[k Ib~
RecelVed &\nspected
,.. ')11-''~UG 0.> /..'-"BEFORE THE
FCC Man Room
Federal Communications Commission
Washington, DC 20554
In the matter of
LightSquared Subsidiary LLC File No. SAT-MOD
20101118-00239
Request for Modification of its Authority for an Ancilliary Terrestrial Component
Comments by Charles W. Rhodes
I, Charles Rhodes, was the Chief Scientist of the Advanced Television Test Center Inc.
throughout its lifetime, 1988~ 1996. I was responsible for the technical aspects of the
testing of Advanced Television Systems in the laboratory facilities of the ATTC which I
was also responsible for. In 1995, we tested the Digital TV System devised by a
consortium of firms who has developed proprietary and competitive systems to replace
the analog TV system (NTSC) adopted by the FCC in 1954. Under the leadership of Mr.
Richard Wiley, chairman of the Industry Advisory Committee to the FCC in the matter of
Advanced Television Systems, these firms worked to devise one digital TV system to be
proposed by the industry to the FCC.
In that work, we were vitally concerned with the possibility of interference between the
analog and DTV signals during the transition from analog to digital transmission. During
that transition, DTV signals would transmit on channels not already in use. We were
especially interested in the DTV~DTV interference after the transition.
Since retirement, I have remained interested and active in the progress of DTV during the
transition, and more recently, in the new all-digital world we helped create. I have
published a number of technical papers on the topic ofDTV-DTV interference based
upon my own experimental investigations.
When I learned of the plan for terrestrial broadcasting in the MSS band adjacent to the
space-to-earth segment of the GNSS band, I realized that there was a potential
interference problem similar to that ofmultiple DTV signals on nearby channels that I
had studied and published on.
In my own laboratory, I have conducted a series of simulations of the three phases ofthe
LightSquared scheme using multiple DTV signals. This work was done in May & June
20 II. The results convinced me that there would be jamming of GPS receivers by the
proposed terrestrial transmissions in the MSS Band. It came as no surprise that the
Technical Working Group (TWG) reported similar findings for actual GPS receivers with
the actual LTE Signals being radiated in the MSS Band.
Page 2
Simulations of higher order distortions which produce noise in the GPS
Band
I believe that the interference reported by the TWG for two 5 MHz LTE signals in the
MSS Band was primarily due to third order inter-modulation products generated by that
pair ofLTE signals. The recent proposal by LightSquared to initially radiate only one
LTE signal in the MSS Band (see above) will not result in third order distortion products
which fall within he GPS receiver passband. However, my simulation has shown that
there will be higher order distortion products which will fall within the GPS Band as
shown in my Figure 1 attached.
The GPS system employs Direct Sequence modulation which uniquely permits reception
of faint signals well below the noise generated by the receiver. [ 1 ]. GPS signals are
received at about - 130 dBm. This compares to the minimum usable received DTV signal
(which does not use Spread Spectrum technology such as Direct Sequence modulation)
power of- 84 dBm. This is 46 dB less power required for GPS reception than required
for DTV reception.
The signal power ofthe planned implementation by LightSquared which may enter GPS
receivers near a LightSquared transmitter may be - 30 dBm. The received GPS signal is
at - 130 dBm, a difference of 100 dB.
I have also simulated the single ten MHz wide LTE signal in the MSS Band from 1526
MHz to 1536 MHz to see if it might also cause interference to GPS reception as shown in
Figure 1.
This interfering signal would overload many GPS receiver input transistors and they will
generate higher order distortion products some ofwhich fall in the GPS Band as shown in
Figure 1.
Marker # 1: 1526 MHz (simulated) Marker # 2: 1536 MHz (simulated) and Marker # 4:
1575 (simulated)..
My colleague, Mr. Linley Gumm has simulated this same problem on a Computer and
obtained similar results. This tends to confirm the simulation carried out in my laboratory
with multiple DTV signals.
Simulation Strategy
These simulations used two DTV signals on channels 30 and 31 in the UHF Band. The
bandwidth between - 3 dB points on the TV signal are 5.38 MHz. The two signal's
passband extends from 566.310 MHz to 577.07 MHz (10.76 MHz bandwidth). The
simulation offset between these is 959.69 MHz. The lower edge of the channel 30 DTV
Page 3
signal is 566.31 +959.69 MHz === 1526.00 MHz. The upper edge of the second DTV
signal is 567.07 MHz + 959.69 MHz = 1536 MHz. The GPS band center = 1575.42 MHz
so in my simulation it is 1575.42 MHz - 959.69 MHz =615.73 MHz. The spectrum plot
shows the GPS bandwidth of+/- 12 MHz centered on 1575.42 MHz.
Conclusions that can be drawn from these simulations
What this simulation demonstrates is that the noise floor in the simulated GPS band is
from 8 to 18 dB above the instrumentation noise floor. Ifwe take the average of 13 dB
across the GPS Band, that is the jamming effect of this signal. This 13 dB increase in the
noise floor across the GPS Band is not the only factor that will cause jamming ofGPS
reception. The other cause and it could be the more significant cause is called de
sensitization. This is another odd order distortion product but it applies to all signal
frequencies. The undesired signal received power can exceed that ofthe GPS signal by
100 dB. It is most unlikely that any known transistor will not suffer de-sensitization when
two signals as different in power as this are present. Even with a filter between the
antenna and the RF amplifier transistor, the power difference between the LightSquared
signal and the GPS signals is so great that some de-sensitization will be experienced.
What cannot be concluded from these simulations
The Low Noise Amplifier transistor used in GPS receivers is usually biased to minimize
noise, not to maximize dynamic range. Furthermore, few such devices are specified for
higher than 3rd order inter-modulation performance. As shown above, higher order non
linearities in these transistors can generate higher order distortion products which fall
within the GPS Band resulting in jamming. Only by laboratory testing of GPS receivers
can their robustness to jamming be determined.
Future Increases in Radiated Power in the MSS Band
If LightSquared were allowed to increase their EIRP, whatever de-sensitization there is
would be increased. A 3 dB increase in EIRP would result in a 9 dB increase in de
sensitization because this is a 3rd order distortion product. That would probably end the
usefulness ofmost GPS receivers.
Filters for mitigating jamming
If a filter were available that could be placed at the antenna so as to reduce the Undesired
signal by 40 dB, it would still be 60 stronger than the GPS signal. Even with a 40 dB
filter, I believe that there will still be some de-sensitization ofGPS receivers. The extent
to which de-sensitization and high order inter-modulation products impact the reliability
ofGPS reception has not been determined. This filter must not attenuate the GPS signal
Page 4
by more than say 1 dB because that amounts to a de-sensitization of the GPS receiver of
1 dB. There are filters that can provide the out-of-band rejection (40 dB) but to the best of
my knowledge their insertion loss exceeds 1 dB.
Reducing signal overload by increasing current in the LNA
The dynamic range over which a transistor operates in a linear mode (does not generate
distortion products) can be made larger by biasing that transistor to operate at much
higher currents than are commonly employed in portable GPS receivers and in cellular
telephones. In short, better performance concerning interference is possible, but at a
considerable cost in the time the battery in hand-held receiving devices can operate
before re-charging.
Conclusions
The testing performed so admirably by the TWG was designed to provide Pass / Fail
results. It was not the intent of those tests to determine the mechanism(s) by which the
system failed. It appears to the writer that non-linear distortions in the GPS receiver LNA
are the causes. De-sensitization lowers the GPS signal power, while higher order
intermodulation products increase the noise floor in the GPS Band. The relative
importance of these two causes is unknown. It would be extremely dangerous for the
FCC to proceed with Final Authorization for LightSquared to proceed to build its infra
structure based on what is known about GPS receiver performance with one ten MHz
L~a1W:R~urt~;~g;n;~~i
References:
Dixon, Robert C. "Spread Spectrum Systems with commercial applications"
ISBN 0471-59342-7
Kaplan, Elliot D. "Understanding GPS Principles and Applications"
ISBN 0-89006-793-7
Rohde, Ulrich L. and
Newkirk, David P. "RF/Microwave Circuit Design for Wireless Applications"
Wiley Interscience Series ISBN 0-471-29818-2
Rhodes, Charles W. "Interference Mitigation/or Improved DTVReception"
IEEE Transactions on Consumer Electronics, May 2005
Volume 51, No.2 ISSN 0098-3063 pages 463 onwards
Rhodes, Charles W. "New Challenges to Designers ofDTVReceivers Concerning
Interference" IEEE Transactions on Consumer Electronics,
Volume 53, No.1 ISSN 0098-3063
Page 5
References (continued):
Rhodes, Charles W. "Interference to DTVReception due to Non-linearity ofReceiver
Front-Ends" IEEE Transactions on Consumer Electronics
February, 2008 Volume 54, Number 1 ISSN 0098-3063
Rhodes, Charles W. "Interference Between Television Signals due to
Intermodulation in Receiver Front-Ends"
IEEE Transactions OIl Broadcasting, March 2005
Volume 51 Number 1, ISSN 0018-9316
Bendov, Oded Interference to DDTReception by First Adjacent Channels
IEEE Transaction on Broadcasting, March 2005
Volume 51 Number 1 ISSN 0018-9316
Martin, S.R. "RF Performance ofDTV Converter Boxes- An Overview of
FCC Measurements" IEEE Transactions on Broadcasting"
December 2010 Volume 56 Number 4 ISSN 0018-9316
Log
10
dBI
~ j
v,....
·
;0
GI'8
Band ---l~
/
""".
"I'/",-'
.'
I
vi
rv....'
~
· 1'1"'\_, ~ ~, " ;',
" I :"'" ~
~ -.'v· .. 1.J~1'~~'._ ~
:
?
?
?
? ~
I ?
~ ~
>
?
?
?
?
?
?
...
. " T
,..fv
',,"
1528 1538 1548 1558 1568 1578 1588 1598 1808
MHz MHz MHz MHz MHz MHz MHz MHz MHz
Green: Simulated LTE Blue: 3rd Order Distortion Purple: Higher Orders of
Signal Products Distortion Products
Figure 1. Simulated spectrum of 10MHz LTE signal 1526-1536 MHz