence and the rays representing the incident and reflected waves will be in the same plane. The intensity of the reflected wave will vary according to the dissimilarity of the refractive indices of the media: the more dissimilar they are the greater the reflected intensity will be. There will also be a transmitted wave into the medium on which the wave is incident. The intensity of the transmitted wave will bear in inverse relationship to the reflected wave: the more dissimilar the media the smaller the transmitted wave will be. There will always be a degree of reflection associated a degree of refraction. A special case of reflection occurs when the second medium is in fact a good conductor, such as a metal. From Maxwell's equations, a wave cannot exist in a perfect conductor so it is totally reflected. The intensity of the reflected wave is almost identical to that of the incident wave. This type of reflection would occur if the wave was incident on a metal framed building, water tank or bridge, for example.  

Refraction is observed under similar circumstances to reflection: when a wave passes from a medium of one intrinsic impedance to another medium of a differing intrinsic impedance. Usually the two media may also be described by their (differing) refractive indices or dielectric constants. (Dielectric constant may also be known as relative permittivity.)  However, at the boundary between the two media, the phenomenon of refraction describes how the wave is transmitted into the second medium and may be described by Snell's Law. As with reflection,  the extent of both media and the boundary between them must be in the order of many wavelengths. In radio wave propagation through the atmosphere, refraction usually occurs with a wave propagation through differing air densities, caused by differing temperatures, altitudes, or ion densities.

Diffraction of a wave may occur around an obstacle to an extent depending on the electrical and physical nature of the edge of the obstacle related to the wavelength itself. Diffraction is most pronounced when the edge of the obstacle is irregular and the irregularities have dimensions of the order of wavelengths. If the edge is smooth, the level of diffraction tends to reduce. The phenomena has been studied for many years since its first discovery in the context of visible light waves by Huygens in the 17th. century. Where diffraction does occur, the diffracted wave is formed from the sum of a number of small wave sources at the edge of the obstacle. A diffracted wave can therefore appear colloquially to 'bend' around an obstacle. This apparent bending is actually the result of the summation of several apparent wave sources around the obstacle.