Page 1 of 85
ERA TECHNOLOGY
Conducted and Radiated
Measurements for Low Level UWB
Emissions
B S Randhawa
ERA Report 2006-0713 (Issue 2)
ERA Project 7G0232303
FINAL REPORT
Client
:
Client Reference :
Ofcom
Mike Parkins
ERA Report Checked by:
Approved by:
S P Munday
M Ganley
Project Manager
Programme Manager
Ofcom Measurement Resource
Ofcom Measurement Resource
January 2007
Ref. BR/vs/62/02323/Rep-6060
2
ERA Report 2006-0713 (Issue 2)
©
Copyright ERA Technology Limited 2007
All Rights Reserved
No part of this document may be copied or otherwise reproduced without the prior written permission of ERA
Technology Limited. If received electronically, recipient is permitted to make such copies as are necessary to:
view the document on a computer system; comply with a reasonable corporate computer data protection and backup policy and produce one paper copy for personal use.
Distribution list
Client
Project File
(1)
(1)
DOCUMENT CONTROL
If no restrictive markings are shown, the document may be distributed freely in whole, without alteration, subject to
Copyright.
ERA Technology Ltd
Cleeve Road
Leatherhead
Surrey KT22 7SA
UK
Tel : +44 (0) 1372 367000
Fax: +44 (0) 1372 367099
E-mail: info@era.co.uk
Read more about ERA Technology on our Internet page at: http://www.era.co.uk/
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
3
ERA Report 2006-0713 (Issue 2)
Summary
a)
Introduction
The proposed European limits for Ultra Wideband (UWB) emission levels are significantly tighter
than the original Federal Communications Commission (FCC) limits, and it has been suggested that
the measurement techniques mandated by the FCC might not have sufficient sensitivity to measure
mean Effective Isotropic Radiated Power (EIRP) limits as low as –85 dBm/MHz in the 1.6 to
3.8 GHz. ERA were asked to investigate if practical, repeatable and accurate measurements could be
made at this low level using conducted and radiated test measurements. The objectives of the project
were to:
1. Focus on the advantages and disadvantages for the following test facilities and recommend a
measurement environment for cost effective compliance testing.
a. Open Area Test Sites (OATS)
b. Semi Anechoic Room (SAR)
c. Fully Anechoic Room (FAR)
d. GTEM or TEM Cells
e. Reverberation Chamber.
2. Characterise the different types of UWB signals.
3. Assess the minimum EIRP level that could be measured using standard measuring equipment
based on:
•
Environment
•
Measurement distance
•
Antenna type
•
Detector type
•
Resolution bandwidth.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
4
ERA Report 2006-0713 (Issue 2)
b)
Test Environment
The radiated measurements were performed in a Fully Anechoic Room. This measurement
environment was chosen over other environments due to the following reasons:
1. The measurement procedures developed by the National Telecommunications and
Information Administration (NTIA) recommend installing the UWB device in an environment
where measurement system multi-path and external radio signals are eliminated or minimised.
The best possible choice is a high-performance anechoic chamber.
2. Communication papers in Japan describe techniques used to measure emissions from UWB
devices also recommend measuring the emission in a high performance anechoic chamber.
Below 1000 MHz a semi-anechoic chamber may be used.
The OATS test environment was not considered because it has several deficiencies such as weather
and ambient vulnerability, ground plane mutual coupling and reflection effects as well as incomplete
angular coverage of the equipment under test radiation patterns. The ambient environment can be a
problem when measurements have to be performed at low sensitivity levels in the frequency range
where intentional emitters are present, such as GSM 900 & 1800 or 3G. The RF ambiance of the
OATS can add considerable uncertainty to the measurements.
Calculations have shown that the GTEM cell may have a 2 dB greater sensitivity compared with a
FAR (See Appendix A). However, the available volume is small and not entirely suitable for
equipment with attached cables. Experiments have shown that the results in the GTEM cell were
sensitive to location of the EUT within a few centimetres, giving variations of 2 dB up to 2.5 GHz and
7 dB above 2.5 GHz [22]. Also, the cross-polarisation performance is considered inferior to an
anechoic chamber or OATS. Over a limited frequency band the field level of the longitudinal mode
can exceed the level of the intended vertical field. The size of EUT is limited to approximately onethird height between the septum and floor. Finally, it is difficult to determine the measurement
uncertainty because of the cross-polarisation performance being inferior to an anechoic chamber or
OATS.
The ERA Stage B Report [25] on “Emission measurements in the range 1 to 6 GHz in Fully Anechoic
Room and Reverberation Chamber Facilities” concludes that there is a reasonable correlation between
the FAR and the reverberation chamber results, further work is required to investigate the appropriate
gain factor as suggested in BS EN 61000-4-21. There are also good correlations between the
reverberation mode- stirred and mode-tuned measurements.
On average the reverberation chamber gave a reduction in field strength of 7 dB to 10 dB +/- 5 dB,
but there was clear and apparently uniform frequency dependence. This could lead to a factor, which
could be included in EN 61000-4-21 to obtain representative fields when using the reverberation
chamber, but further work is required to determine the value and precision of the gain figure to be
applied.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
5
ERA Report 2006-0713 (Issue 2)
The measurement results in a 1 MHz RBW using a peak detector also showed that the reverberation
chamber had approximately 15 dB to 20 dB better sensitivity compared with the FAR. However,
EN 61000-4-21, states that the main disadvantage of the reverberation chamber is that the
measurement system must have a sensitivity of 20 dB lower than the actual mean to get an accurate
average measurement and intermittent signals may be artificially lowered due to insufficient
sampling. Thus, effectively eliminating any sensitivity advantage gained.
Other work has shown the EUT in the reverberation chamber to give a very good comparison with the
FAR data, but 4.5 dB higher. This may be expected, because the reverberation chamber captures the
total energy, whereas the measurements in the FAR measure emissions in a single azimuth plane [22].
Also, reverberation chambers, although referenced in a basic standard are currently accepted in only a
few product standards. Utilisation of reverberation chambers for emissions testing will require
acceptance of total radiated power as a pass/fail criterion. There is also a lowest usable frequency
defined by the cavity dimensions, tuner effectiveness, and cavity quality factor.
Directivity and polarisation effects are not measurable in a reverberation chamber. The high Quality
Factor in reverberation chambers may impose constraints on pulse testing. Depending on test
conditions, correlation of test results to other test techniques may be difficult or impractical.
c)
UWB Signal Characterisation
Conducted measurements have shown that the peak power of a pulsed UWB varies as 20.log10
(Resolution Bandwidth) (RBW) and the Root Mean Square (RMS) power varies as 10.log10 (RBW)
for RBW > PRF.
Also, the results show that the peak and RMS powers vary with a trend of 10.log10 (RBW) for
dithered UWB signals, which is consistent with NTIA Report 01-383 predictions.
Multi-band (MB)-OFDM UWB signals show a similar trend for the measured peak power when
compared to the RMS levels. The results reveal that the peak levels are 4 to 6 dB higher compared
with the RMS results. Both sets of results match well when compared with the Additive White
Gaussian Noise (AWGN) results, which verifies the 10.log10 (RBW) trend. Thus, proving that the
MB-OFDM UWB signal has noise-like characteristics.
d)
Practical Measurement Limits of UWB Signals
Conducted and radiated measurements for a MB-OFDM UWB device were carried out in a FAR
using the test set up described in Sections 11.2 and 11.3. The table below summarises the practical
minimum mean EIRP levels that can be measured.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
6
ERA Report 2006-0713 (Issue 2)
Table 1:
Minimum mean MB-OFDM UWB EIRP levels that can be practically measured
Measurement Method
Minimum Mean EIRP
(dBm/MHz)
Conducted
-861
Conducted with a pre-amplifier
-1011
Radiated at a 1 m separation distance
-501
Radiated with a pre-amplifier at a 1 m separation distance
-74-691,2
1
UWB signal 6 dB above the noise floor of the analyser when measured in a FAR with a DRWG
horn antenna. 2 Between the 1.6 and 3.8 GHz band.
From the table above, it can bee seen that an UWB EIRP limit of –85 dBm/MHz can be achieved
using the conducted measurement set-up, by including a pre-amplifier. The standard conducted
measurement set-up falls within 1 dB of the minimum EIRP limit. However, he radiated test set-up
using a pre-amplifier falls short by 16 dB for frequencies at the top end of the 1.6 to 3.8 GHz band, as
proposed by the ECC. At low measurement frequencies at 1.6 GHz, the mean radiated UWB EIRP
density that can be measured is –74 dBm/MHz.
The conducted measurement set-up using the pre-amplifier is only possible if the antenna of the UWB
device is external and not integrated on the board of the UWB device. Attempting to measure the
UWB signal by connecting an external cable to the antenna port on the board of the device will
produce impedance mismatches and give incorrect results.
The radiated measurements are limited by a combination of the noise floor level of the spectrum
analyser or any other receiver, the noise figure of the pre-amplifier and the sensitivity of the
measurement antenna. In effect, the lower the measurement system sensitivity, the lower the UWB
signal level that can be measured. Smaller bandwidths could be used and the MB-OFDM UWB signal
could be scaled using a 10.log10 (Bandwidth Correction Factor). However, the noise floor of the
spectrum analyser would also go up by the same margin, thus giving no advantage.
The single octave horn antenna has a 6-7 dB better performance in terms of antenna gain compared
with the Double Ridged Waveguide (DRWG) horn antenna. However, the maximum dimension of
35 cm for the single octave horn antenna, compared with 25 cm with the DRWG horn requires the
single octave horn antenna to be roughly twice the distance from the DUT. Thus, incurring an extra
6 dB of path loss compared with using a DWRG. The standard radiated test set-up alone is
insufficient to measure UWB signals with low mean EIRP densities of –85 dBm/MHz.
The Autonomous Interference Monitoring System (AIMS) currently being developed by Mass
Consultants Ltd. for Ofcom, has shown that it is quite possible to measure signals 2 to 3 dB above its
monitoring system sensitivity. The emphasis in that project is on the measurement of the C/(I+N) ratio
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
7
ERA Report 2006-0713 (Issue 2)
and detecting the presence of UWB in the presence of other signals. This is a different problem to
that of measuring out-of-band UWB power, but the two issues are not unrelated. Development of the
necessary algorithms is currently ongoing and it is expected that the system will be completed in
March 2007.”
The monitoring system uses an R&S FSQ26 spectrum analyser with an external wideband preamplifier, giving an overall system noise figure, at the input to Antenna Interface Unit, of between
2.3 dB and 7 dB producing system sensitivities between -110 dBm and -107 dBm in the 0.2 to 8 GHz
frequency band. Theses sensitivities are comparable to those measured in the conducted test set-up as
shown in Table 13, also using a pre-amplifier. The difference in measuring the UWB signal only 2 to
3 dB above the noise floor of the analyser using MASS Consultants monitoring system gives a 3 to
4 dB improvement compared with measuring the signal 6 dB above the noise floor. This improvement
would also apply for radiated measurements as the sensitivity of the analyser with the pre-amplifier
would be the same as in the conducted test set-up.
Using an average detector would improve the noise floor of the spectrum analyser further by 1-2 dB,
but the UWB signal with its noise-like characteristics
Comparison of the R&S FSU spectrum analyser with the R&S ESPI test receiver gave similar noise
floor levels within 1 dB, also giving no advantage in term of measurement sensitivity.
Higher gain antennas such as parabolic dish antennas with a typical gain of 25 – 30 dBi could be used.
However, at frequencies between 1 to 5 GHz, the maximum dimension of the measurement antenna
(around 1 m) would increase the separation distance to 20 m at a frequency of 3 GHz. The additional
path loss of 26 dB would cancel out any effective gain over using a DRWG antenna or single octave
horn antenna.
Measurements in a GTEM may improve the sensitivity and radiated EIRP by 2 dB based on
calculations shown in Appendix A. The reverberation chamber may produce a 4 dB to 5 dB
improvement, based on a better sensitivity performance of 15 dB, but an average reduction in field
strength of 10 dB compared with the FAR [25].
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
8
ERA Report 2006-0713 (Issue 2)
This page is intentionally left blank
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
9
ERA Report 2006-0713 (Issue 2)
Contents
Page No.
1
Introduction
17
2
EMC Standards Review
18
2.1
Current Emissions Standards at Frequencies above 1 GHz
19
2.1.1
IEC/CISPR
19
2.1.2
FCC
20
2.1.3
ETSI
24
2.1.4
Summary of current emission standards
24
3
Current UWB Emission Levels
25
3.1
FCC and NTIA Recommendations
25
3.2
ECC Recommendation
27
4
UWB Applications
30
5
UWB Signal Characteristics
31
6
UWB Technology
34
6.1
Introduction
34
6.2
DS-UWB
36
6.3
Multi-Band OFDM
38
7
Measurement Techniques
40
7.1
Full Bandwidth Time Domain Measurements
40
7.2
Bandwidth Limited Measurements
42
7.2.1
42
Spectrum envelope measurements
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
10
ERA Report 2006-0713 (Issue 2)
8
9
7.2.2
Peak power measurements
44
7.2.3
FCC Part 15 average power measurements
44
7.2.4
RMS power measurements
44
7.2.5
APD measurements
45
Measurement Environments
46
8.1
Open Area Test Site
46
8.2
Anechoic Rooms
47
8.3
GTEM Cells
47
8.4
Reverberation Chamber
49
Measurement Considerations
51
9.1
Test Environment
51
9.2
Detectors
52
9.3
Sensitivity of Equipment
52
9.4
Radiation Pattern
53
9.5
Far Field Conditions
53
10
Test Methodology
54
11
Measurement Results
55
11.1
Characterising UWB Signals
55
11.1.1 Pulsed UWB
55
11.1.2 Dithered UWB
56
11.1.3 Simulated OFDM-UWB
57
11.1.4 MB-OFDM
60
Conducted Low Level UWB Emissions
62
11.2
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
11
ERA Report 2006-0713 (Issue 2)
11.3
12
13
Radiated Low Level UWB Emissions
66
Conclusions
74
12.1
Introduction
74
12.2
Test Environment
74
12.3
UWB Signal Characterisation
76
12.4
Practical Measurement Limits of UWB Signals
76
References
APPENDIX A Calculation of Minimum Measurable UWB EIRP
78
81
A.1 Minimum Measurable EIRP using a FAR
82
A.2 Minimum Measurable EIRP using a GTEM Cell
83
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
12
ERA Report 2006-0713 (Issue 2)
Tables List
Page No.
Table 1: Minimum mean MB-OFDM UWB EIRP levels that can be practically measured .................. 6
Table 2: Summary of the emissions mask for UWB proposed by Vodafone ...................................... 17
Table 3: CISPR 22 amendment, proposed limits above 1 GHz ............................................................ 20
Table 4: FCC limits for digital devices................................................................................................ 21
Table 5: Equivalent FCC limits for UWB ............................................................................................ 22
Table 6: Comparison of CISPR and FCC limits at 3 m for frequencies above 1 GHz ........................ 24
Table 7: Maximum EIRP limits for devices using UWB technology in bands below 10.6 GHz ......... 28
Table 8: Radio Systems operating between 1 and 3 GHz.................................................................... 30
Table 9: Possible set of DS-UWB signal parameters........................................................................... 38
Table 10: Simulated OFDM-UWB signal parameters .......................................................................... 58
Table 11: Wisair MB-OFDM UWB signal parameters ........................................................................ 60
Table 12: Conducted measurements of MB-OFDM UWB signals close to the noise floor ................ 62
Table 13: Conducted measurements of an amplified MB-OFDM UWB signal .................................. 64
Table 14: Radiated measurements of a MB-OFDM UWB signal without a pre-amplifier ................. 68
Table 15: Minimum mean UWB EIRP levels that can be measured below 3.8 GHz in a 1 MHz RBW
using a HP 8449B pre-amplifier ................................................................................................... 70
Table 16: Radiated measurements of a MB-OFDM UWB signal using an ERA pre-amplifier .......... 71
Table 17: Radiated measurements of a MB-OFDM UWB signal using a DRWG and single octave
horn antenna.................................................................................................................................. 71
Table 18: Comparison of S/N ratio for pre-amplifier configurations .................................................. 72
Table 19: Minimum mean MB-OFDM UWB EIRP levels that can be practically measured ............. 73
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
13
ERA Report 2006-0713 (Issue 2)
Figures List
Page No.
Figure 1: FCC Digital device and UWB field strength limits at 3m using a RBW of 1 MHz.............. 23
Figure 2: Indoor UWB emission masks defined by the FCC................................................................ 26
Figure 3: Outdoor UWB emission masks defined by the FCC............................................................. 27
Figure 4: FCC and draft ECC UWB mean EIRP limits compared with CISPR average emission limits
in a 1 MHz band............................................................................................................................ 29
Figure 5: UWB pulse spacing modes.................................................................................................... 32
Figure 6: Spectral characteristics of different pulse spacing modes ..................................................... 33
Figure 7: IEEE 802 organisation........................................................................................................... 35
Figure 8: DS-UWB wavelet and its spectrum in 6-10.6 GHz range ..................................................... 37
Figure 9: Single channel band occupancy of an OFDM UWB system................................................. 39
Figure 10: Expanded view of a multi-band OFDM UWB channel....................................................... 39
Figure 11: Full bandwidth measurement system set-up........................................................................ 41
Figure 12: Bandwidth limited measurement system set-up .................................................................. 42
Figure 13: Typical APD curve .............................................................................................................. 45
Figure 14: APD graph of AWGN produced by a R&S FSU46 spectrum analyser in a 1 MHz
bandwidth...................................................................................................................................... 46
Figure 15: Diagram of the set-up to measure a pulsed UWB signal..................................................... 55
Figure 16: Measured power as a function RBW for a pulsed UWB device with 5 MHz PRF ............. 55
Figure 17: Diagram of the set-up of a dithered UWB signal ................................................................ 56
Figure 18: Measured power as a function RBW for a dithered UWB device with 5 MHz PRF .......... 57
Figure 19: Bandwidth limited measurements for a simulated OFDM-UWB signal using a spectrum
analyser ......................................................................................................................................... 58
Figure 20: Measured RMS power as a function RBW for a simulated OFDM-UWB signal............... 59
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
14
ERA Report 2006-0713 (Issue 2)
Figure 21: Measured peak power as a function RBW for a simulated OFDM-UWB signal................ 59
Figure 22: Bandwidth limited measurements for a MB-OFDM UWB signal using a spectrum analyser
...................................................................................................................................................... 60
Figure 23: Measured RMS power as a function RBW for a MB-OFDM UWB signal ........................ 61
Figure 24: Measured peak power as a function RBW for a MB-OFDM UWB signal ......................... 61
Figure 25: MB-OFDM UWB signal and its 3 bands conductively measured in a 1 MHz RBW ......... 63
Figure 26: Bandwidth limited measurements for amplified MB-OFDM UWB signals using a spectrum
analyser ......................................................................................................................................... 64
Figure 27: Amplified MB-OFDM UWB signal and its 3 bands measured conductively in 1 MHz
RBW using RMS detection........................................................................................................... 65
Figure 28: Set-up for radiating UWB emission measurements ............................................................ 66
Figure 29: Radiated MB-OFDM UWB signal measured at a distance of 1 m in a 1 MHz RBW ........ 67
Figure 30: Radiated MB-OFDM UWB signal measured at a distance of 1 m in a 1 MHz RBW using a
HP pre-amplifier ........................................................................................................................... 69
Figure 31: Radiated MB-OFDM UWB signal measured at a distance of 1 m in a 1 MHz RBW using
an ERA pre-amplifier.................................................................................................................... 70
Figure 32: GTEM cell equivalent electric circuit ................................................................................. 83
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
15
ERA Report 2006-0713 (Issue 2)
Abbreviations List
ACF
Antenna Correction Factor
APD
Amplitude Probability Distribution
BW
Band Width
BWCF
Band Width Correction Factor
CEPT
Conference of Postal and Telecommunication administrations
CISPR
International Special Committee on Radio Interference
DRWG
Double Ridged Waveguide Horn
DS
Direct Sequence
ECC
European Communications Committee
EIRP
Effective Isotropic Radiated Power
ETSI
European Telecommunications Standards Institute
FAR
Fully Anechoic Room
FCC
Federal Communications Commission
FWHM
Full Width Half Maximum
IEC
International Electro-technical Commission
ITE
Information Technology Equipment
OFDM
Orthogonal Frequency Division Multiplex
MSSI
Multi Spectral Solution Inc.
NTIA
National Telecommunications and Information Administration
PRF
Pulse Repetition Frequency
RBW
Resolution Band Width
RF
Radio Frequency
RMS
Root Mean Square
SAR
Semi Anechoic Room
UWB
Ultra Wide Band
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
16
ERA Report 2006-0713 (Issue 2)
This page is intentionally left blank
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
17
ERA Report 2006-0713 (Issue 2)
1
Introduction
The proposed European limits for Ultra Wide Band (UWB) emission levels are significantly tighter
than the original Federal Communications Commission (FCC) limits, and it has been suggested that
the measurement techniques mandated by the FCC might not have sufficient sensitivity to measure
mean Effective Isotropic Radiated Power (EIRP) limits as low as –85 dBm/MHz between 1.6 GHz
and 3.8 GHz frequency band.
Vodafone responded in March 2005 to OFCOM’s consultation [1] on a position to adopt in Europe on
ultra wideband devices in 3.1-10.6 GHz [2]. In this response, Vodafone have proposed the emissions
mask for UWB equipment, which is summarised Table 2.
Table 2:
Summary of the emissions mask for UWB proposed by Vodafone
Frequency Range
Below 960 MHz
Proposal
Based on CISPR 16 quasi-peak detector with a bandwidth of 120 kHz, as
described in FCC Report and Order 02-48 (see note below)
960 MHz 1.6 GHz
-75 dBm/MHz, as in FCC rules (see note below)
1.6 - 2.7 GHz
-85 dBm/MHz
2.7 - 3.1 GHz
Linear slope (in dB) from –85 dBm/MHz to –51.3 dBm/MHz
3.1 – 5 GHz
Based on results of ECC TG3 impact analysis
5 – 6 GHz
6 - 10.6 GHz
Based on results of ECC TG3 impact analysis, taking into account that this
band is not planned for use by UWB PAN applications.
Based on results of ECC TG3 impact analysis.
10.6 - 11.6 GHZ
Linear slope (in dB), starting from –51.3 dBm/MHz
Above 11.6 GHz
The limitations of measurement techniques should be taken into account in
determining an appropriate value.
Note: These requirements below 1.6 GHz are only adequate if there is also a requirement that the
necessary bandwidth of any UWB emission is entirely within the frequency range 3.1 – 10.6 GHz.
Vodafone believes that the out-of-band mask should fall to a mean EIRP density of –85 dBm/MHz at
2.7 GHz. However, it is important that the practical limitation proposed by Vodafone, is capable of
being measured.
In Vodafone’s response to Ofcom’s consultation on UWB, the mobile operator believes that suitable
measurement techniques do exist, but they may need to be validated before they can be specified in
the European regulatory framework. In order that this issue does not delay the finalisation of the
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
18
ERA Report 2006-0713 (Issue 2)
European framework (and especially Harmonised Standards), ERA proposes to study methods of
measurement for UWB emissions with a mean EIRP density of –85 dBm/MHz.
Vodafone believes that such a study could make a substantial contribution to enabling UWB products
to be placed on the EU market with emission limits that are tighter than the FCC limits.
The objective of this report is to:
1. Review current measurement procedures for UWB emissions in the frequency and in the time
domain.
2. Investigate, if practical, repeatable and accurate measurements can be made at this low level
for the following test environments:
a. Open Area Test Sites (OATS)
b. Semi Anechoic Room (SAR)
c. Fully Anechoic Room (FAR)
d. GTEM or TEM Cells
e. Reverberation Chamber.
3. Focus on the advantages and disadvantages of above test facilities and recommend a
measurement environment for cost effective compliance testing.
4. Characterise various UWB signals based on pulsed, dithered and OFDM technology.
5. Devise a measurement procedure that is capable of achieving a sensitivity of –85 dBm/MHz
for both conducted and radiated UWB field measurements up to a frequency of 5 GHz.
2
EMC Standards Review
The ever increasing use of radio communication services operating at frequencies above 1 GHz has
resulted in a perceived need to develop Electromagnetic Compatibility (EMC) emissions test methods
and limits to protect services such as GSM 1800, DECT, 3G and Blue-tooth.
It is recognised by standards bodies such as International Special Committee on Radio Interference
(CISPR) that the driver for emission limits at frequencies above 1 GHz comes from the cellular radio
operators who have invested heavily in the licenses to operate. Although there have been few
recorded interference complaints, the operators need to ensure that there is adequate protection of
their radio services, particularly in the light that the clock speeds in PCs and multimedia equipment
are rapidly increasing and although there may be few problems at present, that may not be the case in
the future if there are no controls.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
19
ERA Report 2006-0713 (Issue 2)
The main downside to the controls is the additional costs faced by the manufacturers in performing
extra tests, and designing products for compliance. However, the current requirements for testing are
easily extended to 6 GHz. Most EMC test laboratories already having the capability and design
methods for compliance with emission limits are well known and would need only minor adjustment
in most cases.
It should also be noted that emission limits are long overdue. In comparison, Radio Frequency (RF)
field immunity requirements at frequencies above 1 GHz have been included in the basic standard
EN 61000-4-3 for a number of years.
The key issues for the provision of emission limits at higher frequencies than considered for some
time relate to the product characteristics and the services to be protected. Electronic equipment is
continuing to operate at faster clock speeds with PCs now operating at speeds exceeding 1 GHz and it
is envisaged that clock speeds may continue to increase up to the 10 to 20 GHz range. The
development of UWB covers a large section of the frequency range above 1 GHz, and limits are
required to protect radio services including GSM 1800 and 3G.
2.1
Current Emissions Standards at Frequencies above 1 GHz
2.1.1 IEC/CISPR
Over the last ten years there have been developments in CISPR to produce test methods and limits for
emissions from Information Technology Equipment (ITE), and only in 2005 the Final Draft
International Standard (FDIS) was approved so that CISPR 22 could be amended. Over this period
there has been a re-organization in CISPR and there are now three sub-committees dealing with the
subject, CISPR A for the basic test methods, CISPR H for the limits and CISPR I for product standard
considerations, CISPR I now covering both ITE, TV and multi-media products.
The proposed amendment to CISPR 22, CISPR/I/151/FDIS, was recently approved and presents the
limits and test methods for emission tests above 1 GHz. The proposed radiated emission limits are
presented in Table 3.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
20
ERA Report 2006-0713 (Issue 2)
Table 3:
CISPR 22 amendment, proposed limits above 1 GHz
Frequency range
(GHz)
Average limit (dBuV/m)
Peak limit (dBuV/m)
Class A ITE
Class B ITE
Class A ITE
Class B ITE
1-3
56
50
76
70
3-6
60
54
80
74
Note: Class A ITE are for commercial and industrial use, Class B is for residential use
The instrumentation and site requirements are referenced in CISPR 16-1-1, 16-1-1-4, and 16-1-1-4.
These requirements may be refined in the future when the current work in CISPR SC A has been
completed.
A 1 MHz resolution bandwidth is proposed for all measurements above 1 GHz, with peak and average
detectors as defined in CISPR 16-1-1 clause 8.2. The preferred test distance is 3 m provided the far
field conditions are achieved, i.e., d > 2D2/ λ, where D is the maximum aperture dimension of the
measuring antenna and λ is the wavelength of the measurement. Measurements may be made at 1 m
or 10 m and the data would be adjusted to the 3m values using free space propagation conditions.
2.1.2 FCC
In the USA, there have been requirements for emissions measurements at frequencies above 1 GHz
for many years. These are set out in the Code of Federal Regulations, 47 Telecommunications, with
limits for digital devices included in Part 15, Sec. 15.109. The recommended test methods and
procedures are according to ANSI C63.4 [3]. There are also requirements for UWB systems at
frequencies above 1 GHz as presented in Part 15, Sec. 15.517, Sec. 15.519 and Sec. 15.521.
2.1.2.1 Unintentional Radiators
The Federal Communications Commission (FCC) has proposed limits for frequencies above 1 GHz,
which are stated in Part 15, Sec. 15.109 and presented below in Table 4. This covers equipment
categories of PC, computers and other digital devices.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
21
ERA Report 2006-0713 (Issue 2)
Table 4:
FCC limits for digital devices
Frequency
range, GHz
Field strength limit (dBμV/m)
Class A (at 10m)
Above 0.96
Class B (at 3m)
Peak
RMS
Average
Peak
RMS
Average
N/a
49.5
N/a
54
For Class B equipment, e.g. residential PCs, the field strength limit is 7 dB more relaxed than the
37 dBμV/m limit of EN 55022 (at 10 m), at frequencies below 1 GHz.
Measurements are performed on an OATS that satisfy the requirements of ANSI C63.7 [4] at
frequencies below 1 GHz, or can be performed in an absorber lined shielded enclosure provided, it
can be demonstrated that reflections do not adversely affect accuracy.
The measurement antenna can be a linearly polarized device with a similar beam-width to the
antennas used at frequencies below 1 GHz. ANSI C63.4 states that at frequencies above 1 GHz, the
measuring antenna may not have a wide beam-width and there is a requirement to move the antenna
vertically and horizontally when observing one face of the EUT in order to ensure that the maximum
radiation emission levels are captured in the search.
The measurements are made with a receiver or spectrum analyser having a bandwidth of 1 MHz and
an average detector (as specified in ANSI C63.2: 1996 [5]).
At frequencies above 1000 MHz, the radiated limits given are based upon the use of measurement
instrumentation employing an average detector function. Where average radiated emission
measurements are specified in this part, there also is a limit on the radio frequency emissions, as
measured using instrumentation with a peak detector function, corresponding to 20 dB above the
maximum permitted average limit for the frequency being investigated. This applies unless a
different peak emission limit is specified in the rules, e.g., Sec. 15.255, operation within the band 57 –
64 GHz; Sec.15.509, (Technical requirements for ground penetrating radars and wall imaging
systems), and Sec 15.511, (Technical requirements for surveillance systems). The average detector
used has a long time constant and significantly reduces the measured levels for any short duration
transients.
2.1.2.2 Intentional radiators
Part 15, Sec. 15.209 sets out limits for intentional radiators such as cordless phones and spread
spectrum devices and the proposed limits are identical to the Class B limits for unintentional radiators,
i.e. digital devices (See Table 4).
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
22
ERA Report 2006-0713 (Issue 2)
2.1.2.3 Personal radio services
The requirements for emissions from radio transmitters are complex but as an example, Sec. 95.635 of
Part 95, Personal Radio Services, dealing with unwanted radiation from radio systems, also applies
the same Class B limits at frequencies above 1 GHz for equipment operating in the MIC band.
2.1.2.4 UWB
Radiated emission requirements for UWB systems are set out in Code of Federal Regulations, 47
Telecommunications, Part 15, Sec. 15.517, 15.519 and 15.521 [6]. The limits are stated as EIRP (in
dBm), and have been converted to field strengths at 3m using the approved conversion, i.e.
E(dBμV/m) = EIRP(dBm/MHz) + 95.26 dB and are presented in Table 5 for both indoor, and hand
held UWB systems.
Table 5:
Equivalent FCC limits for UWB
Frequency
range, GHz
Field strength limits at 3m (dBμV/m)
Indoor UWB
1
Hand-held UWB
Peak
Average
Peak1
Average
0.96-1.61
40
20
40
20
1.61-1.99
62
42
52
32
1.99-3.1
64
44
54
34
3.1-10.6
74
54
74
54
>10.6
64
44
54
34
1
Section15.35 of the FCC rules states that when average radiated emission
measurements are specified in the regulations, the radio frequency emissions,
measured using instrumentation with a peak detector function, can be no more
than 20 dB above the maximum permitted average limit.
The measurements are made as for digital devices with a resolution bandwidth of 1 MHz, and an
average RMS detector.
The plots have been limited to an upper frequency of 18 GHz to represent the requirements for UWB
systems at frequencies above 10.6 GHz. The upper limit on frequency for digital devices and similar
equipment is related to the fifth harmonic of the clock frequency in a computer device, or 40 GHz,
whichever is lower.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
23
ERA Report 2006-0713 (Issue 2)
The FCC limits at frequencies above 1 GHz for the main categories of equipment are presented in
Figure 1 below.
70
Field strength (dBuV/m)
60
50
Class A Digital Device
40
Class B Digital Device
Indoor UWB Device
30
Hand Held UWB Device
20
10
0
0
2
4
6
8
10 12 14 16 18 20
Frequency (GHz)
Figure 1: FCC Digital device and UWB field strength limits at 3m using a RBW of 1 MHz
It is interesting to note the emissions from Class B (residential) digital devices set by the FCC are
identical to the UWB limits in the 3.1 to 10.6 GHz frequency band also set by the regulatory body.
Between 1.99 GHz and 3.1 GHz the field strength at 3 m drops to 34 dBμV/m giving an EIRP of
–61.3 dBm/MHz, 23.7 dB greater than that proposed in Table 2, for handheld UWB devices.
2.1.2.5 ISM
Part 18 of the CFR sets out the emissions limits for Industrial, Scientific and Medical (ISM)
equipment but there are no requirements for measurements at frequencies above 1 GHz.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
24
ERA Report 2006-0713 (Issue 2)
2.1.3 ETSI
Many standards published by the European Telecommunications Standards Institute (ETSI) contain
requirements for the control of radiated emissions at frequencies above 1 GHz. For example, EN 302
066-1: 2005 [7] is for ground probing radar and has ERP power limits for radiated undesired
emissions of –30 dBm in the frequency range 1 GHz to 18 GHz. For Digitally Enhanced Cordless
Telephones (DECT), EN 300 172-2 also has a spurious emission level of 1 μW above 1 GHz [8].
2.1.4 Summary of current emission standards
The test methods stated in the main standards, FCC and CISPR 22 are very similar, requiring E-field
measurements on an OATS, or in an absorber lined room meeting Normalized Site attenuation (NSA)
requirements, i.e. a properly constructed Semi-anechoic Room. Generally at frequencies above
1 GHz a 3 m distance is acceptable and standard measuring instrumentation such as spectrum
analysers are acceptable.
Table 5 presents a comparison of the average detector emission limits at a distance of 3 m for
frequencies above 1 GHz as proposed by CISPR and the FCC respectively.
Table 6:
Comparison of CISPR and FCC limits at 3 m for frequencies above 1 GHz
Frequency range
CISPR Average limit (dBuV/m)
FCC Average limit (dBuV/m)
Class A ITE
Class B ITE
Equivalent Class
A ITE
Class B ITE
1-3 GHz
56
50
60
54
3-6 GHz
60
54
60
54
At frequencies above 3 GHz the limits are identical and below 3 GHz the CISPR limits are 4 dB more
severe, compared to the FCC limits. It is likely that these limits, (if fully agreed) will be then retained
until there is strong pressure in the form of feedback from the manufacturers, or the operators. Any
proposals would take several years to be implemented.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
25
ERA Report 2006-0713 (Issue 2)
3
Current UWB Emission Levels
3.1
FCC and NTIA Recommendations
In the United States, the FCC, in coordination with the National Telecommunications and Information
Administration (NTIA), has developed rules for unlicensed UWB devices under Title 47 of the Code
of Federal Regulations, Part 15 [6]. The regulation sets out guidelines under which an intentional,
unintentional, or incidental radiator may be operated without an individual license. Part 15 stipulates
that unlicensed devices are subject to the condition that no harmful interference is caused to licensed
services and that harmful interference to unlicensed devices must be accepted. It is recognised and
stated in Part 15.5 (c) that; “the limits specified in this part will not prevent harmful interference in all
circumstances.”
The FCC regulation sets out a “mask” with an upper limit of –41.3 dBm/MHz as the amount of power
that can be radiated at any particular frequency by a UWB device within the core band of 3.1 –
10.6 GHz. The amount of power that can be radiated outside of the core band by a UWB device rolls
off to –61.3 dBm/MHz at 2.1 GHz and 11.6 GHz [6].
The FCC has defined a UWB device to be any intentional radiator of RF energy, which has a 10 dB
bandwidth of 25% of the strongest frequency within that 10 dB bandwidth or a 10 dB bandwidth of
equal to or greater to 500 MHz [6]. These devices do not conform to the usual frequency allocation
table and associated FCC regulations, because of these extremely large bandwidths.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
26
ERA Report 2006-0713 (Issue 2)
Figure 2: Indoor UWB emission masks defined by the FCC
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
27
ERA Report 2006-0713 (Issue 2)
Figure 3: Outdoor UWB emission masks defined by the FCC
Figure 2 and 3 show the mask of the average radiated power from an approved unlicensed commercial
UWB for general indoor and hand outdoor UWB devices respectively. Unlicensed UWB devices with
radiation frequencies below 960 MHz are regulated under a different section of Part 15 of the FCC
rules. In this frequency region, the emissions are measured with a receiver using a quasi-peak
detector; therefore, the limit for emissions at frequencies below 960 MHz is in μV/m. This mask also
assumes an isotropic receive antenna.
3.2
ECC Recommendation
Decision ECC/DEC/(06)04 [9] on the harmonised conditions for devices using UWB technology in
bands below 10.6 GHz has been finally adopted by the European Communications Committee (ECC)
at its meeting 20 – 24 March 2006 in Oulu (Finland).
The baseline solution for UWB operation in Europe is twofold and is based on a consensus reached
between Conference of Postal and Telecommunication administrations (CEPT) administrations:
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
28
ERA Report 2006-0713 (Issue 2)
•
Operation in band 6 – 8.5 GHz, subject to maximum power density levels and without
requirement for additional mitigation.
•
Operation in band 3.1 – 4.8 GHz, subject to maximum power density levels and to the
implementation of adequate mitigation techniques, which requirements are to be defined and
effectiveness is to be validated.
No consensus has been reached on the inclusion of regulatory provisions for a phased approach in
the band 4.2 – 4.8 GHz without mitigation techniques.
Draft ECC Decision ECC/DEC/(06)EE has been developed in complement to Decision
ECC/DEC/(06)04 and is currently focused on the principle of a possible harmonised transition
measure applicable to frequency band 4.2 – 4.8 GHz since no technical requirements are currently
available for mitigation techniques considered in the frequency band 3.1 – 4.8 GHz.
Additionally to the technical requirements detailed in Table 7, new UWB devices are exceptionally
permitted (until 30 June 2010/2012) to operate in the frequency band 4.2 - 4.8 GHz with a maximum
mean EIRP density of –41.3 dBm/MHz and a maximum peak EIRP density of 0 dBm/50 MHz
without the requirement for additional mitigation.
Table 7:
Maximum EIRP limits for devices using UWB technology in bands below 10.6 GHz
Frequency range (GHz)
Maximum Mean EIRP
density (dBm/MHz)
Maximum Peak EIRP
Density (dBm/50MHz)
(Note 1)
Below 1.6
–90 dBm/MHz
–50 dBm/50MHz
1.6 to 3.8
–85 dBm/MHz
–45 dBm/50MHz
3.8 to 4.8
–70 dBm/MHz
–30 dBm/50MHz
4.8 to 6
–70 dBm/MHz
–30 dBm/50MHz
6 to 8.5
–41.3 dBm/MHz
0 dBm/50MHz
8.5 to 10.6
–65 dBm/MHz
–25 dBm/50MHz
Above 10.6
–85 dBm/MHz
–45 dBm/50MHz
Note 1: The peak EIRP can be alternatively measured in a 3 MHz bandwidth. In this case,
the maximum peak EIRP limits to be applied is scaled down by a factor of 20log(50/3) =
24.4 dB.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
29
ERA Report 2006-0713 (Issue 2)
-30
-40
EIRP (dBm)
-50
CISPR Class A
-60
CISPR Class B
FCC UWB Indoor
FCC UWB Outdoor
-70
Draft ECC UWB
-80
-90
1
3
5
7
9
11
-100
Frequency (GHz)
Figure 4: FCC and draft ECC UWB mean EIRP limits compared with CISPR average emission
limits in a 1 MHz band
Figure 4 above shows the comparison between CISPR average emission limits above 1 GHz for Class
A (Industrial) and Class B (Residential) and mean UWB limits set by the FCC and proposed by the
ECC. The plot shows that the draft ECC limits proposed are 24 dB below the FCC outdoor limits for
frequencies between 1.99 GHz and 3.1 GHz and 15 dB below 1.6 GHz.
UWB Interference measurements to Global Positioning System (GPS) receivers by the NTIA showed
that for UWB signals examined with a PRF of 20 MHz, the maximum allowable EIRP levels required
to ensure EMC with all of the GPS receiver applications considered ranged from -41.0 to
-76.9 dBm/MHz for the CW-like (non-dithered) UWB waveforms, and from -30.0 to -68.6 dBm/MHz
for the noise-like (dithered) UWB waveforms.
The measurements carried out in the NTIA report were based on a receiver sensitivity -130 dBm.
GSM 1800, DECT, 3G and RLAN all have higher sensitivity levels (Table 8) compared with a typical
GPS receiver and should therefore in theory, operate with FCC UWB EIRP limits of –75.3 dBm/MHz
(in a bandwidth no less than 1 kHz, as stated in Code of Federal Regulations, 47 Telecommunications,
Section 15.509, Part e), 10 dB more than the proposed draft ECC limit of –85 dBm/MHz.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
30
ERA Report 2006-0713 (Issue 2)
Table 8:
Radio Systems operating between 1 and 3 GHz
Radio System
Frequency (GHz)
Sensitivity (dBm)
ETSI Standard
GPS
1.5
-130
N/a
GSM 1800 Mobile
1.8
-102
TS 100 910 V.8.19.0
DECT
1.9
-83
EN 300 175-2
IMT-2000 (3G) Mobile
2.1
-117
RLAN
2.4
-80
1
1
EN 301 908-2
EN 300 328
Measured in dBm/3.84 MHz
4
UWB Applications
The ECC defines an UWB system as a radio system with a spectrum that occupies a spectrum greater
than 20% of the central frequency or with a bandwidth greater than 500 MHz [10]. According to ECC
Report 64 there are three types of applications for which UWB devices can be categorised:
1. Type 1: Communications and measurement systems.
2. Type 2: Imaging systems.
3. Type 3: Automotive radars.
Type 1 UWB devices include:
•
Home entertainment and networking (indoors, high density, low average utilisation).
•
Cellular phones multimedia interfaces (outdoor and indoor, high density, medium utilisation.
•
Wireless Personal Area Networks (WPAN) (indoor, hot-spot, low/medium average
utilisation).
•
Wireless Local Area Networks (WLAN) (e.g. similar to RLAN with enhanced capacity;
indoor, hotspot, high utilisation).
The ECC Report has also identified scope for combined data communication and measurement
systems (e.g. measurement and location recording devices) (outdoor and indoor, low density, low
utilisation) for Type 1 UWB devices.
Type 2 UWB devices will be mainly used for both indoor and outdoor applications and be based o
pulsed technology. These devices will have a low-density population with low-to-medium utilisation
for the following applications:
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
31
ERA Report 2006-0713 (Issue 2)
•
Ground Penetrating Radar (GPR).
•
In-wall imaging.
•
Through-wall imaging.
•
Medical imaging.
•
Surveillance devices.
•
Industrial liquid level gauges.
Type 3 UWB devices include automotive sensors, collision avoidance sensors and smart airbags.
These types of devices will most likely be based on pulsed UWB technology.
Based on the information provided by the ECC report, 98% of the systems are Type 1, of which 88%
of the systems are used for indoor applications and 10% are used for outdoor applications
Type 1 UWB applications are typically used for transmitting digital signals over a wide range of
frequencies with a data rate up to 500 Mbps over short distances in the region of 10 m. However
UWB technology can be also be used in applications requiring low or medium data rates such as
motion sensing and intrusion detection security systems. The transmission of digital signal will be
based on either Multi-band (MB)-Orthogonal Frequency Division Multiplex (OFDM) or Direct
Sequence (DS)-UWB technology, depending on market forces.
5
UWB Signal Characteristics
UWB signals are difficult to define. One definition of UWB signals describes the spectral emissions
as having an instantaneous bandwidth of at least 25% of the centre frequency. Other names for UWB,
or terms associated with it, include: impulse radio, impulse radar, carrier-less emission, time-domain
processed signal, and others.
Terminology and definitions aside, the UWB signal is, in general, a sequence of narrow pulses
sometimes encoded with digital information. UWB signal pulse widths are on the order of 0.2 to 10 ns
and longer. Some have an impulse-like shape and others have many zero crossings. One form of
modulation is Pulse-Position Modulation (PPM) where, for example, a pulse that is slightly advanced
from its nominal position represents a “zero,” likewise; a slightly retarded pulse represents a “one.”
Another form of modulation is On - Off Keying (OOK) where, for example, an absent pulse
represents a “zero.” In addition to the modulation scheme, the pulses can be dithered. In other words,
the pulse will be randomly located relative to its nominal, periodic location (Absolute Dithering) or
relative to the previous pulse (Relative Dithering). For example, 50% absolute dithering describes a
situation where the pulse is randomly located in the first half of the period following the nominal
pulse location. Finally, some UWB systems employ gating. This is a process whereby the pulse train
is turned on for some time and off for the remainder of a gating period.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
32
ERA Report 2006-0713 (Issue 2)
The frequency domain characteristics (emission spectrum) of a UWB signal are dependent on the
time-domain characteristics described above. The pulse width generally determines the overall shape
– envelope – of the emission spectrum. The bandwidth of the pulse spectrum generally exceeds the
reciprocal of the pulse width. If the pulse train is uniformly spaced, the emission spectrum will have a
series of lines. If dithering is used, there will be a smooth component of the emission spectrum in
addition to the line component. The higher the dithering, the greater the power contained in the
smooth component versus the line component. Some types of modulation can also reduce the spectral
line amplitude. Band limiting changes the characteristics of the UWB signal further.
Four different pulse spacing modes (UPS, OOK, ARD, and RRD) are illustrated in the following
Figure, whereby the vertical dashed lines represent the ticks of a clock. Gating is represented by the
removal of the pulses in the shaded areas; in the case of the UPS example, there are 4 pulses
generated during the gated-on time followed by 8 clock ticks for which there are no pulses (to give a
duty cycle of 33%).
Figure 5: UWB pulse spacing modes
The frequency domain characteristics (emission spectrum) of a UWB signal are dependent upon the
time-domain characteristics described above. The pulse shape/width determines the overall spectral
envelope, where the bandwidth is approximately equal to the reciprocal of the pulse width. The
manner in which the pulses are sequentially spaced determines the fine spectral features within the
confines of the envelope.
Spectral plots are shown in the figure below for four different UWB signals (UPS, OOK, ARD, RRD)
as they are passed through a band-pass filter. UPS has the power gathered up into spectral lines at
intervals of the PRF. The greater the PRF, the wider the line spacing, and the greater the power
contained in each spectral line. OOK also has spectral lines spaced at intervals of the PRF that are
superimposed on a continuous noise-like spectrum. Dithered signals have spectral characteristics
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
33
ERA Report 2006-0713 (Issue 2)
inherently different from either UPS or OOK. For these measurements, ARD has a pulse spacing that
is varied by 50% of the referenced clock period. RRD has a pulse spacing that is varied by 2% of the
average pulse period. Both of these dithered cases have spectral features that are characteristic of
noise (i.e., no spectral lines). The reader is referred to [11] for a more in-depth discussion of the
spectral characteristics of UWB signals.
Figure 6: Spectral characteristics of different pulse spacing modes
Another feature worth noting is the phenomenon of spectral lines spreading due to gating. The
spectrum of the gated UWB signal is the result of convolving the non-gated signal with that of a
rectangular function, the latter of whose Fourier transform amplitude has a sinc-squared envelope
characteristic. It follows that the single line of the non-gated cases is spread out into a multitude of
lines confined by the sinc-squared envelope, where the spacing between lines, or line spread spacing
(LSS), is equal to the reciprocal of the gating period; null spacing, or line spreading null-to-null
bandwidth (LSNB), of the main lobe of the sinc-squared function is equal to two times the reciprocal
of gated-on time.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
34
ERA Report 2006-0713 (Issue 2)
6
UWB Technology
6.1
Introduction
An UWB system is defined as one in which the fractional bandwidth is equal to or greater than 0.25 or
has a UWB bandwidth equal to or greater than 500 MHz, regardless of the fractional bandwidth.
UWB bandwidth is defined as the frequency band bounded by the points that are 10 dB below the
highest radiated emission, as based on the complete transmission system including the antenna [12].
The upper boundary is designated fH and the lower boundary is designated fL. The fractional
bandwidth equals
2( f H − f L ) ( f H + f L )
Eq. 1
A very wide bandwidth means better multi-path mitigation, interference mitigation by using spread
spectrum techniques, improved imaging and ranging accuracy, more users and higher data rate. A
lower centre frequency for a given bandwidth allows better materials penetration.
The UWB technology is still advancing, and according the ETSI Technical Report [13] due to the
large variety of UWB radio applications, it is very unlikely that a single type of signalling and
modulation scheme will be used in the future.
Nevertheless, according to the ETSI technical report and a study by Mason Communications Ltd [14],
it appears that the IEEE 802.15 standard committee of the Institute of Electrical and Electronics
Engineers (IEEE) has been investigating various UWB applications and technologies [15] particularly
for mobile and communication devices. The IEEE 802 Organisation is shown in Figure 7 below.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
35
ERA Report 2006-0713 (Issue 2)
IEEE 802 LAN/MAN Standards Committee
(wireless areas)
WLAN IEEE
802.11
WPAN IEEE
802.15
802.15.1
Bluetooth
WMAN IEEE
802.16
Regulatory TAG
IEEE 802.18
802.11
Coexistance
802.15.3 – High Data
Rate, 2.4 GHz
802.11.4
ZigBee
Task Group 4a (UWB
- Low Data Rate)
WBWA IEEE
802.20
Coexistence TAG
IEEE 802.19
Task Group 3a Alt
PHY (UWB)
Figure 7: IEEE 802 organisation
IEEE 802.15.3a of this standard committee was formed to identify medium to high data rate
applications and IEEE 802.15.4 for the low data rate applications. Task Group 3a has defined two
rival approaches for the physical layer (PHY) standard:
•
DS-UWB – Direct Sequence UWB, the Dual Band Direct Sequence technique
•
Multi-band OFDM - Orthogonal Frequency Division Multiplexing approach
These two physical layer standards are capable of supporting data rates between 110 Mbps and
480 Mbps over short ranges less than 10 m with low power technology.
In March 2005 the FCC granted a waiver that benefits both so-called Multi-band OFDM (MBOFDM), which "hops" a 500+ MHz wide OFDM symbols from one band to another, and gated DSUWB, which occupies a much wider switch of spectrum but switches on and off while in operation.
The waiver allows both technologies to take their measurements for compliance with the
-41.3 dBm/MHz limit in their normal operating mode, i.e. with the hopping or gating turned on. The
practical effect is to boost the useful operating power by several dB.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
36
ERA Report 2006-0713 (Issue 2)
While there are a number of competing standards which risk to make universally compatible UWB
products problematic in the short-term, Wireless USB - certified by the USB Implementors Forum
(USB-IF) - has selected WiMedia (MB-OFDM) as its underlying radio. The 1394 group on the other
hand has not standardized either the DS-UWB solution or the WiMedia (MB-OFDM) radio.
Bluetooth stakeholders have selected WiMedia Alliance's MB-OFDM for integration with current
Bluetooth wireless technology. UWB signaling is a candidate for the alternate physical layer protocols
for the high data rate IEEE 802.15.3a standard (where a DS-UWB and an MB-OFDM solution have
been proposed) as well as the low data rate IEEE 802.15.4a "ZigBee" wireless personal area network
(WPAN) standards where a DS-UWB pulsed-type system is being developed. The IEEE 802.15.4a
standard aims at providing a physical layer wireless communication protocol with ranging capabilities
for low-power applications such as sensor networks. The narrow duration of the direct sequence
modulated UWB pulses enables achievement of the stringent ranging accuracy (<1 m) requirements.
Unfortunately, in early 2006 an industry working group announced that it would disband, because the
two competing factions could not agree on a single UWB standard. Therefore consumers will once
again face a "VHS-versus-Beta" format war until one type of UWB technology gains a clear lead.
6.2
DS-UWB
DS-UWB is designed to occupy at least 1.5 GHz of spectrum in the 3.1 – 5.15 GHz and 3.7 GHz in
the 5.8 – 10.6 GHz frequency range. Unlike conventional Direct Sequence – Spread Spectrum (DS –
SS), the UWB approach uses non-sinusoidal wavelets (e.g., pulse type waveforms) tailored to occupy
the desired spectrum in an efficient manner. Figure 8 shows a sample wavelet, which occupies the
latter and larger bandwidth of the spectrum.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
37
ERA Report 2006-0713 (Issue 2)
Figure 8: DS-UWB wavelet and its spectrum in 6-10.6 GHz range
Various modulation schemes can be applied to the DS-UWB approach; BPSK, QPSK or Multi-level
Bi-Orthogonal Keying (M-BOK) depending on the data rate required. However, other modulation
schemes such as Pulse Position Modulation (PPM) and Pulse Amplitude Modulation (PAM) can also
be used.
It is designed to occupy at least 1.5 GHz of spectrum in the 3.1-5.15 GHz and 3.7 GHz in the 5.8 10.6 GHz frequency range. Unlike conventional DS-SS it uses non-sinusoidal wavelets tailored to
occupy the desired spectrum.
The UWB bandwidth spreading is accomplished by sending the wavelets at the 1.368 Gcps chip rate.
Modulation of the sequence of wavelets whitens the spectrum, which means that it disrupts
regularities that would otherwise result in spectral lines. M-BOK modulation comprises 24- and 32length ternary orthogonal sequences (-1,0,+1) forming symbol wavelets. Either 1, 2, 3 or 6 bits are
sent with each code symbol. For example, a 64-BOK modulation carries 6 bits at a time (26 = 64) to
form a symbol. The 24-length ternary codes are used with 2-BOK, 4-BOK and 8-BOK, while 32length codes are used with 64-BOK. M-BOK modulation has the property that as M increases without
bound, the modulation efficiency approaches the Shannon-limited value of –1.59 dB [15].
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
38
ERA Report 2006-0713 (Issue 2)
Symbols are sent at 42.75 MSym/s resulting in a channel bit rate of 256.5 Mbps. The data are encoded
with a rate 0.44 error-correction code resulting in the 112 Mbps information data rate seen in the table
below.
Table 9:
Possible set of DS-UWB signal parameters
Information Data
Rate
112 Mbps
224 Mbps
448 Mbps
Modulation
64-BOK
QPSK/64-BOK
QPSK/64-BOK
Symbol rate, Msys/s
42.75
42.75
42.75
Coding rate
0.44
0.44
0.87
Code Length
32
32
32
Channel chip rate
1.368 Gcps
1.368 Gcps
1.368 Gcps
With QPSK, there are up to 12 bits per symbol. This doubles the data information rate to 224 Mbps.
Finally a 0.87 rate error-correction code is used with QPSK to achieve 448 Mbps data rate
information. Many other combinations of modulation depths (M-BOK) and coding rates are possible.
The M-BOK codes, along with FEC, are especially effective in multi-path propagation. This example
of UWB uses wavelets that are approximately 1.5 GHz and 3.6 GHz, respectively, when an impulse
approach in which the impulses are sent with minimal spacing between them.
6.3
Multi-Band OFDM
OFDM technology is four decades old and it appears in many communications services, both wired
and wireless [15]. An OFDM system can support 110, 200 and 480 Mbps with additional data rates
possible (See IEEE 802 03/449). The technology meets the UWB requirement of a 500 MHz
bandwidth by using 128 carriers with QPSK modulation. The carriers are efficiently generated using a
128-point Fast Four Transform (FFT). Therefore, a composite OFDM signal can occupy a channel
with total bandwidth of 528 MHz (See Figure 9).
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
39
ERA Report 2006-0713 (Issue 2)
Figure 9: Single channel band occupancy of an OFDM UWB system
The composite signal occupies the 528 MHz channel for an information length of 242.42 ns plus an
additional 60.61 ns cyclic prefix time before switching to another channel within a 9.5 ns guard time.
Given the frequency band from 3.1 GHz to 4.8 GHz and the FCC requirement that UWB signals have
to be at least 500 MHz, only three sub-bands can be used in the initial deployment of multi-band
OFDM systems. Figure 10 shows a Multi-band OFDM signal using three 528 MHz channels at a time
in a time-frequency hopping manner, thus occupying a total of 1,585 MHz bandwidth of spectrum.
Figure 10: Expanded view of a multi-band OFDM UWB channel
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
40
ERA Report 2006-0713 (Issue 2)
This OFDM approach allows the UWB system to occupy a very large bandwidth for the purpose of
increasing the total radiated power. Frequency hopping technology in the form of spread spectrum
technology makes OFDM robust to narrow-band interferers. The technology can also comply with
local regulations by dynamically turning off certain channels or tones in the software. This flexibility
and dynamic nature of a multi-band OFDM system allows it to potentially co-exist effectively with a
wide range of current and future wireless technologies.
Since the channel or tone spacing is 4.125 MHz, the resolution of the multi-band OFDM system is
much narrower than the band resolution of 500 MHz for pulse-based multi-band systems. Any narrow
band interference will at most affect a couple of OFDM tones. The information in these tones can be
recovered through the forward error correction codes.
7
Measurement Techniques
UWB is a relatively new class of signal, which can occupy large bandwidths in excess of 500 MHz.
Ofcom is considering to allow UWB to be able to transmit on an unlicensed basis, even in frequency
bands occupied by licensed radio transmission services. Therefore there is a particular need to
characterise a UWB signal radiated across its full emission bandwidth. This bandwidth could possibly
extend from 3.1 to 10.6 GHz. This section describes the measurement process needed to characterise
UWB signals as described by NTIA Report 01-383 [16].
Two types of measurements can be performed to characterise UWB signal emissions:
1. Full bandwidth time domain measurements
2. Bandwidth limited measurements
The latter type of measurement is more applicable when observing the effects of UWB signals in the
IF section of a receiver.
7.1
Full Bandwidth Time Domain Measurements
Full bandwidth measurements can be made in the time domain using an oscilloscope with a
bandwidth of 20 GHz and a pair of wide band horn antennas. Figure 11 below shows the
measurement set-up for full bandwidth measurements. These measurements can be made on a
conducted and radiated basis and the results compared. For conducted measurements an attenuator is
used to prevent overloading and damage to the measurement instrument from too strong a signal
level.
A comparison of the test results gives insight into if a UWB source used in conducted measurements
is representative of itself when used in radiated measurements. When comparing results between
conducted measurements and radiated measurements a correction factor must be applied to the
radiated results data to take into account of the horn antennas decrease in effective aperture as a
function of frequency compared with an isotropic antenna.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
41
ERA Report 2006-0713 (Issue 2)
Radiated
Path
UWB
Source
Oscilloscope
Attenuator
Conducted
Path
Figure 11: Full bandwidth measurement system set-up
For the radiated experiment, the UWB source is characerised by setting up the measurements in an
anechoic chamber. The UWB device is allowed to radiate via a wide band horn antenna to the
receiving oscilloscope connected to another identical wide band horn antenna (
Figure 11). The separation between the transmitting and receiving antennas is set at 1 m.
Both sets of measured results are fast Fourier transformed and the RF emissions spectrum of the
UWB source can be used to determine the band-width at -3, -10, -20 dB points. The inverse of the
width of the 10 and 20 dB spectrum should approximately be equal to the pulse width emitted by the
UWB source [16].
Given that there are i sample points in the time-domain waveform and x is the ith sample point, total
peak power (for a 50 Ω terminated load) can be calculated at the maximum ith value using:
Ppeak =
xi2
50
Eq. 2
The average power can be calculated by using:
Ppeak
1
=
PRI
⎛ xi2 ⎞
∑i ⎜⎜ 50 ⎟⎟ × Δt
⎝ ⎠
Eq. 3
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
42
ERA Report 2006-0713 (Issue 2)
Where xi is the ith time-domain sample, Δt is the time sample of the oscilloscope trace and PRI is the
pulse repetition interval between adjacent pulses.
7.2
Bandwidth Limited Measurements
In the majority of cases, UWB signals will interfere with radio receiver equipment whose IF
bandwidth will be much less than the interferer. Therefore, the characteristics of UWB signals must
be measured using equipment that is bandwidth limited.
In this section, NTIA techniques are described to measure the following UWB parameters:
•
Emission spectra as function of IF measurement bandwidth
•
Peak, average and RMS power
7.2.1 Spectrum envelope measurements
The general laboratory test configuration for measuring UWB emissions using band limited
equipment is shown in Figure 12.
Radiated
Path
UWB
Source
Pre-amp
Spectrum
Analyser
Oscilloscope
Wideband
Detector
Attenuator
Conducted
Path
Figure 12: Bandwidth limited measurement system set-up
The majority of measurements can be made on a radiated basis (with appropriate corrections for
antenna gain, effective aperture etc), with conducted measurements performed as necessary.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
43
ERA Report 2006-0713 (Issue 2)
For the radiated experiment, the UWB source is characerised by setting up the device in an anechoic
chamber. This will eliminate or minimise any interference multi-path and external radio signals to the
measurement system.
The measurement antenna is placed 1 m to the UWB source, as the UWB device emissions may have
a relatively low-amplitude. The measurement antenna must be calibrated for the distance at which
measurements are to be performed. Again, wideband horn antennas have proven to be workable in
this respect, as they typically can be acquired with calibration curves made at 1 m. If the measurement
antenna does not have a constant effective aperture (i.e. if the antenna gain does not increase as
20.log10(frequency), the resulting spectra must be corrected to the constant effective aperture case.
Section C.4 of Appendix C of the Institute of Telecommunications Service (ITS) report describes the
necessary spectrum corrections for non-constant effective aperture measurement antennas [16].
If the UWB emissions are found to be weak and not observed above the noise floor of the spectrum
analyser, then a pre-ampilier can be used at the output measurement of the receiving antenna. In an
optimised measurement system, the sum of the pre-amplifier gain and pre-amplifier noise figure
should be nearly equal to the noise figure of the spectrum analyser across the frequency range to be
measured.
Although UWB device emissions may have low amplitude within the convolution bandwidth of the
measurement system, the total power convolved by the front-end pre-amplifier may be high enough to
cause overload of that component. The pre-amplifier may also experience overload due to ambient
signals, if the measurements are not performed in an anechoic chamber. If overload is experienced,
then appropriate RF filtering will be required between the antenna and the pre-amplifier.
A spectrum analyser may be used to measure both the bandwidth-limited spectra and time domain
information (when operated in zero hertz span). As noted above, the pre-amplifier noise figure and
gain should be optimised to work with this noise figure.
For IF bandwidths greater than the spectrum analyser bandwidth a wideband detector connected to an
oscilloscope can be used to determine the UWB emission levels instead (See Figure 12).
The NTIA have shown that swept-frequency measurements have been shown to be fast and practical
for UWB device emissions. Only at bandwidths narrower than about 30 kHz does the sweep time
exceed a few seconds across the requisite several gigahertz of spectrum [16].
NTIA Report 01-383 [16] suggests that the UWB signal level should be at least 10 dB above the
measurement receiver inherent noise level for the largest measurement bandwidth. The radiated
emission spectrum of each UWB device will be measured to determine the -10 and -20 dB
bandwidths. The inverse of the width of the 10 and 20 dB spectrum should approximately be equal to
the pulse width emitted by the UWB source.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
44
ERA Report 2006-0713 (Issue 2)
7.2.2 Peak power measurements
A spectrum analyser can used to determine peak power in either a spectrum trace or at a single
frequency in a zero hertz span mode. A significant problem is to measure the total peak power in the
emission bandwidth of an UWB device. This is not generally feasible in a direct measurement
therefore the widest resolution bandwidth will be used and an appropriate bandwidth correction factor
applied (See ITS Ultra – Wideband Measurement Plan [17]).
7.2.3 FCC Part 15 average power measurements
Regarding measurements using an average detector, the FCC’s measurement procedure in an average
logarithm detector process is not equivalent to a Root Mean Square (RMS) detector process.
Measurements have shown that the average logarithm detector is largely insensitive to energy
contained in low-duty-cycle, high amplitude signals. This results in Part 15 measurement values that
can be substantially lower (10-15 dB) than the RMS power in a UWB signal. Although the NTIA
recognises that no single average detector function adequately describes the interference effects of
UWB signals, the measurement results indicate that a RMS detector function better quantifies the
potential interference affects of UWB signals than the average-logarithmic detector function used for
Part 15 compliance [18].
7.2.4 RMS power measurements
A power meter can be directly used to measure the RMS power UWB signal received by the spectrum
analyser as shown in Figure 12. The NTIA recommend a power meter utilizing a bolo-meter type
sensing head [16]. The UWB signal must be at least 10 dB above the thermal noise in the
measurement head, so that the power being averaged is primarily an indication of the UWB device
emission and not of the thermal noise in the power meter.
Alternatively, spectrum analyser with RMS detection mode capabilities can be used to measure the
RMS power.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
45
ERA Report 2006-0713 (Issue 2)
7.2.5 APD measurements
Amplitude probability distribution (APD) measurements show the percentage of time that emissions
from a device exceed a given power threshold. APD plots are typically produced on Rayleigh scales,
as shown in Figure 13.
Figure 13: Typical APD curve
The horizontal scale shows the probability of exceeding a given power on the vertical scale. The
vertical scaling is in terms of power relative to the thermal noise level and is therefore proportional to
bandwidth. It is therefore ‘correct’ for the AWGN power, which scales linearly with bandwidth.
AWGN appears as a downward sloping line on the right hand side of the APD.
Interference non-noise-like e.g., pulsed, on the other hand, appears as the raised section on the left
hand side of the APD. The power in the interference is proportional to the square of bandwidth, so
this section will be more pronounced if the measurements have been made with a wider bandwidth.
The far left hand end of the APD has a plateau region, which represents very low probability, high
amplitude interference.
An APD curve can show the entire time-occupancy distribution at a frequency and in a selected
bandwidth. It may be processed to show peak level in the bandwidth and several averages, including
root-mean-square (RMS) (that is, linear average power).
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
46
ERA Report 2006-0713 (Issue 2)
The R&S FSU46 spectrum analyser used in the UWB measurements incorporates APD functionality
as shown in the figure below.
Figure 14: APD graph of AWGN produced by a R&S FSU46 spectrum analyser in a 1 MHz
bandwidth
8
Measurement Environments
8.1
Open Area Test Site
The OATS is a common test facility at EMC establishments. It is used for emissions compliance
measurements for measurements below 1GHz. Ideally, an OATS has no reflecting material present
other than the reflecting ground plane. In theory, the OATS is suitable for measurements at extended
frequencies. However, the imperfection of the ground plane means that in reality the performance at
higher frequencies may be reduced.
The OATS test method has several deficiencies such as weather and ambient vulnerability, ground
plane mutual coupling and reflection effects as well as incomplete angular coverage of equipment
under test radiation pattern [19]. The ambient environment can be a problem when measurements
have to be performed at low sensitivity levels in the frequency range where intentional emitters are
present such as GSM 900 & 1800 or 3G. The RF ambiance of the OATS can add considerable
uncertainty to the measurements.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
47
ERA Report 2006-0713 (Issue 2)
8.2
Anechoic Rooms
An anechoic room is a shielded metal enclosure, usually whose internal walls and ceiling is covered
with radio absorbing material, normally of the pyramidal urethane foam type. If absorbing material
covers the floor, which is also usually metallic, the anechoic room is considered as a FAR, if not then
as a SAR. The advantage of these environments is that the shielding environment improves the
uncertainty of the measurement by protecting the interior of the chamber from external source of
disturbances. Nevertheless, uncertainty contributions are introduced by the use of anechoic rooms
such as reflectivity of the absorbing material, receiving and transmitting antenna images.
A Semi-anechoic Room attempts to simulate an ideal Open Area Test Site whose primary
characteristic is a perfectly conducting ground plane of infinite extent. The level of interference from
ambient signals are reduced by the shielding, whilst the radio absorbing material reduces unwanted
reflections from the walls and ceiling which can disturb the measurements.
A Fully Anechoic Room attempts to simulate free space conditions. The anechoic room generally has
several advantages over other test facilities such as minimal ambient interference, minimal floor,
ceiling and wall reflections. However, there are some disadvantages, which include limited measuring
distance mainly due to available room size and the cost.
Both absolute and relative measurements can be performed in an SAR and in FAR. In general, it is
recommended to measure the emission of low-level signal in a high performance anechoic chamber.
Along with the FCC and NTIA recommendation, CISPR recommends that measurements below
1 GHz be made in a Semi-anechoic Room and measurements above 1 GHz be made in a Fully
Anechoic Room.
8.3
GTEM Cells
The GTEM cell does not measure an OATS equivalent field strength at a given distance but the total
radiated power from the equipment in a given receiver bandwidth. From this measured power, a field
strength is calculated, by considering the source as dipole or multi dipole. This field strength is then
used for compliance evaluation against field strength limits. The mathematical process is called the
correlation algorithm. With limits directly specified in radiated power the GTEM cell can be used
directly.
A GTEM cell generally operates with one mode of polarisation; therefore, it is necessary to rotate the
equipment through three or more directions. Cables of the equipment have to be routed in order to
allow these rotations. However, measurements performed by York EMC services [21] indicate that in
some cases a single GTEM test position can show good correlation even for non-dipole like
equipment under test. In that case, the position of the maximum radiated emission is used and of
course well known.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
48
ERA Report 2006-0713 (Issue 2)
The European Standard EN 61000-4-20 [20] indicates that the TEM waveguide can be used both for
emission or immunity measurement. Annex A of the standard defines the emission testing
procedures. The method describes in this annex is applicable only for small Equipment Under Test
(EUT)s, which have to comply with disturbance limits given in terms of an OATS field strength at a
specific distance. Power and signal cables attached to the Equipment Under Test are outside the
defined acceptable volume and can give rise to unwanted variations in measured values due to internal
reflections.
In this standard, equipment is defined as a small if the largest dimension of the case is smaller than
one wavelength at the highest test frequency (for example, at 1 GHz λ = 0.3 m), and if no cables are
connected to the EUT. All other EUTs are defined as large and the procedures for large equipment are
under consideration.
The main advantages of using GTEM cells are [21]:
•
They fully closed and do not suffer from ambient noise.
•
No antenna set-up is required, since the cell itself functions as a receiving structure. This
makes the test set-up simpler than on an OATS, and avoids the need to change the receiving
antennas for different frequency ranges.
The main disadvantages of using GTEM cells are:
•
The cross-polarisation performance is considered inferior to an anechoic chamber or OATS.
Over a limited frequency band the field level of the longitudinal mode can exceed the level of
the intended vertical field.
•
The size of EUT is limited to approximately one-third height between the septum and floor,
and any attached cables will go outside the acceptable volume.
•
It is difficult to determine the measurement uncertainty because of the cross-polarisation
performance being inferior to an anechoic chamber or OATS.
A comparison of EMC radiated emissions results between a FAR and GTEM have shown good
agreement, mostly within ± 5 dB up to 12 GHz, but above 12 GHz there is around 10 dB more signal
in the GTEM cell [22]. A higher estimated field level is expected from the GTEM cell because data is
taken from three orientations. However, NPL experiments have shown that the results in the GTEM
cell were sensitive to location of the EUT within a few centimetres, giving variations of 2 dB up to
2.5 GHz and 7 dB above 2.5 GHz [22].
From the emission measurements point of view, the GTEM cell generally over-tests the EUT [23].
Calculations have shown that the GTEM cell will have similar sensitivities compared with a FAR
(See Appendix A).
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
49
ERA Report 2006-0713 (Issue 2)
8.4
Reverberation Chamber
A reverberation chamber is a shielded enclosure with the smallest dimension being large with respect
to the wavelength at the lowest useable frequency. The chamber is equipped with a mechanical
tuning/stirring device, which is used to redistribute the field energy in the chamber.
The chamber is excited with RF sources; the number of tuner steps defines the specified field
uniformity. For a sufficient number of positions of the mechanical tuner, the environment is
statistically uniform and statistically isotropic. In some chambers, several mechanical tuners have to
be installed to obtain the desired field uniformity at required frequencies. Stepping motors or
continuous motors can be installed, however, stepping motors are recommended. The frequency range
of tests is determined by the size and construction of the reverberation chamber, as well the
effectiveness of the mechanical tuners. Room-sized reverberation chambers typically operate from
200 MHz to 18 GHz without limitations. It can be used both for radiated emission and immunity tests.
A reverberating chamber is also characterised by a quality factor, which determines its time constant.
This is an important parameter for pulse testing. Indeed, if the chamber time constant is greater than
0.4 times the required pulse width for more than 10% of the test frequencies, then absorbers need be
added or the pulse width increased. Short pulse durations can be distorted by this quality factor of the
chamber.
Annex A of the European standard EN 61000-4-21 [24] indicates that the amount of RF power
radiated by a device placed in the chamber can be determined by measuring the amount of power
received by the reference antenna and correcting for the insertion loss of the chamber. The power
radiated from a device can be calculated using either average or peak received power from a given
number of tuner steps and/or tuner rotations. Results based on mean power is better because there is a
lower uncertainty, nevertheless, the measurement system must have sensitivity of 20 dB lower than
the actual mean to get an accurate average measurement and intermittent signals may be artificially
lowered due to insufficient sampling.
Annex B describes the chamber calibration for mode tuned (stepped tuner rotation) operation and
Annex C the mode-stirred (continuous tuner rotation) operation. The mode-stirred technique can be
faster than the mode-tuned technique; indeed the equipment under test is exposed adequately to the
continuously changing field within the chamber. However, the mode stirred technique is only adapted
for equipment under test with short response time.
Annex E of the European standard defines the procedures for measuring radiated intentional and
unintentional emissions with a reverberation chamber. As with radiated immunity testing, chamber
calibration data described in Annexes B and/or C is used to determine radiated emissions levels. This
annex indicates that the mode-stirred procedures should only be applied for non-modulated signals
using peak detector. The motion of the tuner is at the origin of amplitude variations of the signal,
therefore the testing time will increase if a peak detector is used. Furthermore, this mode is not
applicable when using an average or other weighting detector.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
50
ERA Report 2006-0713 (Issue 2)
The radiated mean power within the measurement bandwidth is measured for modulated emissions, if
a RMS detector is used. For an emission spectrum wider than the measurement bandwidth, the total
radiated mean power is measured by integrating the mean power spectral density over the emission
spectrum.
The main advantages of using a reverberation chamber are:
•
The total radiated power can be measured without rotating the equipment under test, whereas
a GTEM cell may require several positions to ensure a full capture of the total radiated power.
•
Emissions from a low powered radiating device can be magnified, by using the mode stirred
technique and be measured more easily.
•
Excellent environment to test repeatability – good and accurate comparison technique.
•
Set-up time is short – as cable positions and location of equipment within the chamber are
insignificant, providing the EUT and cables ≥ λ/4 away from the chamber boundaries and
paddle wheel.
•
Lower measurement noise floor, for same analyser/set-up used in a FAR – this is related to
both the chambers ‘Q’ factor and the antennas characteristics within the chamber.
The main disadvantages of using a reverberation chamber are:
•
The measurement system must have sensitivity of 20 dB lower than the actual mean to get an
accurate average measurement and intermittent signals may be artificially lowered due to
insufficient sampling.
•
Reverberation chambers are currently accepted in only a few standards. Utilisation of
reverberation chambers for emissions testing will require acceptance of total radiated power
as a pass/fail criterion.
•
There is a lowest usable frequency defined by the cavity dimensions, tuner effectiveness, and
cavity quality factor.
•
Directivity and polarisation effects are not measurable. The high quality factor in
reverberation chambers may impose constraints on pulse testing. Depending on test
conditions, correlation of test results to other test techniques may be difficult or impractical.
ERA Stage B Report on “Emission measurements in the range 1 to 6 GHz in Fully Anechoic Room
and Reverberation Chamber Facilities” [25] concludes that there is a reasonable correlation between
the FAR and the reverberation chamber results, further work is required to investigate the appropriate
gain factor as suggested in BS EN 61000-4-21. There are also good correlations between the
reverberation mode- stirred and mode-tuned measurements.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
51
ERA Report 2006-0713 (Issue 2)
On average the reverberation chamber gave a reduction in field strength of 7 dB to 10 dB +/- 5 dB,
but there was clear and apparently uniform frequency dependence. This could lead to a factor, which
could be included in EN 61000-4-21 to obtain representative fields when using the reverberation
chamber, but further work is required to determine the value and precision of the gain figure to be
applied.
The measurement results in a 1 MHz RBW using a peak detector also showed that the reverberation
chamber had approximately a 20 dB better sensitivity compared with the FAR. However, as already
stated in EN 61000-4-21, main disadvantage of the reverberation chamber is that the measurement
system must have sensitivity of 20 dB lower than the actual mean to get an accurate average
measurement and intermittent signals may be artificially lowered due to insufficient sampling. Thus,
effectively eliminating any sensitivity advantage gained.
Other work has shown the EUT in the reverberation chamber to give a very good comparison with the
FAR data, but 4.5 dB higher. This may be expected, because the reverberation chamber captures the
total energy, whereas the measurements in the FAR measure emissions in a single azimuth plane [22].
9
Measurement Considerations
9.1
Test Environment
There are likely to be some considerations given to convergence of the test facilities for all EMC
radiated emission (and immunity) measurements and there is likely to be convergence on two basic
facilities, Fully Anechoic Rooms and Reverberation Chambers. In IEC/CISPR there have been
several attempts at a unified test facility for both emissions and immunity testing where a single set up
and EUT configuration for both tests would be used. This would be valuable where existing
commercial facilities can be updated with minimal changes.
The reverberation chamber, which has significant advantages for some EMC immunity test work,
mainly in the military sector where very high fields are required to simulate the effects of powerful
transmitters at close distances. Work has been done to investigate the possibility of emissions
measurements and the main advantage is that the results are in the form of the total effective radiated
power from an EUT which can be interpreted to assess the possible interference potential of the
equipment. These limits have yet to be fully developed but will permit an effective assessment of a
product which does not require a search for high gain lobes and should give more accurate
assessments. The disadvantage is that the facility costs are quite high and will not be appealing to
commercial businesses who wish to invest in facilities.
Many standards published by the European Telecommunications Standards Institute (ETSI) contain
requirements for the control of radiated emissions at frequencies above 1 GHz [7] [8]. Both OATS
and indoor test sites may be used for the radiated emission measurements. For indoor sites a FAR is
recommended. An acceptable alternative is the SAR with a conductive ground plane, which gives rise
to a ground reflection and consequently requires antenna height variations to determine the maximum
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
52
ERA Report 2006-0713 (Issue 2)
signal for both EUT measurements and calibration. Similar arrangements apply on the OATS, but
there are additional requirements to extract the wanted observed signals from the ambient.
From the ERA/BT work [26], the recommended test facilities for radiated emission measurements
above 1 GHz are a FAR, a SAR with floor absorber or an OATS with floor absorber.
The measurement procedures developed by the NTIA [16] recommend installing the UWB device in
an environment where measurement system multi-path and external radio signals are eliminated or
minimized. The best possible choice is a high-performance anechoic room.
Communication papers in Japan describes the measurement techniques emissions from UWB devices
discussed [27] also recommends measuring the emission in a high performance anechoic room. Below
1000 MHz a Semi-anechoic Room may be used, in addition a reverberation chamber may be
employed for average power measurements. In [27] a separation distance of 3 m is used for the
measurements, nevertheless, this distance is be reduced for low-level UWB signals. The NTIA [16]
have performed the radiated emission measurements at 1 m.
9.2
Detectors
Work has been done in CISPR to determine the most optimum detector function for the assessment of
potential interference to digital modulated radio services. The use of the quasi-peak detector, which
was developed to include the subjective annoyance factor in the assessment of interference to
amplitude modulated (AM) radio services, primarily in the long and medium wave broadcast bands,
was extended to the frequency range 30 - 1000 MHz. For many standards such as EN 55022 there
was little impact because the main radiated emission disturbance types were the harmonics of the
clock and the quasi-peak levels were lower than the peak levels but only by relatively small dB.
However, it has long been recognised that the quasi-peak detector is not ideal for modern digital
communications where the modulation schemes are more complex. There has been some work,
which concludes that a complex detector having peak and RMS responses for different pulse
repetition frequencies may be the way forward and there will be more consideration of these issues
later in the project. However, alternative methods such as APD have also been the subject of research
and submission to CISPR.
The ETSI standard quotes CISPR 16-1 for the instrument settings, and consequently measurements
would be performed with a resolution bandwidth of 1 MHz and both peak and average detectors are
called up, the latter by limiting the video bandwidth. ETSI standards also often quote ETSI TR 102
273-2 [28] on the methods for measuring spurious emissions in absorber lined chambers giving much
good practical advice on making the measurements and recording the data.
9.3
Sensitivity of Equipment
The receiver sensitivity is defined as the minimum discernable signal that is just above the receiver
noise floor. The receiver noise floor is bandwidth dependant. The source of noise is due to thermal
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
53
ERA Report 2006-0713 (Issue 2)
effect. The power spectral density of this noise source is described by the Planck distribution and can
be approximated by a white noise spectrum. The theoretical noise floor N floor = 10 log( KTBw 103 ) ,
where K is Boltzmann’s constant; T is the temperature in Kelvin; Bw is the measurement bandwidth.
A multiplication factor of 103 is used to convert from Watts to milli-watts [29] to give the result in
dBm.
For instance for a particular measurement the sensitivity for a measurement bandwidth of 1 MHz, the
theoretical noise floor is –113.9 dBm. At this noise floor additional components within the receiver
raise the noise floor above this level.
This analysis does not consider the ambient level at a particular facility, the gain of the antennas, the
amplifiers, which may be used and then the path loss in free space. Therefore, the minimum
discernable signal M as given in [29], can be calculated with the simple relation:
M (dBm ) = 10 log( KTBw103 ) + 20 log(
4πd
λ
) + Rnf + Gr
Eq. 4
Where Rnf is the receiver noise floor and Gr is the receiver antenna gain. This relation assumes the
antenna has negligible noise figure. Typical noise figure for a spectrum analyser is 23 to 26 dB, giving
a minimum theoretical sensitivity of –91 to –88 dBm in a 1 MHz bandwidth.
9.4
Radiation Pattern
The radiation pattern of the equipment under test can be fairly uniform in the two polarisations around
frequencies of 1 GHz. As the frequency increases the radiation pattern may present more variations
with higher gain lobe. This implies that measurements must be made over the surface around the
equipment or eventually over the entire surface encompassing this equipment with rotational steps,
which decreases as the frequency increases.
It is highly recommended to use a non-conductive turntable for finding the orientation that provides
the maximum response within the measurement equipment. An analysis performed by York EMC
services Ltd [21] also demonstrates that care must be taken in material selection for all elements of a
test facility for use in making measurements at microwave frequencies even for non conductive
material.
9.5
Far Field Conditions
An electromagnetic wave consists of a magnetic field and an electric field. Close to the source of the
electromagnetic wave the relative magnitude between the magnetic field and the electric field depends
on the distance from the source and on the nature of the source itself. In the reactive near-field region
the field is either predominately magnetic or predominately electric. At a certain distance from the
source the two components of the wave become perpendicular each other and the direction of
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
54
ERA Report 2006-0713 (Issue 2)
propagation is at right angles to the plane containing the two components. At this distance the far field
conditions are achieved and the wave is said to be a plane wave. In practice there is also a third region
called radiating near field, where the distribution of the electromagnetic wave becomes more uniform
with the distance from the source. Placing receiving equipment close to the source can lead to
measurement inaccuracies since energy can be coupled by induction as well as by radiation. The
reactive energy is bigger than the radiated energy and therefore the measured result does not coincide
with the emission power.
The boundaries of these three regions are generally well defined for canonical antennas such as a
dipole. Generally, the measuring equipment has to be installed at a distance defined as d = 2(d1+d2)2/ λ
where d1 is the largest dimension of the EUT or substitution dipole and d2 is the largest dimension of
the test antenna. At this distance, it is considered that the far field conditions are achieved.
Furthermore, the measurement standard EN 55016-2–3 [31] define the measurement distance d =
2(D)2/ λ, where D is the largest dimension of either the EUT or the antenna determining the minimum
aperture for the illumination of the EUT, which applies to situations where D >> λ.
This relation shows that for an EUT dimension of 1 m at 3 GHz, the measurement distance is 20 m.
Therefore, it is clear that for many cases, it will not be practical to make measurements at this distance
both from the point of view of test site construction and measurement equipment sensitivity. York
EMC services Ltd [21] concluded that measurements have to be made in the near field conditions
[21]. However, the emissions from many products are due to leaking apertures and discontinuities in
the equipment case or sections of track on the PCB. The dimensions of these sources are much
smaller than 1 m, more like 10 cm, and consequently a test distance of 3 m is in the far field region of
many products.
10
Test Methodology
Bandwidth limited measurements were performed to characterise UWB signal emissions: The
methodology used to carry out the tests is described in Section 7 of this document. A Fully Anechoic
Chamber was decided as the best environment to perform the UWB radiated emissions tests based on
the NTIA test methodology and measurement environment described in Sections 7 and 9.1.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
55
ERA Report 2006-0713 (Issue 2)
11
Measurement Results
11.1 Characterising UWB Signals
11.1.1 Pulsed UWB
Conducted bandwidth limited measurements were made using a MSSI TP1001 pulsed generator
connected to a Marconi 20 GHz programmable attenuator. The output of the attenuator was measured
using an R&S FSU46 spectrum analyser. The attenuator was set to 10 dB and the HP 8116A pulse
generator controlling the Pulsed Repetition Frequency (PRF) of the MSSI UWB device was set to
5 MHz.
HP 8116A
Pulse Gen
MSSI
TFP1001
Attenuator
Spectrum
Analyser
Figure 15: Diagram of the set-up to measure a pulsed UWB signal
It was confirmed that the spectral lines of the UWB signal displayed on the spectrum analyser could
be resolved if the Resolution Bandwidth (RBW) < PRF, using the set-up shown above.
Pulsed UWB @ 3 GHz
Power (dBm)
-20
-30
RMS
-40
Peak
-50
-60
0.01
0.1
1
10
100
RBW (MHz)
Figure 16: Measured power as a function RBW for a pulsed UWB device with 5 MHz PRF
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
56
ERA Report 2006-0713 (Issue 2)
From Figure 16, it can be seen that the peak power varies as 20.log10 (RBW) and the RMS power
varies as 10.log10 (RBW) for RBW > PRF.
For RBW ≤ 0.2×PRF the amplitude of the spectral lines remained reasonably constant when using
peak detection. When centred on a spectral line, the amplitude of the line was the same for peak and
RMS detection and therefore indicating behaviour similar to CW.
This is expected because there will be only one pulse being measured at a time when the RBW is less
than the PRF. As the RBW increases more lines are included in the pass-band and the power increases
linearly with RBW. At 1 MHz and below there is a single line in the pass-band so that the power
remains constant, as shown in Figure 16 above.
For accurate measurements, the sweep time of the detector was set to be greater than SPAN/ RBW2.
The SPAN was set to 50 MHz. For a RBW 1 MHz and below, a Video Bandwidth (VBW) matching
the RBW and 10*RBW was used for peak and RMS detection respectively. For RBW of 3 and
10 MHz a VBW of 10 MHz was used, at RBW of 20 and 50 MHz a maximum VBW of 30 MHz was
used.
11.1.2 Dithered UWB
The pulse generator device was triggered to produce a PRF of 5 MHz using a HP 8116A pulse
generator. The UWB signal was dithered (Figure 17) by applying a FM signal to the HP pulse
generator triggering the TFP1001 device [32].
LF Signal
HP8116A
with FM
MSSI
TFP1001
Dithered
Pulse
Figure 17: Diagram of the set-up of a dithered UWB signal
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
57
ERA Report 2006-0713 (Issue 2)
The FM signal was generated using the LF output from a signal generator. The percentage of
dithering was controlled by the LF frequency from the signal generator and the pulse output frequency
of the HP8116A device. The dithered UWB signal was measured using a spectrum analyser.
Dithered UWB @ 3 GHz
Power (dBm)
-20
-30
RMS
-40
Peak
-50
-60
0.01
0.1
1
10
100
RBW (MHz)
Figure 18: Measured power as a function RBW for a dithered UWB device with 5 MHz PRF
From Figure 18, it can be seen that the peak and RMS powers vary with a trend of 10.log10 (RBW).
The results are consistent with NTIA Report 01-383 predictions.
Exact sweep time constants, VBW and analyser SPANS used for the pulsed UWB results were used
for the dithered UWB measurements.
11.1.3 Simulated OFDM-UWB
An R&S AMIQ in conjunction with an Agilent E4438C signal generator was used to simulate an
OFDM-UWB signal.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
58
ERA Report 2006-0713 (Issue 2)
Signal
Generator
AMIQ
Spectrum
Analyser
Figure 19: Bandwidth limited measurements for a simulated OFDM-UWB signal using a
spectrum analyser
Measurements were made at a frequency of 3 GHz using the following OFDM signal parameters:
Table 10:
Simulated OFDM-UWB signal parameters
OFDM Signal Parameter
Value
RF Frequency
3 GHz
RF Carrier Power Level
-28 dBm
Bandwidth
100 MHz
The OFDM signal created by the AMIQ was connected to the external I/Q ports of the E4438C signal
generator with a RF carrier frequency of 3 GHz. A carrier power of –28 dBm was used to produce an
OFDM signal level as close to FCC RMS limit –41.3 dBm in a 1 MHz bandwidth. Peak and RMS
measurements were made using an R&S FSU46 spectrum analyser connected to the signal generator.
Note, that bandwidth of the signal was limited to the 100 MHz, because the AMIQ could not sample
at a rate greater than 100 MHz for each I/Q port.
The relation between RF signal level and RBW was verified by plotting the power measured as a
function of RBW for peak and RMS detection. The graph below supports the theory of the simulated
OFDM signal generated by the set up in Figure 19 increase as 10.log10 (RBW) for RMS detection.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
59
ERA Report 2006-0713 (Issue 2)
OFDM @ 3 GHz
RMS Power (dBm)
-20
-30
-40
OFDM
AWGN
-50
-60
-70
0.01
0.1
1
10
100
RBW (MHz)
Figure 20: Measured RMS power as a function RBW for a simulated OFDM-UWB signal
OFDM @ 3 GHz
Peak Power (dBm)
-10
-20
-30
OFDM
-40
AWGN
-50
-60
0.01
0.1
1
10
100
RBW (MHz)
Figure 21: Measured peak power as a function RBW for a simulated OFDM-UWB signal
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
60
ERA Report 2006-0713 (Issue 2)
Figure 21 above shows a similar trend for the measured peak power of the simulated OFDM signal
when compared to the RMS results. The plots also reveal that the peak results are 4 to 6 dB higher
compared with the RMS results. Both sets of results match well when compared with the AWGN
results, which verifies the 10.log10 (RBW) trend. Thus, proving that the simulated OFDM-UWB
signal has noise-like characteristics.
11.1.4 MB-OFDM
A Wisair device was used to measure the signal characteristics of MB-OFDM UWB signal as shown
in the diagram below.
MB-OFDM
UWB Source
Spectrum
Analyser
Figure 22: Bandwidth limited measurements for a MB-OFDM UWB signal using a spectrum
analyser
Measurements were made at a frequency of 3.96 GHz using the following MB-OFDM signal
parameters:
Table 11:
Wisair MB-OFDM UWB signal parameters
MB-OFDM UWB Signal Parameter
Value
RF Frequency
3.96 GHz
No of Bands
3
Bandwidth per Carrier
528 MHz
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
61
ERA Report 2006-0713 (Issue 2)
MB-OFDM UWB
RMS Power (dBm)
-20
-30
-40
MB-OFDM
-50
AWGN
-60
-70
0.01
0.1
1
10
100
RBW (MHz)
Figure 23: Measured RMS power as a function RBW for a MB-OFDM UWB signal
MB-OFDM UWB
Peak Power (dBm)
-10
-20
-30
MB-OFDM
-40
AWGN
-50
-60
-70
0.01
0.1
1
10
100
RBW (MHz)
Figure 24: Measured peak power as a function RBW for a MB-OFDM UWB signal
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
62
ERA Report 2006-0713 (Issue 2)
Figure 24 above shows a similar trend for the measured peak power of the simulated OFDM signal
when compared to the RMS results (Figure 23). The plots also reveal that the peak results are 4 to
6 dB higher compared the RMS results. Both sets of results match well when compared with the
AWGN results, which verifies the 10.log10 (RBW) trend. Thus, proving that the simulated OFDMUWB signal has noise-like characteristics.
11.2 Conducted Low Level UWB Emissions
The Wisair device was used to measure lowest possible RMS signal levels from a MB-OFDM UWB
device using the conducted set-up as shown in Figure 22. The external antenna was removed form the
UWB device and a SMA cable was connected directly to the R&S ESPI7 spectrum analyser. Three
Resolution Bandwidths of 10 kHz, 100 kHz and 1 MHz were used to measure the spectrum analyser
noise floor, a MB-OFDM signal at 3.96 GHz and an UWB signal at a frequency approximately 6 dB
above noise floor as the minimum UWB signal that could be measured without the noise of the
analyser of the contributing to the measurement. The measurements were made in a 200 MHz span
and the sweep time was set greater than the required minimum of SPAN/ RBW2, as shown in Table
12.
Table 12:
Conducted measurements of MB-OFDM UWB signals close to the noise floor
RBW (kHz)
Power @ Center Frequency
3.96 GHz (dBm)
Noise Floor of
Analyser (dBm)
Power @ 2.43 GHz ~ 6 dB
above Noise Floor (dBm)
10
-63
-115
-109
100
-51
-102
-95
1000
-42
-92
-86
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
63
ERA Report 2006-0713 (Issue 2)
From the above table, it can be seen that the lowest UWB emission level that can be measured is
–87 dBm in a 1 MHz bandwidth using RMS detection. Smaller bandwidths could be used and the
MB-OFDM UWB signal could be scaled using a 10.log10(BWCF). However, the noise floor of the
spectrum analyser would also go up by the same margin, thus giving no advantage.
Freq = 2.43 GHz
Figure 25: MB-OFDM UWB signal and its 3 bands conductively measured in a 1 MHz RBW
Figure 25 above reveals the spectrum of the Wisair MB-OFDM device as observed in a 4 GHz span
on the spectrum analyser. The plot clearly shows the three 528 MHz bands constituting a MB-OFDM
signal in the lower 3 GHz UWB band.
Next, a HP pre-amp was connected between the MB-UWB device and spectrum analyser. This
allowed the UWB signal to be amplified with a typical gain of 40 dB across the 1.5 GHz bandwidth of
the signal.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
64
ERA Report 2006-0713 (Issue 2)
HP 8449B
Pre-amp
MB-OFDM
Device
Spectrum
Analyser
Figure 26: Bandwidth limited measurements for amplified MB-OFDM UWB signals using a
spectrum analyser
Table 13 below shows the conducted results of an amplified MB-OFDMUWB signal. The results
show that the minimum UWB signal from the MB-OFDM device is 18 dB above the receiver
sensitivity (including a pre-amplifier) in a 1 MHz bandwidth. Taking into account the 40 dBi gain of
the pre-amplifier, a MB-OFDM RMS signal level of –89 dBm was measured in a 1 MHz bandwidth.
In fact a UWB signal of –101 dBm can be measured assuming that the minimum signal level from the
device is 6 dB above the measured analyser sensitivity of -107 dBm in a 1 MHz bandwidth, with the
pre-amplifier connected.
Table 13:
Conducted measurements of an amplified MB-OFDM UWB signal
RBW (kHz)
Receiver Sensitivity with
Pre-amplifier (dBm)
Minimum Observed UWB Signal
with Amplification (dBm)
10
-130
-110
100
-115
-100
1000
-107
-89
Figure 27 below reveals the spectrum of the Wisair MB-OFDM device as observed in a 5 GHz span
on the spectrum analyser. The plot clearly shows the minimum signal from the MB-OFDM device as
low as –49 dBm around 1.68 GHz. Thus giving a conducted EIRP density of -89 dBm/MHz, when
taking into account the 40 dBi gain of the HP pre-amplifier. Note, that the noise floor of the spectrum
analyser varied by +/- 1 dB when performing the measurements.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
65
ERA Report 2006-0713 (Issue 2)
Minimum
UWB Signal
Figure 27: Amplified MB-OFDM UWB signal and its 3 bands measured conductively in 1 MHz
RBW using RMS detection
“The Autonomous Interference Monitoring System (AIMS) currently being developed by Mass
Consultants Ltd. for Ofcom, has shown that it is quite possible to measure signals 2 to 3 dB above its
monitoring system sensitivity. The emphasis in that project is on the measurement of the C/(I+N) ratio
and detecting the presence of UWB in the presence of other signals. This is a different problem to
that of measuring out-of-band UWB power, but the two issues are not unrelated. Development of the
necessary algorithms is currently ongoing and it is expected that the system will be completed in
March 2007.”
The monitoring system uses an R&S FSQ26 spectrum analyser with an external wideband preamplifier, giving an overall system noise figure, at the input to Antenna Interface Unit, of between
2.3 dB and 7 dB producing system sensitivities between -110 dBm and -107 dBm in the 0.2 to 8 GHz
frequency band. Theses sensitivities are comparable to those measured in the conducted test set-up as
shown in Table 13, also using a pre-amplifier. The difference in measuring the UWB signal only 2 to
3 dB above the noise floor of the analyser using MASS Consultants monitoring system gives a 3 to
4 dB improvement compared with measuring the signal 6 dB above the noise floor. This improvement
would also apply for radiated measurements as the sensitivity of the analyser with the pre-amplifier
would be the same as in the conducted test set-up.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
66
ERA Report 2006-0713 (Issue 2)
11.3 Radiated Low Level UWB Emissions
Radiated emission measurements from the Wisair MB-OFDM device were made in an anechoic
chamber using a R&S FSU46 spectrum analyser, an optional HP 8449B pre-amplifier and an A.H
Systems Double Ridged Waveguide (DRWG) horn antenna, as shown in Figure 28.
Other antennas were considered, e.g. the EMCO 3147 log-periodic antenna and the EMCO 3160
single octave horn antenna series. The log periodic antenna gave no advantage over the DRWG horn
antenna in terms of Antenna Correction Factor (ACF), but the single octave horn antenna gave a 5 7 dB/m better performance at a distance of 1 m. However, for practical UWB measurements from 1 to
5 GHz required 3 separate single octave horn antennas. Thus, making it more difficult to perform the
measurements.
The receiving DRWG horn antenna and Wisair antenna were mounted on wooden tripods 0.8 m
above the ground and the received power from the UWB source was measured at distance of 1 m
using the spectrum analyser. The measurements were made using vertical polarisation and the
received power was recorded in a 1 MHz RBW using RMS detection on the analyser. Measurements
were also attempted at separation distance of 3 m.
Screened Room
Spectrum
Analyser
Optional Preamplifier
Anechoic Chamber
DRWG Horn
Antenna
MB-OFDM
UWB Device
d (m)
Figure 28: Set-up for radiating UWB emission measurements
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
67
ERA Report 2006-0713 (Issue 2)
Figure 29 below shows the radiated MB-OFDM UWB signal measured in a 1 MHz bandwidth using a
RMS detector. The plot shows that the UWB signal is approximately 6 dB above noise floor of the
analyser at a frequency of 2.87 GHz.
Figure 29: Radiated MB-OFDM UWB signal measured at a distance of 1 m in a 1 MHz RBW
A smaller bandwidth could be used to measure the MB-OFDM UWB signal and then scaled using a
10.log10 (BWCF). However, the noise floor of the spectrum analyser would also go up by the same
margin, thus giving no advantage (See table below).
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
68
ERA Report 2006-0713 (Issue 2)
Table 14:
Radiated measurements of a MB-OFDM UWB signal without a pre-amplifier
RBW (kHz)
Noise Floor of
Analyser (dBm)
UWB Signal @ 2.87
GHz (dBm)
10
-114
-107
100
-99
-93
1000
-91
-84
Using the expression:
EIRPRMS (dBm / MHz ) = E (dBμV / m ) + 20 log 10(d m ) − 104.8
Eq. 5
Where,
E (dBμV / m ) = PRMS (dBm ) + ACF + 107 + Lcable (dB )
Eq. 6
The minimum mean EIRP that can be measured in RBW of 1 MHz on a spectrum analyser and a
DRWG horn antenna with an ACF of 30.2 dB/m at 2.87 GHz is -50 dBm.
The signal level of the UWB device measured at 3 m was just above the noise floor of the analyser
using an RMS detector and was discarded, because the noise of the analyser was contributing to the
measured UWB signal. Also, NTIA Report 01-383 [16] suggests that the UWB signal level should be
at least 10 dB above the measurement receiver inherent noise level for the largest measurement
bandwidth.
Next, a HP pre-amplifier was inserted between the receiving DRWG horn antenna and the spectrum
analyser. This step was taken to amplify the low level emissions radiated by the Wisair device, which
has a mean EIRP density of –41.3 dBm/MHz as shown in Figure 25. This figure is based on
conducted measurement assuming an antenna gain of 0 dBi for an omni-directional antenna.
Figure 30 below shows the radiated UWB signal measured at a separation distance of 1 m in a 1 MHz
RBW, using RMS detection on the spectrum analyser via a pre-amplifier. The plot shows that UWB
signals as low as –63 dBm (6 dB above noise floor in a 1 MHz bandwidth) can be measured using the
radiated set-up shown in Figure 28.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
69
ERA Report 2006-0713 (Issue 2)
Figure 30: Radiated MB-OFDM UWB signal measured at a distance of 1 m in a 1 MHz RBW
using a HP pre-amplifier
Using the expression:
EIRPRMS (dBm / MHz ) = E (dBμV / m ) + 20 log 10(d m ) − 104.8
Eq. 7
Where,
E (dBμV / m ) = PRMS (dBm ) + ACF + 107 − G pre− amp (dBi ) + Lcable (dB )
Eq. 8
The minimum mean EIRP that can be measured in RBW of 1 MHz on a spectrum analyser with a 40
dBi gain and a DRWG horn antenna with an ACF of 30.2 dB/m at 2.7 GHz is -69 dBm. This
compares well with the theoretical calculations shown in Appendix A. Also, the minimum level
improves slightly at 2 GHz because the ACF of the measuring antenna drops down by several dB, as
shown in the table below. Note, that the sensitivity of the measuring system can be calculated by
subtracting the gain of the pre-amplifier (40 dBi) from the measured noise floor of the spectrum
analyser.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
70
ERA Report 2006-0713 (Issue 2)
Table 15:
Minimum mean UWB EIRP levels that can be measured below 3.8 GHz in a 1 MHz RBW using
a HP 8449B pre-amplifier
Frequency
(GHz)
Noise Floor
(dBm)
UWB Signal
Level (dBm)
ACF (dB/m)
Mean EIRP
(dBm/MHz)
1.6
-69
-63
24.5
-74
2.0
-70
-63
27.8
-71
2.7
-69
-63
30.2
-69
Note: 2 dB measured cable loss is included in the EIRP calculation.
Next, the HP 8449B pre-amplifier was replaced an ERA WBA3-4-06G20N preamplifier with a gain
of 28 dBi and a lower noise figure.
Figure 31: Radiated MB-OFDM UWB signal measured at a distance of 1 m in a 1 MHz RBW
using an ERA pre-amplifier
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
71
ERA Report 2006-0713 (Issue 2)
The figure above shows that a signal level of –70 dBm can be measured at 2.7 GHz using the ERA
pre-amplifier in a 1 MHz bandwidth.
Table 16:
Radiated measurements of a MB-OFDM UWB signal using an ERA pre-amplifier
RBW (kHz)
Noise Floor of
Analyser (dBm)
Signal Level @ 2.7
GHz (dBm)
EIRP (dBm/MHz)
10
-100
-93
-87
100
-87
-79
-73
1000
-79
-70
-64
From Table 16 above, it can be seen that the noise floor measured in a 1 MHz bandwidth for the ERA
pre-amplifier is 10 dB lower when compared the noise floor measured with the HP pre-amplifier (See
Table 15. Using Equations 7 and 8, the minimum EIRP that can be measured using an ERA preamplifier, for the radiated measurement set-up shown in Figure 28, at a distance of 1 m is
–64 dBm/MHz. This is 5 dB higher compared with the HP pre-amplifier results of –69 dBm/MHz at a
frequency of 2.7 GHz, as shown in Table 15. This difference is due to the HP result being 3 dB closer
to the noise floor compared with the ERA result and slightly better gain performance from the HP
amplifier.
The radiated measurement results from the DRWG horn using the HP pre-amplifier were then
compared with an EMCO 3160-2 single octave horn antenna. The maximum dimensions of the
DRWG horn and single octave horn antenna were 25 cm and 35 cm respectively. Using the Rayleigh
distance criterion and physical dimensions of each antenna, the measurements were made at 1 m and 2
m for the DRWG and single octave horn respectively.
Table 17:
Radiated measurements of a MB-OFDM UWB signal using a DRWG and single octave horn
antenna
Antenna
Distance (m)
UWB Signal @
2.7 GHz (dBm)
ACF (dB/m)
EIRP
(dBm/MHz)
DRWG
1
-61
30.0
-67
Octave
2
-61
22.6
-68
The radiated results shown in the table above reveal that the minimum UWB EIRP density that can be
measured is –67 dBm/MHz, for frequencies less than 3.8 GHz. The signals measured were 8 dB
above a noise floor of –69 dBm, in a 1 MHz bandwidth. Therefore, assuming a wanted signal level of
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
72
ERA Report 2006-0713 (Issue 2)
6 dB above the noise floor of the spectrum analyser, the minimum UWB EIRP density that can be
measured would drop to –69 dBm/MHz.
Finally, the HP pre-amplifier and ERA pre-amplifier were connected in series to see if any signal gain
over noise advantage could be obtained. The pre-amplifiers were also switched round to see if the
different noise figures from both devices had any effect to the overall noise floor observed on the
R&S ESPI7 analyser in a 1 MHz RBW. A CW input signal of –80 dBm from an Agilent E4438C
signal generator was used calculate the overall Signal-to-Noise S/N ratio of both the individual preamplifiers and them connected in series. All measurements results shown in Table 18 were made in a
1 MHz bandwidth using a span of 20 MHz.
Table 18:
Comparison of S/N ratio for pre-amplifier configurations
Pre-amplifier
Frequency
(GHz)
Noise floor N
(dBm)
Signal Level S
(dBm)
S/N (dB)
ERA
2.0
-76
-52
24
HP
2.0
-65
-41
24
ERA+HP
2.0
-37
-12
25
HP+ERA
2.0
-38
-13
25
From the table above it can be seen that there is no advantage in connecting both the ERA and HP
pre-amplifiers in series. The S/N ratios for the combined pre-amplifiers are within 1 dB of each other
when compared to the individual devices themselves.
Table 19 below summarises the practical minimum mean EIRP levels that can be measured in a
1 MHz bandwidth using RMS detection on a spectrum analyser for frequencies less than 3.8 GHz.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
73
ERA Report 2006-0713 (Issue 2)
Table 19:
Minimum mean MB-OFDM UWB EIRP levels that can be practically measured
Measurement Method
Minimum Mean EIRP
(dBm/MHz)
Conducted
-861
Conducted with a pre-amplifier
-1011
Radiated at a 1 m separation distance
-501
Radiated with a pre-amplifier at a 1 m separation distance
-74-691,2
1
UWB signal 6 dB above the noise floor of the analyser when measured in a FAR with a DRWG
horn antenna. 2 Between the 1.6 and 3.8 GHz band.
From the table above, it can bee seen that an UWB EIRP limit of –85 dBm/MHz can be achieved
using the conducted measurement set-up, by including a pre-amplifier. The standard conducted
measurement set-up falls within 1 dB of the minimum EIRP limit. However, he radiated test set-up
using a pre-amplifier falls short by 16 dB for frequencies at the top end of the 1.6 to 3.8 GHz band, as
proposed by the ECC. At low measurement frequencies at 1.6 GHz, the mean radiated UWB EIRP
density that can be measured is –74 dBm/MHz.
The conducted measurement set-up using the pre-amplifier is only possible if the antenna of the UWB
device is external and not integrated on the board of the UWB device. Attempting to measure the
UWB signal by connecting an external cable to the antenna port on the board of the device will
produce impedance mismatches and give incorrect results.
The single octave horn antenna has a 6-7 dB better performance in terms of antenna gain compared
with the Double Ridged Waveguide Horn (DRWG) horn antenna. However, the maximum dimension
of 35 cm for the single octave horn antenna, compared with 25 cm with the DRWG horn requires the
single octave horn antenna to be roughly twice the distance from the DUT. Thus, incurring an extra
6 dB of path loss compared with using a DWRG. The standard radiated test set-up alone is
insufficient to measure UWB signals with low mean EIRP densities of –85 dBm/MHz.
All in all, the radiated measurements are limited by a combination of the noise floor level of the
spectrum analyser or any other receiver, the noise figure of the pre-amplifier and the sensitivity of the
measurement antenna. In effect, the lower the measurement system sensitivity, the lower the UWB
signal level that can be measured. Smaller bandwidths could be used and the MB-OFDM signal could
be scaled using the 10.log10 BWCF. However, the noise floor of the spectrum analyser would also go
up by the same margin, thus giving no advantage.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
74
ERA Report 2006-0713 (Issue 2)
12
Conclusions
12.1 Introduction
The proposed European limits for UWB emission levels are significantly tighter than the original FCC
limits, and it has been suggested that the measurement techniques mandated by the FCC might not
have sufficient sensitivity to measure EIRP limits as low as –85 dBm/MHz between 1.6 GHz to
3.8 GHz band. ERA were asked to investigate, if practical, repeatable and accurate measurements
could be made at this low level for using conducted and radiated test measurement. The objectives of
the project were to:
1. Focus on the advantages and disadvantages for the following test facilities and recommend a
measurement environment for cost effective compliance testing.
a. Open Area Test Sites (OATS).
b. Semi Anechoic Room (SAR).
c. Fully Anechoic Room (FAR).
d. GTEM or TEM Cells.
e. Reverberation Chamber.
2. Characterise the different types of UWB signals.
3. Assess the minimum EIRP level that could be measured using standard measuring equipment
based on:
•
Environment.
•
Measurement distance.
•
Antenna type.
•
Detector type.
•
Resolution bandwidth.
12.2 Test Environment
The radiated measurements were performed in a Fully Anechoic Room. This measurement
environment was chosen over other environments due to the following reasons:
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
75
ERA Report 2006-0713 (Issue 2)
a. The measurement procedures developed by the NTIA [16] recommend installing the UWB
device in an environment where measurement system multi-path and external radio signals
are eliminated or minimised. The best possible choice is a high-performance anechoic
chamber.
b. Communication papers in Japan describe the measurement technique used to measure
emissions from UWB devices [27] also recommend measuring the emission in a high
performance anechoic chamber. Below 1000 MHz a semi-anechoic chamber may be used.
The OATS test environment was not considered because it has several deficiencies such as weather
and ambient vulnerability, ground plane mutual coupling and reflection effects as well as incomplete
angular coverage of the equipment under test radiation patterns [19]. The ambient environment can be
a problem when measurements have to be performed at low sensitivity levels in the frequency range
where intentional emitters are present such as GSM 900 & 1800 or 3G. The RF ambiance of the
OATS can add considerable uncertainty to the measurements.
Calculations have shown that the GTEM cell may have a 2 dB greater sensitivity compared with a
FAR (See Appendix A). However, the available volume is small and not entirely suitable for
equipment with attached cables. Experiments have shown that the results in the GTEM cell were
sensitive to location of the EUT within a few centimetres, giving variations of 2 dB up to 2.5 GHz and
7 dB above 2.5 GHz [22]. Also, the cross-polarisation performance is considered inferior to an
anechoic chamber or OATS. Over a limited frequency band the field level of the longitudinal mode
can exceed the level of the intended vertical field. The size of EUT is limited to approximately onethird height between the septum and floor. Finally, it is difficult to determine the measurement
uncertainty because of the cross-polarisation performance being inferior to an anechoic chamber or
OATS.
ERA Stage B Report on “Emission measurements in the range 1 to 6 GHz in Fully Anechoic Room
and Reverberation Chamber Facilities” [25] concludes that there is a reasonable correlation between
the FAR and the reverberation chamber results, further work is required to investigate the appropriate
gain factor as suggested in BS EN 61000-4-21. There are also good correlations between the
reverberation mode- stirred and mode-tuned measurements.
On average the reverberation chamber gave a reduction in field strength of 7 dB to 10 dB +/- 5 dB,
but there was clear and apparently uniform frequency dependence. This could lead to a factor, which
could be included in EN 61000-4-21 to obtain representative fields when using the reverberation
chamber, but further work is required to determine the value and precision of the gain figure to be
applied.
The measurement results in a 1 MHz RBW using a peak detector also showed that the reverberation
chamber had approximately a 20 dB better sensitivity compared with the FAR. However, EN 610004-21, states that the main disadvantage of the reverberation chamber is that the measurement system
must have a sensitivity of 20 dB lower than the actual mean to get an accurate average measurement
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
76
ERA Report 2006-0713 (Issue 2)
and intermittent signals may be artificially lowered due to insufficient sampling. Thus, effectively
eliminating any sensitivity advantage gained.
Other work has shown the EUT in the reverberation chamber to give a very good comparison with the
FAR data, but 4.5 dB higher. This may be expected, because the reverberation chamber captures the
total energy, whereas the measurements in the FAR measure emissions in a single azimuth plane [22].
Also, reverberation chambers are currently accepted in only a few standards. Utilisation of
reverberation chambers for emissions testing will require acceptance of total radiated power as a
pass/fail criterion. There is also a lowest usable frequency defined by the cavity dimensions, tuner
effectiveness, and cavity quality factor.
Directivity and polarisation effects are not measurable in a reverberation chamber. The high Quality
Factor in reverberation chambers may impose constraints on pulse testing. Depending on test
conditions, correlation of test results to other test techniques may be difficult or impractical.
12.3 UWB Signal Characterisation
Conducted measurements have shown that the peak power of a pulsed UWB varies as 20.log10
(RBW) and the RMS power varies as 10.log10 (RBW) for RBW > PRF.
Also, the results show that the peak and RMS powers vary with a trend of 10.log10 (RBW) for
dithered UWB signals, which is consistent with NTIA Report 01-383 predictions.
MB-OFDM UWB signals show a similar trend for the measured peak power when compared to the
RMS results. The results reveal that the peak results are 4 to 6 dB higher compared with the RMS
results. Both sets of results match well when compared with the AWGN results, which verifies the
10.log10 (RBW) trend. Thus, proving that the MB-OFDM UWB signal has noise-like characteristics.
12.4 Practical Measurement Limits of UWB Signals
Conducted and radiated measurements for a MB-OFDM UWB device were carried out in a FAR
using the test set up described in Sections 11.2 and 11.3. From Table 19, it can bee seen that a mean
EIRP UWB limit of –85 dBm/MHz can be achieved using the conducted measurement set-up, by
including a pre-amplifier. The standard conducted measurement falls within the minimum mean EIRP
limit by 1 dB. However, the radiated test set-up using a pre-amplifier can measure UWB EIRP
densities of -69 dBm/MHz, which falls short by 16 dB for frequencies at the top end of the 1.6 to
3.8 GHz band, as proposed by the ECC. At low measurement frequencies of 1.6 GHz, the mean
radiated UWB EIRP density that can be measured is –74 dBm/MHz.
The conducted measurement set-up using the pre-amplifier is only possible if the antenna of the UWB
device is external and not integrated on the board of the UWB device. Attempting to measure the
UWB signal by connecting an external cable to the antenna port on the board of the device will
produce impedance mismatches and give incorrect results.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
77
ERA Report 2006-0713 (Issue 2)
The radiated measurements are limited by a combination of the noise floor level of the spectrum
analyser or any other receiver, the noise figure of the pre-amplifier and the sensitivity of the
measurement antenna. In effect, the lower the measurement system sensitivity, the lower the UWB
signal level that can be measured. Smaller bandwidths could be used and the MB-OFDM UWB signal
could be scaled using a 10.log10 Bandwidth Correction Factor. However, the noise floor of the
spectrum analyser would also go up by the same margin, thus giving no advantage.
The single octave horn antenna has a 6-7 dB better performance in terms of antenna gain compared
with the Double Ridged Waveguide Horn (DRWG) horn antenna. However, the maximum dimension
of 35 cm for the single octave horn antenna, compared with 25 cm with the DRWG horn requires the
single octave horn antenna to be roughly twice the distance from the DUT. Thus, incurring an extra
6 dB of path loss compared with using a DWRG. The standard radiated test set-up alone is
insufficient to measure UWB signals with low mean EIRP densities of –85 dBm/MHz.
The Autonomous Interference Monitoring System (AIMS) currently being developed by Mass
Consultants Ltd. for Ofcom, has shown that it is quite possible to measure signals 2 to 3 dB above its
monitoring system sensitivity. The emphasis in that project is on the measurement of the C/(I+N) ratio
and detecting the presence of UWB in the presence of other signals. This is a different problem to
that of measuring out-of-band UWB power, but the two issues are not unrelated. Development of the
necessary algorithms is currently ongoing and it is expected that the system will be completed in
March 2007.”
The monitoring system uses an R&S FSQ26 spectrum analyser with an external wideband preamplifier, giving an overall system noise figure, at the input to Antenna Interface Unit, of between
2.3 dB and 7 dB producing system sensitivities between -110 dBm and -107 dBm in the 0.2 to 8 GHz
frequency band. Theses sensitivities are comparable to those measured in the conducted test set-up as
shown in Table 13, also using a pre-amplifier. The difference in measuring the UWB signal only 2 to
3 dB above the noise floor of the analyser using MASS Consultants monitoring system gives a 3 to
4 dB improvement compared with measuring the signal 6 dB above the noise floor. This improvement
would also apply for radiated measurements as the sensitivity of the analyser with the pre-amplifier
would be the same as in the conducted test set-up.
Using an average detector would improve the noise floor of the spectrum analyser further by 1-2 dB,
but the UWB signal with its noise-like characteristics would also reduce by the same amount, again
giving no advantage.
Comparison of the R&S FSU spectrum analyser with the R&S ESPI test receiver give similar noise
floor levels within 1 dB, also giving no advantage in term of measurement sensitivity.
Higher gain antennas such as parabolic dish antennas with a typical gain of 25 – 30 dBi could be used.
However, at frequencies between 1 to 5 GHz, the maximum dimension of the measurement antenna
(around 1 m) would increase the separation distance to 20 m at a frequency of 3 GHz. The additional
path loss 26 would cancel out any effective gain over using a DRWG antenna or single octave horn
antenna.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
78
ERA Report 2006-0713 (Issue 2)
Measurements in a GTEM may improve the sensitivity and radiated EIRP by 2 dB based on
calculations shown in Appendix A. The reverberation chamber may produce a 4 dB to 5 dB
improvement, based on a better sensitivity performance of 15 dB, but an average reduction in field
strength of 10 dB compared with the FAR [25].
13
References
[1]
Ofcom, “Ultra Wideband: This document consults on a position to adopt in Europe on ultra
wideband devices in 3.1-10.6 GHz”, January 2005
[2]
VODAFONE RESPONSE TO OFCOM CONSULTATION ON ULTRA WIDEBAND
(UWB), March 2005
[3]
ANSI C63.4 1992 American National Standard for Methods of Measurement of Radio-Noise
Emissions from Low-Voltage Electrical and Electronic Equipment in the Range of 9 kHz to
40 GHz
[4]
ANSI C63.7: 1992 American National Standard Guide for the Construction of Open-Area
Test Sites for Performing Radiated Emission
[5]
ANSI C63.2: 1996 American National Standard for Instrumentation -.Electromagnetic Noise
and Field strength, 10 kHz-40 GHz - Specifications
[6]
FCC, “Revision of Part 15 of the Commission’s Rules Regarding Ultra-Wideband
Transmission Systems, ET Docket No. 98-153, First Report and Order, April 2002
[7]
EN 302 066-1: 2005 ERM: Short Range Devices (SRD): Ground and Wall-probing Radar
applications. Part 1:Technical characteristics and test methods
[8]
EN 300 175-2 - Digital Enhanced Cordless Telecommunications (DECT); Common Interface
(CI); Part 2: Physical Layer (PHL)
[9]
Draft ECC/DEC/(06)04 Decision on the harmonised conditions for devices using UltraWideband (UWB) technology in the frequency band 3.1 – 4.8 GHz, March 2006
[10]
ECC Report 64 “The protection requirements of radiocommunications systems below 10.6
GHz from generic UWB applications”, February 2005, www.ero.dk
[11]
NTIA Report 03-402, “Measurements to Determine Potential Interference to Public Safety
Radio Receivers from Ultrawideband Transmission Systems”, June 2003.
[12]
J.Fisher, N.Lee, B.Firth and P.Lockhart, “Ultra-Wideband and Radar Co-existance”, 1st
EMRS DTC Technical Conference – Edinburgh 2004
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
79
ERA Report 2006-0713 (Issue 2)
[13]
ETSI Technical Report, “Electromagnetic compatibility and Radio spectrum Matters (ERM);
Short Range Devices (SRD); Technical characteristics for SRD equipment using Ultra Wide
Band technology (UWB) Part 1: Communications applications”, January 2004
[14]
Mason Communications Ltd, Final Report for Ofcom, “Value of UWB personal Area
Networking Services to the United Kingdom”, November 2004.
[15]
K.Siwiak, Debra McKeown, “Ultra-Wideband Radio Techonology”, Wiley, 2004
[16]
William.A.Kissick, “The Temporal and Spectral Characteristics of Ultrawideband Signals”,
NTIA Report 01-383, January 2001
[17]
ITS Ultra-Wideband Measurement Plan (Master Plan Task 1.2), June 14th 2001
[18]
NTIA report 01-43 “Assessment of Compatibility between Ultra Wideband Devices and
Selected Federal Systems”, January 2001
[19]
Total-radiated-power-based OATS-equivalent emissions testing in reverberation chambers
and GTEM cells. HARRINGTON, TE, 2000 IEEE Symposium on EMC, Washington, DC,
22 August 2000, p. 23-28.
[20]
BS EN 61000-4-20:2003, Electromagnetic Compatibility, Part 4-20: Testing and
measurement techniques - Emission and immunity testing in transverse electromagnetic
(TEM) wave-guides
[21]
Practical Limits for EMC Emission Testing at Frequencies Above 1GHz Final Report
(AY3601) For The Radiocommunications Agency, A J Rowell, D W Welsh & A D Papatsoris
York EMC services Ltd, UK, 2000
[22]
Alexander-Loh-Arnaut, “Preliminary results above 1 GHz comparing radiated emission
measurements on experimental EUTs in a FAR, GTEM cell and Reverberation chamber”,
CISPR/A/WG2() 05-02 October 2005
[23]
A GTEM Good Practise Guide - The Use of GTEM Cells for EMC Measurements Applying
IEC 61000-4-20”, A Nothefer, M Alexander (National Physical Laboratory) & D Bozec, A C
Marvin, L McCormack, University of York
[24]
BS EN 61000-4-21:2003, Electromagnetic Compatibility, Part 4-21: Testing and
measurement techniques - Reverberation chamber test methods
[25]
Shadi Makhalfa, Tony Maddocks Gavin Barber, Graeme Eastwood, “Emission measurements
in the range 1 to 6 GHz in Fully Anechoic Room and Reverberation Chamber Facilities”,
ERA Technology Report 2006-0352, June 2006.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
80
ERA Report 2006-0713 (Issue 2)
[26]
P S Bansal, A J Maddocks et al, “Limits and method of measurement for emissions at
frequencies above 1 GHz”, ERA Technology Report 2001-0489 (for Radiocommunications
Agency), March 2002
[27]
Measurement Techniques of Emissions from Ultra Wideband Devices" - Jun-ichi Takada,
Shinobu ISHIGAMI, Juichi Nakada, Eishin NaKagawa, Masaharu Uchino and Tetsuya Yasui,
IEICE Trans. Fundamentals Vol.E88-A, No.9, September 2005
[28]
ETSI TR 102 273-2, Improvements on Radiated Methods of Measurement (using test site)
and evaluation of the corresponding measurement uncertainties. Part 2: Anechoic chamber
[29]
ETSI TR 102 273-1, Electromagnetic compatibility and Radio spectrum Matters (ERM);
Improvement on Radiated Methods of Measurement (using test site) and evaluation of the
corresponding measurement uncertainties; Part 1: Uncertainties in the measurement of mobile
radio equipment characteristics; Sub-part 1: Introduction
[30]
The Use of GTEM Cells for EMC Measurements, A Nothefer, M Alexander (National
Physical Laboratory) & D Bozec, A C Marvin, L McCormack (York EMC services Ltd)
[31]
BS EN 55016-2-3:2004, Specification for radio disturbance and immunity measuring
apparatus and methods, Part 2-3: Methods of measurement of disturbances and immunity —
Radiated disturbance measurements
[32]
Jay.J.Ely, Gerald.L.Fuller and Timothy.W.Shaver,
Interference to Aircraft Radio”, NASA et.al
[33]
S.T.Li, J.B.McGee, P.M.McGinnis, J.H.Schukantz, Jr., “Characterisation of a High-Power,
High-Frequency, Soft-Switching Power Converter for EMC Consideration”, Technical
Document 3120, March 2001
“Ultra-wideband
Electromagnetic
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
81
ERA Report 2006-0713 (Issue 2)
APPENDIX A
Calculation of Minimum Measurable UWB EIRP
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
82
ERA Report 2006-0713 (Issue 2)
A.1 Minimum Measurable EIRP using a FAR
This section describes the theory on how to calculate the minimum EIRP that can be measured from a
UWB device radiating in a FAR. The calculations assume that a pre-amplifier and a DWRG horn
antenna are used in conjunction with the spectrum analyser (See Figure 28).
The noise floor of a spectrum analyser N can be as:
N = K B TBFa
Eq. 9
Where, KB is Boltzmans constant equal to 1.38e-23, T is the temperature in Kelvin, B is the
measurement bandwidth in Hz and Fa is the noise figure of the spectrum analyser. In log to the base
10 terms this is:
N (dBm) = −198.6 + 10 log10(T ) + 10 log10( B) + Fa (dB )
Eq. 10
Assuming that the spectrum analyser operates at room temperature of 290 K (20° C), the above
equations simplifies to:
N (dBm) = −174 + 10 log10( B) + Fa (dB )
Eq. 11
The noise figure Nf for a typical analyser is 23 to 26 dB and taking the worst-case value, the noise
floor of the analyser in a 1 MHz bandwidth would be –91 to –88 dBm.
When a pre-amplifier (HP 8449B) is used in conjuction with the spectrum analyser, the noise floor of
the measuring device should increase by same amount as the noise figure Fp of pre-amplifier.
Assuming, a wanted UWB S/N of 6 dB, the minimum UWB signal level S that can be measured on a
spectrum analyser using the HP 8449B pre-amplifier is given by:
S (dBm) = N (dBm ) + S N (dB ) + F p (dB ) − G p
Eq. 12
Where, Gp is the gain of the pre-amplifier, which in this case is 40 dB as shown on the calibration
datasheet for the HP 8449B pre-amplifier. From the HP datasheet the noise figure is 8 dB, but noise
floor of the analyser increased by 20 dB in a 1 MHz bandwidth with an RF attenuation of 0 dB.
Therefore, minimum UWB signal using a HP 8449B pre-amplifier that can be measured on a
spectrum analyser is –105 dBm, 6 dB above a minimum system sensitivity of -111 dBm.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
83
ERA Report 2006-0713 (Issue 2)
When performing radiated measurements using a DRWG antenna as used in the measurement set-up
of Figure 28, the electric field measured at the front end of the antenna can be calculated by:
E (dBμV / m) = S (dBm ) + ACF (dB / m ) + 107
Eq. 13
At a measurement separation distance of 1 m between the UWB device and the receiving DRWG
horn antenna, the ACF is typically 30 at 2.7 GHz. Thus, the electric field strength observed at the
front end of the antenna for a received signal of –101 dBm is 30 dBuV/m. Using equation 14 below
this translates to an EIRP density of –71 dBm/MHz, including a 2 dB cable.
EIRP(dBm ) = E (dBμV / m) + 20 log 10(d ) + 104.8
Eq. 14
The minimum EIRP practically measured in a 1 MHz bandwidth was –69 dBm (See Table 15).
A.2 Minimum Measurable EIRP using a GTEM Cell
Figure 32: GTEM cell equivalent electric circuit
The measured voltage V, across the terminating resistor RT is related to the electric field by the
effective height heff, of the cell acting as an antenna. The equation for the effective height of an
antenna is given as follows [33]:
Ae (R A + RT ) + ( X A + X T )
120πRT
2
heff =
2
Eq. 15
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
84
ERA Report 2006-0713 (Issue 2)
Where RA + jXA is the impedance of the GTEM cell, RT + jXT is the impedance of the termination and
Ae is the effective aperture. Knowing that RA = RT = 50 Ω and XA = XT = 0, the above equation
simplifies to:
heff = 0.73 Ae
Eq. 16
Knowing that the gain G is related to the effective aperture by:
G=
4πAe
λ2
Eq. 17
The effective height can be expressed as:
heff = 0.26λ G
Eq. 18
The generated voltage id divided between the cell’s radiating resistance and the terminating
resistance. Therefore, the measured voltage V at the output of the GTEM cell can be expressed as:
⎛ RA
V = E i heff ⎜⎜
⎝ R A + RT
⎞
⎟⎟
⎠
Eq. 19
Assuming that the UWB device radiated like a dipole, the electric field propagating in free space can
be expressed as:
Ei =
49.2 PT G
d
Eq. 20
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd
85
ERA Report 2006-0713 (Issue 2)
Where, d is the distance form the DUT to the terminating end of the GTEM cell. Substituting,
Equation 20 in to 19 and converting the measured voltage V into power P in dBm across the 50 Ω
termination gives:
⎛ 0.96 × PT G ⎞
⎟
P (dBm ) = 10 log10 ⎜⎜
2 2
⎟
d
f
GHz
⎝
⎠
Eq. 21
If the UWB device radiates with an EIRP density of –85 dBm/MHz, then the power measured by the
spectrum analyser via the HP 8449B with a gain of 40 dBi would be –74 dBm. The calculation
assumes a typical GTEM cell gain G of 10 dBi [33]. This calculated value would be below the
spectrum analyser noise floor when the pre-amplifier is used. The best theoretical UWB EIRP density
that can be measured is –73 dBm/MHz, including a 2 dB cable loss. This value assumes that the
wanted signal is 6 dB above the noise floor of –70 dBm as measured in a 1 MHz bandwidth using an
RMS detector. This value is 2 dB better than the theoretical UWB EIRP calculated in a FAR test
environment.
Ref: P:\Projects Database\Ofcom 2005 - 7G 02323\Project 3 - Vodafone\ERA Reports\Rep-6060 - 2006-0713 (Issue 2).doc
© ERA Technology Ltd