Non Line of Sight Point to Point Wireless Backhaul

Does an outdoor wireless point to point bridge require Line-of-Sight ("LOS") or can a quality wireless Ethernet bridge perform under Non-line-of-Sight ("NLOS") conditions? LOS is when both antennas in a outdoor wireless bridge system must have clear visibility with one another and have no encroachments to the first Fresnel Zone. In a NLOS situation there is either limited visibility from one wireless antenna to the other (near-line-of-sight or" nLOS") caused by a Fresnel Zone encroachment or complete obstruction blocking the visibility between the two wireless antennas. 

The answer depends on the individual path and the throughput requirement. In many cases, if the wireless design and wireless installation is done properly a quality NLOS microwave link will provide good quality high bandwidth.

General Overview of Point to Point Wireless Backhaul: 
A typical outdoor wireless backhaul is used to pass higher throughput over greater distances. Outdoor wireless bridges operate in the SHF (Super High Frequency) band in unlicensed wireless backhaul 5.3GHz, 4.9GHz, 5.4GHz, 5.8GHz, and 24GHz or licensed microwave backhaul 6GHz, 11GHz, 18GHz, and 23GHz. There is also unlicensed 60GHz and registered 80GHz millimeter wave in the EHF (Extreme High Frequency) band. The unlicensed wireless Ethernet bridges typically provide from 10Mbps to 300Mbps aggregate throughput. Unlicensed 24GHz and licensed microwave links offer up to 360+Mbps Full Duplex. 60GHz and 80GHz wireless bridge systems can provide up to GigE Full Duplex (gigabit wireless). The higher frequencies do not do well with penetrating obstructions. 

For an outdoor wireless bridge to work the system gain must be greater that the total Path Loss. Historically, an outdoor wireless bridge required LOS providing first Fresnel Zone clearance. By having no obstructions in the first Fresnel Zone the receive signal are optimized and the out of phase signals are minimized.

General Overview of Non-Line-of Sight Wireless Bridges
When considering a point to point wireless backhaul, whether a licensed microwave link or an unlicensed wireless Ethernet bridge, one of the first questions asked is if there needs to be LOS to get a microwave link. Many don't understand the difference between wireless bridge technologies that they are use to (like cellular and cordless phones or Wi-Fi) compared to an outdoor point to point wireless Ethernet bridge.

Devices like cellular operate in a range from 800MHz to 1900MHz of the UHF (Ultra High Frequency) band. These frequencies do well with penetrating obstructions but have limited throughput capabilities. Most Wi-Fi operates in the 2.4GHz frequency of the UHF band and can provide higher bandwidth but is very limited in distance. Microwave communication signals are highly attenuated by an obstructed path. In a NLOS microwave link the RF signals will get to a destination by: diffraction around an object, reflection off objects, or by penetration through the obstruction. 

For an outdoor wireless bridge, being used for high bandwidth, point to point backhaul to work in a NLOS application there are several requirements that need to be met. Proper power budget, fade mitigation, adaptive link characteristics, and proper demodulation in regards to dispersion. Because of obstructions in a NLOS situation there tends to be a large amount of multipath. Obstructions like trees add to multipath and add attenuation to the overall Path Loss of the microwave link. Trees can be tricky because they are not constant due to movement caused by wind, foliage changes during various seasons, moisture content of the foliage, etc. Constant obstructions like buildings or hills are easier to model and predict. 

General Overview of NLOS Wireless Technology
Current wireless backhaul technologies can help in NLOS cases. MIMO (Multiple Input Multiple Output) antenna signaling and spatial diversity reduces the amount of fade margin required. OFDM (Orthogonal Frequency Division Multiplexing) which divides the data into several parallel data streams helping the fading that occurs with multipath. Adaptive rate modulation also helps by giving the wireless backhaul radio the ability to manage the modulation scheme and bandwidth according to the RSL (receive signal level) optimizing the microwave communication link. Outdoor wireless bridges that can take advantage of these wireless backhaul technologies are the unlicensed wireless systems. Unlicensed wireless backhaul using these technologies can provide up to 300Mbps aggregate throughput (depending again on the characteristics of the microwave link path).

A common question of why a licensed microwave link, which can provide higher, full duplex connectivity, doesn't use OFDM wireless or MIMO antenna solutions and why they can't be used in NLOS (non line of sight) applications. In a NLOS wireless link application point to point wireless Ethernet bridge radios that use OFDM or MIMO take advantage of multipath for their connectivity. Because a licensed microwave link is not to inject any interference on other licensed microwave backhaul operators in the area they must have LOS (line of sight) and not cause heavy multipath. If a licensed microwave radio was to cause a lot of wireless multipath it could potentially reflect into another existing licensed microwave communication radio belonging to another party.

Prior to considering a NLOS wireless backhaul, a wireless site survey and a proper wireless path calculation should be performed. Field test may need to be performed in order to verify if a NLOS microwave link will work or to gather accurate estimates on throughput performance. As with any point to point wireless backhaul, a certified expert should perform the wireless installation.

Add from Glenn De Haes
 .Let me add to that some field experience. OFDM works by all means better in higher frequencies due to the higher reflective coefficiences of materials on those frequencies. This means a NLOS system with OFDM in 5GHz works better than in 2.4GHz. You also notice this even with wifi indoor systems using 11g or 11a.

A key thing with NLOS is the ability to catch as much signal as you can. It is thus easier to have a NLOS link running with two sector antennas instead of parabolic dishes.
MIMO is a nice addition to this since you will gather signals from multiple polarities and with a larger delay spread.

What most people also forget, is that in order to profit from reflections, you first of all need them. NLOS in a city envirnment is thus a bit easier then when there are other obstacles likes trees in the way. These things only absorb power and hardly reflect anything.

We see mostly examples from manufacturers with trees. Mainly to explain the difference between LOS, near LOS and non LOS. Take it from me, OFDM, MIMO whatever, if there are trees in the way, the only thing that can help you is power. We have had installations where we tried 2x2 versus 3x3 MIMO and since you have less gain in an antenna of the same size in 3x3 versus 2x2, the result in most circumstances was worse. 3x3 MIMO does give you a better path fading result so ideally for implementations in cities or amongst buildings; even in LOS.

Bottom line: ALWAYS test a NLOS up front and make sure that your reflections are trustworthy and not originating from something only temporarily available.

Search Interfering in a Radio Link

A short guide to search for interfering links in a Radio Link

Introduction and purpose
With the advent of next-generation networks becomes more and more interference must be searched to verify the availability of frequencies and / or channels are free to set up, expand or replace links in Radio Link

 Network.
This article is intended as a guideline introdutiva to run a check for the research of any interfering signals on the link.
We can identify a dual purpose for this type of activity:

• Search free channels in order to properly allocate the frequencies for the new radio link (or expansion of existing ones).
• Verify potential interferents which may affect the functionality of the radio link connection to existing or to be installed.

Normally the tests are to be carried out in all the stations on the radio link.

Instrumentation and Tour 
By way of reference, we can identify the following types of instrumentation:

• Horn Antenna type reference (in many cases, also known as "piccolo trumpet").
• Low-noise amplifier with accessories (cooling fins, and accessories for DC power supply).
• Spectrum analyzer.
• RF cables in the correct frequency ranges.
• Transitions and steps in relation to the various RF outputs present on the antenna, cables and tools.

One can imagine a scheme of the bench as shown in Figure:
Some specifications:

• Antenna: gain of about 20 dB and angle of irradiation at 3 dB, between 20 and 30 degrees.
• Spectrum Analyzer: Model similar to or higher than the Agilent HP 4408B ESA. In any case, with a minimum level of noise less than -100 dBm.
• Amplifier: Gain of about 30/40 dB, noise figure: within 3 dB
  + Additional accessories such as cooling fins, and DC power
• RF Cable: Cable, low attenuation.
• Series of RF coaxial transitions: According to the specifications terminations on the cables and instruments, typically N-type (m) / N (m), N (f) / N (f), N (m) / SMA (f), etc ... .
• DC power source to the amplifier.
  Typically between 9/15 Volts and 70/300 mA (depends on the specific model).
• Also make sure to have compass, binoculars and support with goniometer for the antenna.

It 'important to remember that a lack of quality in the choice of instruments, in particular the spectrum analyzer and the low noise amplifier, can adversely affect the measurement results.
Have at hand the specific ITU-R of reference, in particular frequency bands correct and funnels RF.
Useful to know some specifications of the radio system that will be installed, in particular the thresholds of degradation and specifications of noise, in order to understand if a measured signal can be considered or not, can interferent.

Installing and Configuring Tour
If possible, it is preferable to carry out the measures will be installed at the same height where the parable of the new radio link.
If this is not possible, stand still as close, otherwise there is a risk of obtaining unreliable results, interfering in how some might be "masked."
Install the antenna Horn preferably on a rotating support (possibly provided with a disc goniometric).
The support system (for example on a tripod) must allow easy rotation of the antenna.
Enable the counter and configure the spectrum analyzer as references below:

• Start Freq.: According to RF channel to be analyzed
• Stop Freq.: According to RF channel to be analyzed
• RBW and VBW: 300 Khz
  (It is suggested to keep the same value in order to simplify the calculations and measurements of power)
• Sweep: Automatic
• Attenuation: 0 dB
• Reference power: -10 dBm dB
• Scale (dB / Div): 10 or 5 dB / Div

The values ​​above are for reference, but typically you will get a background noise appears around - 60/-70 dBm.
The use of a line of aplificazione least 30/40 dB serves precisely to be able to display on the instrument signals having powers even of the order of - 90 dBm.
I would suggest making sure that the sum of the gains of the antenna and the amplifier is roughly about 50 dB. This in order to have a sufficient dynamic detection on the part of the instrument.

Regarding the handling of the antenna, it will perform the measures on both polarizazioni and preferably on an angle as broad as possible.
The analysis at 360 ° (or less) depends on the specific objective of the measure.
Consider, however, that the antenna has an angle of detection of the order of 15/20 degrees, so you might divide the research in most areas around the main direction.
The figure below shows an example.

Sequence of controls and measures
The control will follow 4 main phases:

• Control in the same direction of the remote station in polarization V
• Control in the same direction of the remote station in polarization H
• External control the direction of the remote station (coverage with N segments) in polarization V.
• External control the direction of the remote station (coverage with N segments) in polarization H.

For each field and polarization, the operator must:

• Place the antenna in a reference position.
• Configure the "range" of the frequency spectrum analyzer. Measurement bands may be divided into sub-bands so as to accelerate the speed of the measure (are suggested guideline values ​​around 200/300 Mhz according to the width of the pipe).
• Save the "screens" of the instrument for each measurement by placing a reference (marker) on the maximum value of power detected (*).
• Repeat the measurement for both polarizazioni (H and V)
• . In case of detection of specific interfering signal (or any signal "significant"), collect data of the same in terms of bandwidth / power and if possible identify the main direction.

(*) The detection of received power carried out by a mere reading of "Marker", may be subject to considerable error due to the configuration of the same instrument. The video signal is subject to variations dependent on the Resolution Video, The Band, Sweep, attenuation etc ...
Therefore, the power should be measured with specific functions available on the analyzer or re-calculated with appropriate equations.
I will take short to make any picture on this issue.
When in doubt, take note however (along with recordings of the signals this graphic) of average power values ​​of the measured signals. The necessary corrections may be made at a later time.
Remember, however, that once identified the optimal configuration of the instrument, DO NOT change it for all phases of research, could make an exception specific analysis of an interfering signal significatico.

Analysis signals
Place the antenna on the same azimuth reference provided for the remote station.
Remember that if you use a compass, it can undergo magnetic deviations due to metal structures.
Moreover, in case if we are identifying the directions of remote sites and have only the geographical coordinates, let us remember that in some areas the deviation geografice Magnetic North - True North may be relevant (in Italy is not high).
It is therefore, useful to take the absolute references that can be used subsequently to detect the correct angles and possibly correct the deviations of the compass.
Divide the channeling RF to be analyzed into N sub-bands from a few hundred Mhz to analyze them and start up the exhaustion of the entire RF channeling request.

• Scan in both vertical and horizontal polarization, taking care to also search in elevation.
• For each scan store the track and detect / measure the signal indicated by the instrument. If the spectrum analyzer is provided with a specific measurement option, activate and detect the results. Fill in the relevant documentation required measurement.
• In case of detection of interfering signal (or any signal "significant), detect the data of the same in terms of bandwidth / power and if possible to identify the main direction.
• Once the measure in the first sector, using as a reference the angle of radiation of the antenna and turn it to the right (or left) of N multiple segments of the angle.
• Repeat the sequences described above for each angular segment.

Signal significant
An interfering signal can be considered such if it degrades the threshold BER systems of at least 1 dB. The value of the signal interfering signal will thus have a different weight in relation to several factors dependent on the specific product used to make the radio link (bandwidth, modulation, frequency, threshold, received power etc. ..).
It will thus be difficult to define a priori a proper assessment interference without knowing the characteristics of the Radio Link.
In principle, you could define a rule of thumb that considers all potential interfering signals signals that may have detected a real power greater than - 90 dBm.
So each modulated signal (or CW signal) having the above characteristics could be considered potential interferent and reported to the authorities responsible for the study and planning of radio systems.
If during a scan to locate a particular signal defined as "significant", you must:

• Detect the data of the interfering signal in the normal conditions of measurement (same parameters of banda and canalization used for the verification interferential).
  The detection of the signal must be highlighted for both polarizations.
• At the end of the interferential scanning, center frequency in the detected signal and highlight highlighting (providing data on the relevant documents) the parameters of bandwidth and power.
• If possible, try to identify the source (website, Azimuth, etc. ..)

When these scans organize all saves the images stored by the instrument, organizing them for direction, polarization and band, possibly accompanied by photographs of the reports findings.

The following are some examples of paths without interfering with potentially interfering signal and the same feature.
I will provide to make a further Article with a practical example of calculation of the signals.

Thank you to my friend and Big Telecom Technician Gaetano Elifani for documents.

Power measurements with a spectrum analyzer

Will not go into the merits inherent in the specification or the detailed characteristics of a spectrum analyzer, but I will try to introduce only the elements relevant to the execution of a measure of power (assuming no measurement options normally found in these types of tools Already for some years, such as the measurement of called "Channel Power").

The spectrum analyzers, contrary to power meters (also called bolometers), measure signals in the frequency domain and on a specific portion of band configurable.
Fundamental elements that characterize the result will be reflected in the measure are the resolution bandwidth (Resolution Bandwidth (RBW)), the Band Video (Video Bandwidth (VBW)), the speed sweep (sweep), and of course, the intrinsic specifications of the instrument.

Therefore you can easily misinterpret the readings, as it is very dependent on the configuration of the instrument (and its characteristics).

Consider the following:
If we measure a signal with very narrow bandwidth (essentially a pure unmodulated carrier, CW)with a power of -30 dBm, the reading with bolometer (measuring instrument broadband) will correspond to the same value; the same signal (CW) measured with a spectrum analyzer set at a ratio RBW / VBW = 1 will give approximately the same value of the bolometer and that is - 30 dBm.
In this case in fact the width of the carrier band CW is so small, therefore the risk of error in the reading on the analyzer is minimized.

If instead we measure a signal having a bandwidth of more substantial (eg a modulated signal in phase with bandwidth of 30 MHz), the bolometer will always give us the same value, while the spectrum analyzer will give us the reading given by "+ Prx | J0 |. "
Where J0 is the error due to the distribution of power density that depends on the ratio between the bandwidth of the signal to be measured and the resolution bandwidth set on the analyzer used for the measurement.
In our case the error potra be approximately calculated with the following formula:
Assuming the example above, and a 3 MHz RBW (RBW = VBW with), we have:For which we will read in a video signal with a power of less than 10 dB compared to the actual value of power read with a "Bolometer", therefore:
If the actual signal is - 30 dBm spectrum analyzer can give me a reading of about - 40 dBm.
Naturally more shake bandwidth and most get worse the spectral density ratio between the signal to be measured and the bandwidth used to measure it. By way of example, with the same signal but with a RBW of 300 Khz will be calculated approximately 20 dB of error.
There is, therefore, to consider the use in some cases, a bit 'naive and misleading that is made ​​of spectrum analyzers, especially when they are to be used as a power meter. 

By way of example, if one analyzes a signal having the characteristics:

• Center Frequency = 7471.00 MHz
• Span = 56 MHz
• Ref: -74 dBm
• dB / div = 3 dB
• RBW = 10 kHz
• VBW = 10 Khz
• Marker = Measured Signal Level - 94 dBm

Assuming a measure applied to a reference signal of 14 MHz of bandwidth (typically an 8 PSK modulation on a radio system PDH), we will have that the real value of equivalent power will be:
For which the level that displays the analyzer will be lower than 31.46 dB with respect to the real value. The actual value will be - 62.54 dBm.
The above also becomes essential if the goal of the measure is to seek real signals of the order of -90 dBm. In fact, with the above described conditions will be impossible only with the spectrum analyzer to measure signals with intensity so low.

In case of search for signals modulated by a large bandwidth, is not expedient to reduce the resolution bandwidth of the instrument, but should keep it out of relatively large values ​​and act on the sensitivity of the measurement, by using low noise amplifiers (LNA Low Noise Amplifier) ​​with low level of intermodulation and a high gain antenna. The use of antennas with a gain of at least 15 dB and low noise amplifiers with amplification medium of about 40 dB then make the checks possible.
In the calculation of the received power equivalent, in the case of checking interference must be considered as a value of Band B, the following rules. 

• If you do the calculation and the radio link to be used, use the bandwidth of the system to be installed.
• In case you want to make a specific measure of power of a signal, in this case one must consider the overall bandwidth of the signal being analyzed.

In interferential, total estimates of the interfering power () will therefore:

Where:
The above can be considered valid with an accuracy of a few dB to 1 but maintaining the ratio between RBW / VBW (ie assuming RBW = VBW).

Otherwise you can easily introduce errors of assessment even of the order of a few tens of dB.

Useful demos and remote control of a spectrum analyzer, Agilent is available on the website:

http://wireless.agilent.com/videos/econtent/Remote/

Feature Sensitivity test bench and in the analyzer bench Measure for Research interference, the quality of the amplifier LNA and the spectrum analyzer have a major impact on the overall quality of the measures.
In a generic way we can verify that the sensitivity instrinseca of the spectrum analyzer at a given resolution of Banda (RBW) turns out to be: 
Where "Nsa (RBWrif)" and the average sensitivity of noise stated by the manufacturer to a given value of RBW reference (Displayed average noise level).
By way of example for spectrum analyzers Agilent HP 4408B ESA the reference values ​​may be:

• 10 MHz to 1.0 GHz ≤ -116 dBm
• 1 GHz to 2.0 GHz ≤ -115 dBm
• 2 GHz to 6.0 GHz ≤ -112 dBm
• 6.0 GHz to 12.0 GHz ≤ -110 dBm
• 12.0 GHz to 22.0 GHz ≤ -107 dBm
• 22.0 GHz to 26.5 GHz ≤ -101 dBm

The above measured with a RBW reference at 1 KHZ and VBW of 30 Hz and a level of -70 dBm.

If we consider, therefore, that the measuring bench for interfering research will consist of an antenna and an amplifier Horn reference LNA, more of the same spectrum analyzer, interference minimum level will be measured:

Conclusion
Let us remember that in case of measurements of modulated signals the power value displayed by the instrument will be lower than the real power; there will be to perform a correction that will depend on the bandwidth used by the analyzer to measure the signal and the real bandwidth of the signal (or band equivalent to consider).
In order to correct this error, the value of Pm (Power measured) will be: 

Where:

• Pm = equivalent power measured (basically the input power).
• Prx = equivalent power reading on the monitor of the instrument (average value measured over 100 points).
• B = equivalent bandwidth RF signal on which the study is carried interferential (specific band of radio product) or bandwidth of the RF signal measured
• RBW = resolution bandwidth set on the spectrum analyzer

The formula above, allow with a good approximation to calculate the effective power in ingesso to the instrument. The above is true and only the values ​​of RBW and VBW are equal (in the case other than the error could also be of some tens of dB).

One must also consider the lines of amplification and attenuation of the bench.
Whereas the test bench for a search interfering is composed of elements Amplifiers (antenna and amplifier) ​​and elements Attenuators (cables and adapters), considering negligible errors due to other factors, the equivalent value of Pi interfering power (dBm) measured in the different sections of measurement, is determined using the following basic formula:
where:
Pm = equivalent power measured.
Ga = antenna gain reference.
Glna = Gain Low Noise Amplifier.
Ac = total attenuation of the measurement leads and transitions.

Thank you to my friend and big Telecom Technician Gaetano Elifani for the documents.