Friday, April 29, 2016

OneWeb: Key Characteristics and Aero Application

The public has its first good glance into what is OneWeb

You can download the Technical Narrative from the following link:

OneWeb Constellation (from OneWeb)

I will make corrections to this paper as they emerge, and am happy for any help.

Given the transient nature of the coverage pattern, at least for mid to low latitude locations, no more than 250 MHz are available at any location. Furthermore, in some locations FS interference issues may prevent downlink full spectrum availability, resulting in even less bandwidth available.

I readily admit I am unsure the extent of issues around equatorial service and am basing my conclusions on a very brief verbal discussion with Brian Holz at Sat2016 (to which I may have misunderstanding). The equatorial zone (rumored to be +/- 15 degrees latitude) is exclusively left to FSS overlay except where there may be a coordinated allowance.

Equatorial Service (+/- 15 deg latitude) GSO or Coordination Agreement

  • 720 satellites (18 orbits, 40 satellites per orbit)
  • 50+ ground stations operation over Ka band
    • Visible ground station to all parts of the orbit (true global coverage)
  • 1200 km polar-inclined orbit
  • 16 elliptical beams (with +/- 10 deg N/S variable pitch)
  • Satellite beam about 700-900 miles across
  • 10.7 - 12.7 GHz service downlink (2 GHz total)
  • 12.75 - 13.25  GHz band not used in US  (see note 1 below)
  • 14 - 14.5 GHz service uplink (500 MHz total)
  • User Terminal - Single polarity transmit (LHCP) and receive (RHCP)
  • Each forward beam operates 250 MHz TDM
    • I REPEAT >>>250 MHz per beam<<<<
  • Each return beam operates 1.5 - 20 MHz TDMA/FDMA
  • 5 GHz usable spectrum yields 2:1 frequency reuse
  • Ka feeder links aggregate equivalent feeder spectrum using dual circular polarities
  • User terminal sized 30 - 70 cm (12 - 28 inches)
    • Dual mechanical apertures or electrically-steered arrays (see note below)
  • Satellite is at least 55 deg above horizon for any user terminal
    • higher minimum elevation at higher latitude
  • The free space loss to a user from LEO is about 30 dB less than from GEO
  • -13.4 dBW/4 kHz downlink PSD
  • -1 G/T at satellite
  • -77 dBW/m2 uplink PFD to saturate Ku transponder
  • Ku and Ka interference managed by suppressing emissions in the direction of FS/FSS assets
  • Ka uplink interference managed by dodging the GSO +/- 6 degrees
    • Ground station diversity
    • 2.4m or larger ground station terminals bigger than other Ka user terminals
  • Progressive Pitch enhances mid to low latitude interference protection to FSS
    • Pitch leads beam towards equator
    • Beam shuts down near equator (rumored to be +/- 15 deg latitude?)
    • Pitch trails beam leaving equator
  • OneWeb promotes first come - first served priority against FSS and NGSO
    • actively coordinating with other operators
    • frequency sharing (band splitting) is a likely outcome
Forward Channel uses dedicated SSPA per User Beam

Return Channel uses SSPA per polarity per Gateway

Link Budget

I took a stab at a link budget.  There is not much to go on, and in particular I have largely ignored the feeder link and focused on the service link.

The downlink PSD is stated to be -13.9 dBW per 4 kHz.

The uplink SFD to saturate the transponder is -77 dBW/m2.  The uplink G/T is -1 dB/K.

I set the forward channel at 225 Msps, with 1.10 spacing that is about 250 MHz.  I set the return channel at 18 Msps, with 1.10 spacing that is about 20 MHz.

Assuming clear sky, 1200 km separation.

The downlink performance is set by the user terminal G/T.  I don't know how to judge the C/I, so I left it at the same value I would for FSS.

Without a spectral mask or limit to HPA, I could drive any aperture uplink to saturate the satellite transponder.  I used a 50W HPA in all cases and a symbol rate to yield the same return C/N for each aperture (and thus the same spectral efficiency).  The 50W HPA matched the 28" aperture nicely in delivering maximum satellite PFD, so I used the maximum symbol rate for this condition to set the C/N reference.

13 beams covers an area about equal to Idaho, Montana, and Wyoming.  328,000 sq. miles into 13 beams is about 25,000 sq. miles per beam.  

Continental US (land areas) is about 2,959,000 sq. miles.  That works into about 118 beams for all of CONUS.  118 beams in the forward direction provide a total of 29,500 MHz for CONUS.  In the return it works out to 14,750 MHz for CONUS.  Assuming spectral efficiency of 2 yields 59 Gbps forward.  Return SE of 1.8 yields about 26 Gbps.  This is homogenous coverage.  Even if two beams overlap, the underlying service is limited to one beam for the times when only one beam is in view.

A single FSS GEO HTS satellite can provide 59 Gbps to CONUS, and with beam hopping can apply considerably more service to a given hotspot.

Scan loss at 55 deg is about 19%.
28" aperture shrinks to 25" equivalent.
18" aperture shrinks to 16" equivalent.
12" aperture shrinks to 11" equivalent.

Beam Pattern

Each satellite has 16 highly-elliptical beams that are about 35-40 miles N-S and 700-900 miles E-W in extent.  

With 18 orbits, there are 36 crossings at the equator.  The equatorial circumference is about 25,000 miles, making each orbit about 700 miles (1100 km) apart at the equator.  The orbits converge with increasing latitude.

The satellites travel N-S about 3 minutes apart.  Half the day the satellites go north and the other half of the day they go south.

The Earth rotates under the orbits, causing each satellite to cover a bit more westerly as it crosses a given point.  There are 36 orbital crossings in one day, or an orbit travels E-W across a fixed location in about 40 minutes. 

Coverage from five OneWeb satellites each with 16 highly elliptical user beams (from OneWeb)

Progressive Pitch

OneWeb Progressive Pitch, Satellite approaching equator (from OneWeb)

Pitching/Tilting the Beam towards the Equator

Equatorial Shutdown (from OneWeb)

Does the equatorial shutdown cause a gap in equatorial service? Can you regain equatorial coverage by the leading/trailing tilted satellites to fill in the equatorial shutdown?   Maybe.

Composite coverage

There is always a challenge when you shift from leading to trailing tilt requiring the satellite to shut down for some interval. 

Swapping from leading tilt to creates an Interference Zone

Issues with using a single mechanically steered antenna

Using a single mechanically steered aeronautical antenna for OneWeb creates challenges and possible limitations.  I am taking some issue with the assertion from OneWeb that buffering can smooth the handoff, that a single antenna can switch pointing directions quickly enough, and that OneWeb does not recognize the keyhole issue.

Excerpt from OneWeb Technical Narrative

I wrote the following white paper exploring these issues a few weeks ago. Some of my assumptions are off based on the newly released narrative, notably I thought they planned on 20 orbits, but now it appears to be only 18; also I was expecting a smaller user terminal than the 12" - 28" range announced. However, the concerns raised are relevant and the geometries expressed reasonably accurate.

OneWeb by design encounters keyhole beam steering conditions frequently.

OneWeb by design will handoff between satellites at least once every three minutes.

Handoff and keyhole are real challenges for a single mechanically-steered aperture.

Stay tuned!

Peter Lemme
peter @

Copyright 2016
All Rights Reserved

No comments:

Post a Comment