The World’s Longest All-IP Microwave Link

Successfully Implementing a 193-km Microwave Link over Water to Deliver 99.9995 Percent Availability


For most designers of microwave transmission paths, engineering a reliable link over water can be a daunting task. Reflections off the water surface can play havoc with the received signal, leading to high levels of interference resulting in fading and ultimately a high level of errors and signal interruptions. For these types of paths, performance calculations using commercially available software planning tools will be insufficient to ensure superior path performance. In these cases, experience and understanding of the key parameters that influence microwave performance are critical.

Recently, Aviat Networks and our agent, Telecomunicaciones y Sistemas S.A. (TELSSA), deployed an Eclipse microwave link for Central American Corporation for Air Navigation Services (COCESNA) in Honduras that crosses over 193 km, most of which is over water. With careful design and installation, this link is now operating successfully.

COCESNA is responsible for the air traffic control over Central America territory and oceanic areas, therefore, availability of service is a critical issue.

Key Design Considerations

A key design criteria for designers to keep in mind is the prevalent propagation conditions to meet or exceed the expected Error Performance Objective (EPO), particularly for the design of very long terrestrial microwave systems traversing large and warm bodies of water, like the ones prevailing in Central America. To meet the specified EPO for any path, designers must have a thorough understanding of the following key aspects:

a)    The effect of antenna diameter
b)    Precise alignment
c)    Accurate diversity spacing to counter reflections
d)    Deterministic prediction of the variations of atmospheric conditions
e)    Multi-path propagation delay

With these parameters in mind, achieving a microwave path that delivers fiber-optic level error performance over water is possible. The mathematical models to predict the link’s EPO don’t have a meaning if the design variables above are not fully grasped, respected and implemented.

Understanding Local Propagation Conditions

Understanding the behavior of the atmospheric conditions determines the type of design needed to accommodate how the link will perform over time in any region, regardless of the temporal, seasonal and climatic variations along the path. In Central America, with its warm days and nights, standard propagation is prevalent, because the troposphere in this region is “assumed” to be thermodynamically unstable and in a state of constant turbulent mixingducting (a key propagation impairment over long paths) has not been observed.

Under these conditions and based on Aviat Networks’ experience, long (exceeding 80 km), over-water, semi-horizontal and windy paths exhibit excellent performance, even with relatively low ground elevations above mean sea level at both sites.

The Design

For the 6GHz, QPSK, 40 Mbps, 193-km path in Honduras, which was more than double the length of a typical long microwave path, experience, propagation considerations and exact installation were crucial in the design. Even though the path inclination was relatively small, Aviat Networks’ prediction of the propagation conditions’ “soft” fluctuations along the path determined that the link was viable. To span such an enormous path, sufficient site elevation is a must to prevent diffraction or blockage. In this design, Site A, “Las Cruces,” is at 1600 meters above mean sea level and Site B, “Dixon Hill,” at 250 meters on Roatan Island. The reflections and delay were optimally neutralized with precise antenna spacing and up-tilt.

Microwave Path Profile

Figure 1. Microwave Path Profile showing antenna elevations and path clearance over effective earth curvature.

To make the design even more unusual and challenging, the short radar tower facility at Site B Dixon Hill lacked space to mount a second space diversity antenna. As a result, Aviat Networks/TELSSA installed the second antenna on a separate structure, located 15 meters in front of the main antenna. Consequently, the design had to take into account and negate any additional signal delay arising from the horizontal antenna offset.

Diversity Antenna Installation

Figure 2. Diversity antenna installation showing 15 meter horizontal separation at the radar site Dixon Hill.

The resulting link was designed to provide an overall availability of 99.9995 percent availability, equating to just 32 seconds of accumulated outage every year.


Once the design was done, the challenge was not over. Ensuring superior link performance took careful and exact installation, with precise alignment of the antennas by an experienced crew to achieve a stable received signal level. A few days after the initial alignment, the crew revisited the 1600-meter site to adjust the antenna installation to counter very strong wind shears that threatened to cause instantaneous twisting of both antennas that could degrade performance.

Site A 1600m elevation

Figure 3. The view looking out from Site A, Las Cruces, (elevation 1600 meters) across the water.

The Result

With the design and installation complete, further adjustments were made for local weather conditions. The microwave link was then monitored over several days to measure actual performance to verify that it met or exceeded the design requirements.

The measured unfaded link RBER (Residual Bit Error Rate) performance test showed a result of 1×10-13, or just one bit error for every trillion bits received. This level of error performance is equivalent to what you would expect if you connected two Eclipse terminals on a bench, directly connected by a cable! The link is also exceeding the 6-9’s availability target.

“This IP link was a challenge for us from the moment of the project’s inception, but we have always relied on link engineering support from Aviat Networks to complement installations carried out by our engineers and technicians—who have great experience in complex hops—which has led to excellent results,” says Jose Raul Gomez, CEO TELSSA.

At 193 km, this Eclipse link is probably the longest all-packet microwave link in the world and demonstrates that making challenging paths work well takes experience, superior product performance and quality, careful design and precise installation, with results that canand did, in this caseexceed customer expectations.

-Ivan Zambrano
Senior Network Engineer
Aviat Networks


    Sunil Kumar May 4, 2011/

    Great achievement. Well done. Proud to be an Aviat Employee.

    Shibly Sadik Al Zaidy May 4, 2011/

    Obviously a great achievement! Wish I was working with that installation team!
    I was always dreaming to do this kind of precise installation job.
    Excellent!!! Feeling very proud to be an Aviat employee!

    William Leclerc May 5, 2011/

    Dear Ivan,

    Your “news” is abolutely great. I love it!

    Peter A Coward May 5, 2011/

    Fantastic technological achievement, fascinated and intrigued when reading this blog, by Aviat’s true genius.

    Tony Maietta May 5, 2011/

    Great job. I’d like to see the pathloss calculations.



    Mirza Baig May 5, 2011/

    Great achievement. Good Job. Proud to be an Aviat Employee.

    Umar Khan May 5, 2011/

    Great achievement. Congrats!! What is the dish size used for main and standby??

    John Churchill May 5, 2011/

    Loved reading your report.
    I have engineered a number of links to offshore platforms.
    The maximum path length was/is 73 km (high – low path) and using hybrid space / frequency diversity. And carefully modeling path geometry over range of k = +.5 to k = -.5 also achieved great results – at times the diversity combiner switch has to cope with very high transitions from A – B path but modeling showed that by carefully choosing frequency and antenna spacing, the reflection fading on one path does not correlate with fading on the other path, hence – providing the diversity combiner switch is functioning – traffic is unaffected.
    Your 193 km link definitely sets the record.

    Pavel Kamburov May 5, 2011/

    Great job,

    I am proud to work with these professionals who made these links.

    Pavel Kamburov

    ashmar May 6, 2011/

    Appreciable achievement! This breaks our record in india 162 Kms link in 7Ghz.

    Rasheed Amuda May 6, 2011/

    My 2006 Tutor has done it. Well done Ivan and co. Reflection analysis is one of the 2 most important considerations for a path like this. I learned that from you.

    Joro Joe Sallau May 6, 2011/

    A very impressive achievement given the 99.9995 availability and over water tx.

    Gerard Casey May 6, 2011/

    Congratulations!!!! I’m interested in knowing what was the diameter size of the antennas employed due to the long distance of the link? Also, was a QPSK signal used as well due to the distance or was it 16 or 128 QAM?

    Again congratulations!

    Freddie (MTN-SA) May 6, 2011/

    Ivan presented this feat to us in person during his last visit to MTN-SA. Congratulations!

    It would be very interesting to get an update on recorded link reliability (i.t.o. SES & Availability) a year after it has been put into operation.

    aviatnetworks May 10, 2011/

    Hi, Tony,

    Thanks for your comment. It was a very challenging but rewarding project. Unfortunately, the pathloss calculations are based on proprietary data, and we cannot release them online. You might get a rough idea of what they are by using a wireless engineering tool such as Starlink, though there was additional tweaking of the numbers by hand. But if you would like a closer discussion regarding the calculations offline, we would be happy to give your request more consideration. Let us know.

    Aviat Networks

    aviatnetworks May 10, 2011/

    Thanks, Umar,

    For this wireless installation, we used four 3-meter antennas for main and standby.

    Aviat Networks

    aviatnetworks May 10, 2011/

    Hello Gerard,

    For this long distance installation over water, we employed QPSK for more robustness and used four 3-meter diameter antennas.

    Let us know if you’d like anymore information.

    Aviat Networks

    Masood Yousfani May 10, 2011/

    Stunning job…great work done. I wish I was part of that planning team, being a microwave planner….

    Deep May 11, 2011/

    Great Achievement.
    Can anyone share the Link Budget?

    hasnat babar May 11, 2011/

    WOW, Great job man. Hats off to you! Can you please provide some more details like polarity, TX power and antenna spacing? It’s only SD? Have you de-correlated the specular reflections? 15 M diversity antenna offset? How did you de-correlate that?

    Hasnat Babar

    aviatnetworks May 12, 2011/

    Thanks, Deep,

    We checked with the customer and they wish to keep the link budget confidential.

    Aviat Networks

    Hassan May 13, 2011/

    Good job!

    I am interested in knowing radio & antenna system parameters used in this link.

    Ahsn Ghumman May 13, 2011/

    Really a great achievement…a new benchmark set in the field of MW.

    Ahsan Ghumman…TXN Microwave Engineer, Pakistan.

    aviatnetworks May 13, 2011/

    Thanks, Hassan,

    The only additional information we are able to reveal is that the antenna diameters were 3 meters and a QPSK modulation was used in the link. If you would like any additional details, perhaps we can talk offline. Let us know.

    Aviat Networks

    Paulo Nunes May 16, 2011/

    Hi all,

    I am fortunate enough to know part of the team behind this great achievement and for them all I leave my congratulations.
    All the best!

    aviatnetworks May 16, 2011/

    Hi Paulo,

    Thanks! Ivan and team are really glad this got your attention! We hope to have something further on this at a later date.

    Stay tuned!

    Aviat Networks May 19, 2011/

    Great…!! Just a Buff..!!

    I’m working with NSN and learned lot from my team and projects I undertook…but anytime, if given a chance, to be a part of these activities is a dream of mine.

    Congratzzz Team…!!

    Many Thanks..!
    Ashish Jain

    Sayan Asanar May 19, 2011/

    Great Achievement! To implement an MW link over water for 193 km and with 99.9995% availability is stunning. Congratulations to all the members of that team.

    We would like to know, the antenna gain and the channel bandwidth used for this link.

    Thanks and regards

    Sayan Asanar

    Amit May 20, 2011/

    I am a transmission maintenance engineer. I need to know that the link will have the same performance all the time or not? I’m interested in knowing the diameter of the antennas employed due to the long distance of the link. Maybe 3.2 meters or more. Also, was QPSK used as well due to the distance or was it 16 or 128 QAM? Why not more capacity?

    Congratulations to the team to design such a link!!!

    Jean May 21, 2011/

    Dear, Ivan,

    Thank you for this interesting information, and could you please indicate what K-factor value was used in the design for the Earth bulge clearance. As well, could we have some more explanation on why the diversity antenna was installed away horizontally and not vertically, as usual, with the main antenna.


    Joseph Mwangi May 22, 2011/

    Ivan & Team,

    This is a great feat keeping in mind that the path is over water and the link’s throughput is 40Mbps! There is a lot for us to learn here; please share with us the link’s performance stats.

    Best Regards,

    aviatnetworks May 24, 2011/

    Hello Amit,

    A microwave communications link’s performance is standardized by American and ITU Error Performance Objectives that predict the time the link will be “off the air” due to standard atmospheric multi-path conditions. Consequently, we expect that this link will meet or exceed the customer’s objective. For the antennas, 3-meter units were used. Due to the geo-climatic conditions and capacity requirements, QPSK was implemented. Why not more capacity, you ask? There’s a possibility that more capacity can be supported, but the customer did not specify this objective at this time.

    Aviat Networks

    aviatnetworks May 24, 2011/

    Hello Sayan,

    Thanks for your positive comments. To answer your questions succinctly, the antenna gain is 43.3-dB and the RF channel bandwidth is 28 MHz.

    Aviat Networks

    M. Jüngst May 27, 2011/

    Fascinating achievement! 20 years ago Richard U. Laine and Ross Lunan from Harris MCD wrote several top value articles on the design techniques for difficult links. How far was this knowledge applied to this link design and which new knowledge—if any—was used? It is worth it to supplement these classic articles with a new part describing this record link design.


    Gautam June 14, 2011/

    What equipment was used? Was it an IDU GE with 20x or an INU-based system and what type of ODU was used.

    aviatnetworks June 24, 2011/

    Hi Hasnat,

    Over water paths should be assigned vertical polarization to reduce the reflective fade depth; for this reason, vertical polarization was used here. The antenna spacing was less than 10-m at each end; with this, the reflections were totally de-correlated. Transmit power was around 1-W; however, this is by no means “relevant” in path design in the majority of cases. What’s important is solid path engineering and all its implications to span microwave radios over any type of geo-climatic conditions and meeting the Error Performance Objective. The horizontal offset was due to the fact that the radar tower was too short to accommodate both dishes.

    Aviat Networks

    aviatnetworks June 24, 2011/

    K=4/3. The horizontal offset was fortuitous because the radar tower was too short.

    aviatnetworks June 24, 2011/


    “Wireless fiber optics performance” is possible with microwave when many experience-rich factors converge. This type of behavior is due to a good assessment of the propagation conditions, solid path design and optimum antenna separation and alignment. These three parameters are of paramount importance to span any geo-climatic conditions and still meet the expected Error Performance Objective according to Vigants or ITU.

    Aviat Networks

    aviatnetworks June 24, 2011/

    Dear M.J.,

    Definitely, the technical descriptions contained in the papers written by the “old” masters are necessary to predict the performance of difficult Ethernet paths. What’s important here, as in any link design, is the knowledge of the propagation conditions to assess how they vary over time to adapt the antenna system to the geo-climatic conditions. The papers written by Dick Laine are of great relevance in the design of these types of challenging links.

    Aviat Networks

    aviatnetworks June 24, 2011/

    ODU 300hp, INU with DAC ES. This is a native Ethernet link, 40 Mbps.

    Kamal Vaughan August 28, 2011/

    Great achievement!!!!

    I also have a query similar to that posted by Hassan but would like additional details as to the radio and antenna equipment used and the initial model used to design the link.

    Thanks Kamal

    Shashi Kant August 29, 2011/

    Thats really a Great News and Record as well
    Hearty Congrats to all team members involved in this historic activity.

    Shashi Kant

    Yasser September 27, 2011/

    Engineering and successfully implementing such an impressive link is one thing. But to share the experience and results with the rest of the community out there is another. Thank you and congratulations!

    Gautam Goel December 13, 2011/

    How is the link performing?

    What is the current performance status of the Ethernet traffic and link performance?

    Gautam Goel

    aviatnetworks December 14, 2011/

    Dear Gautam,

    According to a complimentary screen I received on Dec. 5, no Errored Seconds (ESs) were detected during a period of 4.77 days. Dribbling ESs derive the residual BER (RBER) or quality of a microwave link. No ESs during a measurement period of several days as in this case reflects powerful error correction algorithms working flawlessly supported by robust path engineering and fine antenna alignment.

    Ivan Zambrano
    Senior Network Engineer
    Aviat Networks

    Agus P. Simorangkir June 19, 2017/

    Greaaat achievement…. it will be good solution in Indonesia – country with more than ten thousands Island…
    it will be a honour to know if some day Aviat will come to Indonesia and promote the technology.


    Agus P. Simorangkir
    Linkedin: agus p. simorangkir

    mohamed haris November 9, 2017/

    great installation.

    i had a same kind of situation with limited space and structure availability to establish a link over water for 70KM.

    we used 7 & 6 GHz frequency diversity with packet redundancy to provide stable link with very less packet loss or error.

    193Km link is a big achievement congrats!

    umberto Bini March 7, 2018/

    Hi guys, great achievement. I implemented a 90 Km link over water between Greece and Italy at 4Ghz with 3.7mtrs antennas in Space diversity and 140Mbs capacity. It was very difficult to achieve the performance objectives and the antenna alignment was a very difficult work. However 193 Kms is something else. I have been in Honduras in 2007 and am familiar with propagation issues. Do you have feed back how the link is performing during the rainy season ? Let me know please. Thanks Umberto Bini

Comments are closed.

Subscribe to our newsletter