140 Knowles Drive, Los Gatos, CA 95032 Tel: 408-399-7771 Fax: 408-317-1777 http://www.firetide.com Introduction to Basic Reflective Multipath In Short-Path Wireless Systems DISCLAIMER - This document provides basic guidelines and practices for the diagnosis of Firetide mesh performance issues. It is not a substitute for Firetide Certification Training. Integrators of Firetide products must address the specific requirements of each project in their system designs, and should be Firetide Certified to assure competent system designs. Firetide, Inc. is not responsible for any system design using Firetide products unless specifically contracted to provide such design. J. Frank Jimenez Director, Solutions Engineering Firetide, Inc. May 2, 2013 Revision 1
Introduction to basic reflective multipath Page 2 of 7 Introduction Reflective multipath is commonly overlooked as a factor in wireless system designs involving short paths at microwave frequencies. Many inexperienced wireless system integrators simply perform link budget calculations to determine how far their system should be able to communicate considering only free space loss between the antennas. The only parameter that they are considering is what will my RSSI be? More experienced wireless system integrators are also aware that at microwave frequencies, RF line-of-sight requires signal clearance beyond simply that necessary to establish visual line-ofsight. The criteria for RF line-of-sight, is visual line-of-sight plus additional clearance for 60% of the first Fresnel zone, commonly referred to as 0.6F1. If this amount of clearance is not provided with respect to all path obstacles, each violation will introduce additional diffraction loss. Since the effect of signal clearance violations is cumulative, these can add up quickly, consuming calculated fade margin resulting in a system that will not work reliably if at all. Knowledge of the above can lead to systems being designed and deployed with excessive antenna height if the integrator does not perform the proper RF path engineering to determine proper antenna height. In some cases an integrator may know he needs additional clearance for Fresnel zone(s) but not how to determine the specific clearance. So to make sure he does provide enough clearance, he installs antennas as high as he can. This is typically where most reflective multipath problems begin. In other cases, the integrator has not performed a proper path survey to identify locations and confirm heights of path obstructions, so that proper antenna heights can be determined. Without a path verification survey, it is impossible to engineer the proper antenna heights for a system. The result is implementation of incorrect antenna heights resulting in either diffraction loss or reflective multipath problems. Reflective multipath problems can occur on paths of virtually any length, from links between buildings across a street or parking lot from each other, to links that are miles long. The example below could involve a short path just a few hundred feet long. Path reflection point
Introduction to basic reflective multipath Page 3 of 7 Simplistically, the above reflection problem could be resolved by relocating one of the antennas toward the back of its building to block its line-of-sight view of the path reflection point while maintaining line-of-sight to the opposite antenna. Another means of resolving the previous reflective multipath problem would involve adjustment of antenna heights. If the reflected signal is received perfectly in-phase with the direct-path signal, the power of the two signals is summed without any distortion, just as two singers singing in perfect harmony. This is where Fresnel zones become important. Fresnel zones explained A signal reflecting off of any even-numbered Fresnel zone will end up at the receive antenna exactly 180 degrees out-of-phase with the direct-path signal, causing severe phase distortion, and depending on its strength, possible signal cancellation. A signal reflecting off of any odd-numbered Fresnel zone will end up at the receive antenna exactly in-phase with the direct-path signal, resulting in no phase distortion, and its power will add to the power of the direct-path signal resulting in an increase in signal strength (think two singers singing in perfect harmony). Because the scope of this paper is to provide you with a high level understanding of reflective multipath at the conceptual level, we will avoid mathematical formulas or technical details. This information is available in another paper titled Fundamentals of Radio Link Engineering. This paper is available without cost for those interested in a more in-depth explanation of the physics and mathematical formulae involved. Since this paper does refer to Fresnel zones, I will provide a simple word definition of what constitutes the first and subsequent Fresnel zones. Fresnel zones are defined by a series of radial points offset from the direct signal path, along the entire length of a signal path. In the case of the 1 st Fresnel zone, these points occur where if a signal transmitted from one end of the path were to reflect off of, the reflected signal will have traveled exactly one-half wavelength (180 degrees) farther in distance than the direct-path signal when it arrives at the receive antenna. In the case of the 2 nd Fresnel zone, the offset distance for these data points increases to add another one-half wavelength to the reflected signal path distance compared to that of the 1 st Fresnel zone. The total distance of travel for a 2 nd Fresnel zone reflected signal would therefore be a full wavelength longer than the direct path signal.
Introduction to basic reflective multipath Page 4 of 7 Each subsequent Fresnel zone simply adds another one-half wavelength to the reflected signal path distance with respect to the previous Fresnel zone. Signal reflection explained It is important to understand the phase shift aspect of the signal reflection process because without the inclusion of its phase effects, even-numbered Fresnel zones that add exactly one full wavelength to a reflected signal would not be a problem. The reflection process inverts the signal, effectively adding 180 degrees (or one-half wavelength) to the reflected signal. This then, added to the full wavelength contribution of an even-numbered Fresnel zone results in a 180 degree opposed reflected signal at the receive antenna. The drawing below illustrates how the top of an incoming signal prior to reflection becomes the bottom of the outgoing reflected signal. The path reflection point Knowing the location of the path reflection point is required in order to determine whether a reflected multipath problem can exist. In the drawing on page 2, the reflection point was simple to identify, because the antennas were at the same height and the terrain in-between flat, the reflection point was exactly half-way between the antennas. If one of the antennas were mounted lower than the other, the reflection point would be located closer to the lower antenna than the higher one. The problem is simply a geometrical one related to the height of the antennas at either end of the path with respect to the reflecting plane (usually the ground surface). The primary path reflection point that must be considered is the specular reflection point. This is generally a single reflection point along the path where the reflected signal would end up exactly at the receive antenna. Signals with a steeper or flatter angle of reflection would either overshoot or undershoot the receive antenna.
Introduction to basic reflective multipath Page 5 of 7 It is important to understand that it is possible for more than a single reflected signal to end up at the receive antenna. In such cases one or more reflective points along the path could be sloped such that a reflected signal off of these points will deliver two or more reflected signals to the receive antenna. This is also common in metropolitan deployments where buildings on either side of a signal path could contribute multiple reflected signals. The graphic below shows how terrain surface angle can cast reflected signals in different directions. Irregularities in the slope of the terrain along a path could result in multiple reflected signals from different points ending up at the receive antenna. Receive Antenna Since the scope of this paper is limited to providing an explanation of what reflective multipath is, and how to recognize it as a potential root cause of system problems, we will not cover the mathematical and engineering aspect of reflection analysis and mitigation. Some system symptoms related to reflective multipath on fixed systems. In wireless systems, reflective multipath can cause any of the following symptoms. 1. Significantly stronger or weaker signal (RSSI) levels than link calculations indicate. 2. Inability to align antennas (constant low signal level). 3. A constant higher than normal packet error rate (wireless packet retries). 4. A signal level (RSSI) that varies by several db on its own over a time period. 5. Intermittent wireless link dropouts at certain times that recover on their own.
Reflective multipath in mobility systems Introduction to basic reflective multipath Page 6 of 7 Reflective multipath signal nulling is especially a problem with mobile systems, which involve RF paths that are continually changing length. As the path distance between a fixed cell node and a mobile node changes, so does the radius of its Fresnel zones. As a mobile node moves toward a fixed node, its Fresnel zones diminish in size as distance decreases. This is the root of the problem. As Fresnel zone radius increases or decreases, different numbered Fresnel zones line up with the reflecting plane (typically ground surface) as shown below in the following two plots. Notice that in traveling from a distance of 300 feet (from 500 feet to 200 feet from a fixed node); Fresnel zone radius has diminished by 10 Fresnel zones, increasing path clearance from the 7 th to the 17 th Fresnel zone. This means that as the mobile node traveled the distance involved, 5 signal nulls would have occurred, as the ground surface corresponded with the 8 th, 10 th, 12 th, 14 th, and 16 th Fresnel zones. Notice that where the 7 th Fresnel zone touches the ground surface, the spacing between Fresnel adjacent zones is just under one foot, compared to the blue dashed line indicating the spacing between 60% of the 1 st Fresnel zone and the direct signal path.
Introduction to basic reflective multipath Page 7 of 7 Notice that where the 17 th Fresnel zone has cleared the ground surface, the Fresnel zone spacing is now just inches apart. As Fresnel zone numbers increase, the spacing between Fresnel zones diminishes, creating more frequent occurrences of signal nulling. Summary As we have covered in this paper, reflective multipath is a very real issue that contributes to mystical problems in wireless systems. In fixed systems, it can result in unstable system operation with sporadic outages or performance degradation problems that can mistakenly be blamed on equipment. Worse yet, some system issues are so unpredictable in terms of occurrence and duration, that they become difficult to resolve since by the time that someone goes onsite, the problem has corrected itself. This is particularly true on longer RF paths where the physics of atmospheric refractivity become a significant factor. These physics will be covered in a separate paper focused on explaining reflective multipath issues involved in long-path wireless systems. Reflective multipath is the main reason why it is so important to properly document every system with as built drawings that accurately document radio link frequencies, site coordinates and antenna heights above ground level. At least with this information, it is possible to analyze the system design for potential reflective multipath root causes should problems occur. A new and growing application for broadband wireless systems is deployment in mobility applications. As we have shown in this paper, reflective multipath is a primary consideration in the design of a mobility system. If the system has not been properly designed, the system will have built in outage mechanisms that cannot be resolved without great expense and time.