Tata Institute of Fundamental Research

Homi Bhabha Road, Navy Nagar, Colaba, Mumbai, INDIA

Colloquium on Sept 30, 2019
New Radio Networks and Emerging Trends: SDR, 5G and IoT

Prof. Dr. Ing. Habil Ulrich L. Rohde

Chairman, Synergy Microwave Corp., NJ, USA

Partner, Rohde & Schwarz, Munich, NJ
Professor, Oradea Univ., Romania
Professor, BTU Cottbus, Germany
Hon. Professor, IIT, Delhi, India
1
 Outline

Received Award at Winn Comm 2017 for pioneer work

in the field of SDR & Modern Communication systems

2
 Outline

• SDR (Software Defined Radio) & IoT

• Analog Frond End: Pros & Cons
• High Dynmaic Range Microwave Monitoring Receivers
• Image Rejection Mixer: Eliminate triple conversion
• Important Characteristics of A/D converters
• Important Characteristics of Down Converters
• Characteristics of AGC
• Carrier recovery of Data Communication
• Spectrum Analysis of Communication Receiver
• 5G
• Radio Monitoring Receiver
• Modern Radios-Futuristic Trends & Conclusions

3
 Radio Communication Standard
Problem !
• Myriad of standards exist for terrestrial communications
• Cell phone communication standards change every few years
• Satellite ground station would like to listen to multiple spacecraft, some launched in the
1970s
• Spectrum space is a precious resource
– Each frequency is “owned”
– How do we deal with new technologies like ultra wide band (UWB)?

Solution: SDR (Software Defined Radio) SDR

• Flexible radio systems that allow communication standards to migrate
• Flexible methods for reconfiguring a radio in software
• Flexible, intelligent systems that communicate via different protocols
at different times

Emerging Trend: Cognitive Radio Solution

4
 SDR & IoT
Transforming our world
Through intelligent connected platforms

Mobile IoT

Technology leadership in
system on chip, systems
design, and connectivity
Last 30 Next 30 years
years Interconnecting their
worlds
Interconnecting
people
SDR & IoT fundamentally changed how people live, work and stay connected
Courtesy: Qualcomm Technologies
5
Rohde & Schwarz: Software Defined Radios

Courtesy: Qualcomm Technologies

6
Radio Follow Moore‘s Law

7
 Why SDR ?

High Performance Radio Needed

Everywhere !
8
 Why SDR ?
First-Responder Communications Failures

11 September 2001

Hurricane

• SDR will facilitate radio interoperability

9
 Why SDR?
Deep Space Communications

• SDR allows old and new protocols

10
 Why SDR?
Spectrum space as a scarce resource

 SDR will enable spectrum reuse

11
SDR will Dynamically Access Available Spectrum
All Spectrum May Be Assigned, But…
Next Generation Radios: Enable
Technology and System Concepts for
Dynamically Access Available Spectrum

Sense
Sense
Real
Realtime,
time,Low-
Low -
Low-
power, wideband
power, wideband
monitoring
monitoring
Autonomous
…Most Spectrum Is Unused! Adapt
Adapt Dynamic
Characterize
Characterize
Transition Rapid
Rapidwaveform
waveform
Maximum Amplitudes Transition Spectrum determination
Heavy Use Heavy Use network
networkto
tonew
new determination
emission plan
Utilization
emission plan

React
React
Amplidue (dBm)

Formulate
FormulateBest
Best
Sparse Use Medium Use
Course
CourseofofAction
Action

Goal: 10 Folds increase in spectrum access

Frequency (MHz)

12
 SDR (Software Defined Radio)
• Definition:
A Software Defined Radio (SDR) is a communicaton system, where the major part of
signal processing components, typically realized in hardware are instead replaced by
digital algorithms, written in software (FPGA).

SDR
• Want to make all parameters digitally tunable
- What Parameters?
• RX/TX Frequency
• Bandwidth
• Impedance Match

• First Reported Publication (February 26-28, 1985):

Ulrich L. Rohde, Digital HF Radio: “A Sampling of Techniques, presented at the Third
International Conference on HF Communication Systems and Techniques”, London,
England, February 26-28, 1985, Classified Session (U.S Secret).
http://en.wikipedia.org/wiki/Software-defined_radio
13
 Benefits of Software Defined Radio
• Ease of design
– Reduces design-cycle time, quicker iterations
• Ease of manufacture
– Digital hardware reduces costs associated with manufacturing and
testing radios
• Multimode operation
– SR can change modes by loading appropriate software into memory
• Use of advanced signal processing techniques
– Allows implementation of new receiver structures and signal
processing techniques
• Fewer discrete components
– Digital processors can implement functions such as synchronization,
demodulation, error correction, decryption, etc.
• Flexibility to incorporate additional functionality
– Can be modified in the field to correct problems and to upgrade
14
 Benefits of SDR, Cont’d.

• Flexible/reconfigurable
– Reprogrammable units and infrastructure
• Reduced obsolescence
– Multiband/multimode
• Ubiquitous connectivity
– Different standards can co-exist
• Enhances/facilitates experimentation
• Brings analog and digital worlds together
– Full convergence of digital networks and radio science
– Networkable
– Simultaneous voice, data, and video

15
 Technologies that enable SDR
• Antennas
– Receive antennas are easier to achieve wide-band performance than transmit ones
– New fractal & plasma antennas expected in smaller size and wideband capability
• Waveforms
– Management and selection of multiple waveforms
– Cancellation carriers and pulse shaping techniques
• Analog-to-digital converters
– High ADC sampling speed
– ADC bandwidth could be digitized instantaneously
• Digital signal processing/FPGAs
– Number of transistors doubles every 18 months
– More specific purpose DSPs and FPGAs
• Batteries
– More and more power needed (need to focus on more efficient use of power)
– Fuel cell development for handhelds
• Terrain databases
– Interference prediction, environment awareness
• Cognitive science
– A key aspect will be to understand how multiple CRs work with each other

16
Block Diagram: Software Defined Radio

Antenna
RF IF Baseband

Band Pass ADC/DAC

Filter DSP

Variable Local
Frequency Oscillator
Oscillator (fixed)

17
Oscillator Phase Noise - 1

18
Oscillator Phase Noise - 2

19
Oscillator Phase Noise – 3
IC -PLL

20
Frequency Generation

21
Frequency Generation, cont’d.

22
NCO (Numerically Controlled Oscillator)

23
NCO Block Diagram
48 Bits Resolution

24
Block Diagram: Software Defined Radio
Antenna
RF Baseband
IF

ADC/DAC DSP

Local
Oscillator
(fixed)
Antenna

Baseband
RF IF

ADC/DAC DSP

25
 5G: Emerging Cellular Networks
Mobile operators have just commercialized LTE and
5G drivers few of the features that make LTE a true 4G
technology have made it into live networks. So why
is industry already discussing 5G?

• Constant user demands for higher data rates

and faster connections require a lot more
wireless network capacity, especially in dense
areas.
• The industry is expecting demand for 100x
higher peak data rate per user and 1000x more
capacity, and better cost efficiency defined these
as targets for the 5th generation of mobile
networks (5G).
• Internet of Things (IoT) provides new challenges
to be addressed. It is anticipated that millions of
devices will “talk” to each other, including
machine to machine (M2M), vehicle-to-vehicle
(V2V) or more general x-2-y use cases.
Beyond doubt there is a need to improve the understanding of potential new air interfaces at frequencies
above current cellular network technologies, from 6 GHz right up to 100 GHz, as well as advanced antenna
technologies such as massive MIMO and beam forming, very long battery lifetimes (years instead of days)
and very low response times (latency) call for another “G” in the future !!!
26
Present Technology and Planned 5G Spectrum

Table represents existing technology and spectrum for SDR frequency spectrum in
terms of waveforms
27
Radios: What is Desired? Capacity or Cost

Spectral Efficiency (Capacity) Energy Efficiency (Cost)

What is better?
In terms of capacity and cost

Advanced test equipment

bridging between radio &
virtualization

28
5G Business Case

Drive profit by reducing expenses (energy efficiency)

29
 Why 5G? Power Consumption
Cellular Network Energy Consumption (China) Radio Access Network Energy Consumption

2G GSM 3G TD-SCDMA
O&M Site
830,000 base 350,000 base Rent
stations stations 3% 21
Air-Con
%
80 GWH (96 KWH 13 GWH (37 KWH Equipment 31%
51
per BTx) per BTx) % 46% Electricity
7%
41% Tx

CAPEX OPEX

Servic Ratio
Packet Data to 4G TD-LTE Biggest CAPEX/OPEX Expense is Air Conditioning
Size Signaling
e Type (%) 800,000 base
(kB) Ratio (DSR) CMRI, “C-RAN: The Road Towards Green RAN,” Dec. 2013

Text/I
stations
60 1 1 to 3
M 16 GWH (20 KWH Example: China Mobile Network in 2013
per BTx) consumed over 15 Billion KWH
Voice 35 10
Source: IEEE Communications Magazine, Feb 2014
Photo 4 150 65 to 375

Video 1 1500

30
Energy Efficiency: Centralized Baseband Processing
Centralized Control/Processing
BS1: GSM BS2: LTE ... BS3: 5G  Centralized processing resource pool that can support 10~1000
Phy/Mac Phy/Mac Phy/Mac cells
...
RTOS RTOS RTOS Collaborative Radio
 Multi-cell joint scheduling and processing
Hypervisor Real-Time Cloud
 Target to open IT platform
General purpose processor platform  Consolidate the processing resource into a cloud
 Flexible multi-standard operation and migration

Clean System Target

 Less power consuming
 Lower OPEX
 Fast system roll-out
Virtual base station pool (Real-time cloud BBU) -15% Capital Costs

High bandwidth optical transport network -50% Operating Costs

-70% Power Consumption

Architect Equipm Air Switchi Batter Transmis

Total
ure ent Con ng y sion

Tradition 0.65 0.2 3.45

2.0 kW 0.2 kW 0.2 kW
al kW kW kW
Distributed configurable wideband RRU
Distributed configurable wideband RRU
Cloud 0.55 0.01 0.1 0.86
0.2 kW 0.0 kW
Radio kW kW kW kW

31
 Spectral Efficiency
More channels = MIMO (5G FR1) Shannon Channel Information Capacity SNR increase (log2 increase)
Linear increase Capacity (bits/second)

CWlog2 1 
Signal BW (Hz) SNR (S/N)

Massive MIMO

Larger signal BW = mmWave (5G FR2)

Linear increase

Fiber Coverage High Capacity

Mobility, Reliability Massive Throughput

FR1: Sub-6GHz FR2 mmWave: 24-90 GHz

32
Energy Efficiency: Why Massive?

Number of UEs: 1
Number of antennas = 1
120 antennas per UE

Number of BS
transmit antennas
1 120

Normalized output 1 1
𝑃 = =1 𝑃 =
𝑀 𝑃
power of antennas

Normalized output 𝑃 = 𝑃 =1 𝑃 = 𝑃 ~1/1000

power of base station
Source: IEEE Signal Processing Magazine, Jan 2013

Improve energy efficiency: beamforming

33
 How to Beam form?
Beamsteering (Phase
Principle of Beamforming & Beamsteering Sidelobe Suppression
Shift)
1.Fixed antenna
spacing d Broadside
2.Choose direction θ
3.Set phase shifts Δφ

Gain (dBi)
To far-field

Gain (dBi)
θ

φ1 φ2 φ3

φM

Antennas d
...
Phase Shifters
...
Attenuators
...

34
 Beamforming Architectures
nth TRx module
Single Transceiver + Antenna
Circulator
Probe mth antenna or
switch
Receive RF chain Rn
Measurement
equipment p Cm g
m

Transmit RF chain Tn

From Analog .... ... To Digital ... To Hybrid

Analog Beamforming (ABF) Digital Beamforming (DBF): 5G FR1 Hybrid Beamforming (HBF): 5G FR2

35
Massive MIMO = Complex Base Stations

36
 Which is the Optimal Network? 12 x 107

1 TRx + DBF
10 8 TRx + DBF

2G & 4G Capacity 5G optimal capacity for DBF

8
1x 106 massive MIMO
(8 TRx per user)
6
FR1 SE: 30 bps/Hz
1 x 105 EE: 6 x 107 bits/Joule

EE (bits/J)
1 x 103
SE (bps/Hz)
10 20 30 40 50
EE (bits/J)

4 x 107
GSM
LTE Macro
LTE Small Cell 5G optimal capacity for
3 hybrid beamforming
FR2 mmWave (50 antennas)
1 TRx + ABF (8 TRx per array)
SE (bps/Hz) SE: 10 bps/Hz
2 4 6 7 8 8 TRx + ABF EE: 3.5 x 107 bits/Joule

EE (bits/J)
2G Optimal Capacity: 2-4 bps/Hz

4G Optimal Capacity: 8-10 bps/Hz

Sources: IEEE Communications Magazine, Feb 2014 & Jan 2015 SE (bps/Hz)
5 10 15 20

37
5G Devices: New Measurement Paradigms

Active Antenna System (Massive MIMO

or mmWave)

Traditional

Dig I/Q

64 - 128 Antennas 8 - 128 RF Transceivers FPGA + Fiber TRx 5G

38
Basic Over-The-Air (OTA) Test Setup
Passive measurements Active measurements OTA test solutions

2D/3D antenna characterization RF transceiver characterization

System and System and

AUT control AUT control

R&S®AMS32 OTA performance R&S®AMS32 OTA performance

measurement software measurement software
R&S®CONTEST R&S®CONTEST
R&S®PWC200 plane wave converter

R&S®SMW200A vector signal generator

R&S®ATS800R CATR rack
based antenna test system

R&S®ZNA vector network analyzer

R&S®FSW signal and spectrum analyzer R&S®ATS1800C CATR conformance chamber system

39
5G & SDR Inspired Monitoring Receiver
R&S ESMD Wide Band Monitoring Receiver

High Dynamic Range !!!

40
 Monitoring Receivers

High Dynmaic Range Microwave Monitoring Receivers

• Searching for faults in professional radio networks
• Comprehensive spectrum analysis
• Monitoring of user-specific radio services
• Monitoring on behalf of regulating authorities
• Handoff receivers, i.e. parallel demodulation of narrowband
signals and simultaneous broadband spectrum scanning=High
Dynamic Range
• Critical Parameters: Noise Figure, IP2, IP3, and instantaneous
dynamic range

Best solution: Software Defined Radio

41
 Typical Microwave Receiver

Principal Arrangment for Typical Microwave Receivers

The analog front end is downconverting the RF signals into an IF range

<200MHz

42
 Microwave Receiver, Cont‘d.
Principal Arrangment for Typical Microwave Receivers

The digital front end consists of an Analog to Digital converter and a digital
down- converter to reduce the sample rate down to the bandwidth needed by
the application. Sampling rate of AD converters are rising up to 250Msps with
resolutions of 14 or 16 bits.
43
 Microwave Receiver, Cont‘d.

Principal Arrangment for Typical Microwave Receivers

The baseband processing takes over the base band filtering, AGC,
demodulation, and the signal regeneration..

44
 Typical Analog Front End
Possible Drawbacks on the Analog Front End

• Wide band microwave receivers need tripple conversion to prevent image reception
• Several expensive and switchable filters are required for pre- and IF-selection
• Intermodulation and Oscillator Phase Noise are the main issues
• Low noise and high dynamic range are contradictionary

45
 Image Rejection Mixer
Solution to eliminate tripple Conversion

LP 90°

cos(ωot)

Rx in LO + IF out (LSB)

sin(ωot)

LP

Image Rejection Mixer

• An analog Image Rejection Mixer is capable to attenuate the Image by 30...40dB

• Criterions for the image attenuation are amplitude and phase errors in both branches
• The most critical element is the 90° phase shifter, mainly for wide band IF
• The SDR technology allows to move the phase shifter from analog into the digital
part, where it can be realized nearly ideal by means of a Hilbert Transformer

46
 Image Rejection Mixer
Solution with a distributed Image Rejection Mixer

high speed
AD-converters

BPF Decimation BPF 90°

Rx in
0°
LO +/- Baseband
90°

BPF Decimation BPF

cos sin
IF-frequency
70...200MHz NCO

• The preselector filters may be wider, as they are no longer used for image rejection
• The digital parts, following the AD converter, can be realized in a FPGA
• In a wide band receiver, the LO can be tuned in steps from up to 10MHz which is
simplifying the PLL loop filter design. The fine tuning will be done by the NCO
• The image rejection can be further improved by calibration algorithms in the digital
part to values up to 80dB

47
 Down Converters
Digital Down Converter

The digital down converter includes:

• a numerically oscillator (NCO)
• a complex IQ-mixer to convert the IF down to approx. 0Hz (zero-IF)
• several decimation filter stages for reducing the sampling rate
• final lowpass FIR-filters (Finite Impulse Response)
48
 Down Converters, Cont’d.
Digital Down Converter

N
sin(   M  f
A [dB]
20 |H(f)| =
0
sin(   f / R)
-20 R: decimation factor
-40 B
N: filter order
(sections)
-60
M = 1 or 2
-80
fs: input sample
-100 rate
0 1 2 3 4
f/fs B = fs / R

CIC-Filter with R = 16, N = 5, M = 1 (CIC: Cascaded Integrator Comb)

49
 Automatic Gain Control
Automatic Gain Control

The broadband AGC serves to protect the AD converter from overvoltages. The
RF-AGC can be used to set the receiver sensitivity just below the external noise.
The digital processing part is free from distortions, therefore the final AGC can be
placed near the analog output.

50
 Automatic Gain Control, Cont’d.
Automatic Gain Control

The main AGC control is realized near the of the signal processing chain as a feed
forward control.

51
 Typical Multi-Channel Receiver
Multichannel Receivers

N - channel Receiver with N analog front ends

52
 Receiver, Cont‘d.
Multichannel Receivers

N - channel Receiver with only one analog front end and N digital down
converters. The channel frequencies must be allocated inside the preselector
passband.
53
 Receiver, Cont‘d.
Multichannel Receivers

“Polyphase Filters” is often

incorrectly taken to mean some
special kind of filter…
instead, it is merely a special
structure that is handy when
using filters in multi-rate
settings.

Polyphase is a way of doing

sampling-rate conversion that
leads to very efficient
implementations.

If all channels are equally spaced, then a Polyphase Filter bank can replace
the multiple channels in the downconverter
54
Typical Characteristics of Sampled Systems

Important Characteristics of sampled Systems

A(f)

1. zone 2. zone 3. zone

-fs -fs/2 0 fs/2 fs 3fs/2

f

The Sampling Theorem (Nyquist / Shannon)

• A bandlimited signal can be reconstructed, when B < fs/2

• Due to aliasing, replicas in all Nyquist zones will occur
• The aliasing effect can be used to sample a bandlimited signal B in a higher Nyquist
zone (bandpass- or undersampling)
B = (n – 1) ∙ fs/2 ... n ∙ fs/2 whereas n is the zone (1,2, …)

55
 Typical Characteristics of AD Converters
Characteristics of AD Converters
E NO B

18

16

14

12

10

6
6 8 10 12 14 16 18
bits

ENOB: the Effective usable Number Of Bits

SNReff = 1.76dB + ENOB ∙ 6.02dB

(measured in B = fs/2)
56
Characteristics of AD Converters, Cont‘d.
Characteristics of AD Converters
S N R [d B ]
90

85
fin = 10MHz
80
fin = 30MHz
75

70
fin = 100MHz
65

60
fin = 300MHz
55

50
0 0. 2 0.4 0 .6 0 .8 1
R M S Jitt er [p s ]

Degradation of SNR by clock jitter

very important when applying undersampling!

57
 Characteristics of AD Converters
Important Characteristics of AD converters
IMD measured on R&S EM510 without Dithering
Pout [dBm]

-20
1. Order
-40
IMD3

-60
IMD5

-80 IMD7

-100 IMD9

Noise
-120

-140
-100 -80 -60 -40 -20

Pin [dBm]

Higher order intermodulation products as a function of the input signal. The known
relationship of n∙dB/dB (n = order of IM) can not be applied. Therefore an Intercept
point cannot be calculated. In practice, the IM is measured with two tones on -7dBm
58
 Characteristics of AD Converters
Important Characteristics of AD converters

IMD measured on R&S EM510, with Dithering

Pout [dBm]

-20
1. Order
-40
IMD3
-60
IMD5
-80
IMD7

-100
IMD9

-120 Noise

-140
-100 -80 -60 -40 -20

Pin [dBm]
Applying dithering noise has the effect, that the discontinuities are no longer periodic
and therefore the spuriies are reduced.
59
 Characteristics of AD Converters
Important Characteristics of AD converters

Altrernative Methode for IM measuring The Noise Power Ratio

Pn
0dB

In Out
NPR
DUT

-100dB f

fs/2

The NPR methode reflects the true impact of intermodulation from any order
60
 Characteristics of AD Converters
Important Characteristics of AD converters

Theoretical NPR for 10, 12,14 and 16bit AD converter

61
 Carrier Recovery
Carrier Recovery for Data Communication

Example for the carrier synchronisation for a QPSK modulated carrier.

62
 Data Clock Extraction
Data clock extraction for Data Communication

Example for a Timing Error Detector for a QPSK modulated signal according to
Gardner.
63
 Typical Architecture of Communication Receiver
Filter portion of the front-end of the receiver

64
Typical Down-Conversion Architecture

Down-converter of the receiver

65
Typical Down-Conversion Architecture

Down-converter of the receiver

66
 Spectrum Analysis in Receiver
Spectrum Analysis in Communication Receivers

The actual usable bandwidth is reduced by a factor k compared with the sampling rate
fs:
Beff = fs / k
In this exemple k = 1.28
67
 Spectrum Analysis in Receiver, Cont‘d.
Spectrum Analysis in Communication Receivers

Blackman Window

An fs = 12.8Msps allows to process

6250 FFTs per second

Due to the applied window function,

the capability to detect short pulses
at both ends of the window is
reduced
Solution: overlapping FFTs

68
 Spectrum Analysis in Receiver, Cont‘d.
Spectrum Analysis in Communication Receivers

Computing Power for overlapping 2048 bins FFT and fs = 12.8Msps:

≈ 2GFLOPs (Floating Point Operations)

69
 Spectrum Analysis in Receiver, Cont‘d.
Spectrum Analysis in Communication Receivers

Panorama Function with N consecutive FFT slices

for any bandwidth, but lacks in time resolution

70
 Widebnad Monitoring Receiver
Figure 1: R&S ESMD Wide Band Monitoring Receiver

R&S ESMD

High Dynamic Range

71
 Spectrum Analysis in Receiver, Cont‘d.
Spectrum Analysis in Communication Receivers

Narrowband Analysis

Wideband Analysis

72
 Spectrum Analysis in Communication Receiver
Proof of Available Dynamic Range

Aircraft Radio Communication Receiver can be monitored abd demodulated in

the presence of strong FM Radio signal
73
 Spectrum Analysis in Communication Receiver
Multichannel (4) Operation

All 4-Channels can be anlayzed

74
 Antenna for Communication Receiver
Typical Antennas for high dynmaic range Communications Receivers

Input level up to 0 dBm !

75
 “Measure what is measurable,
and make measurable what is not so!”
Galileo Galilei

76
References
Ulrich L. Rohde “radio house”

77
Thank You

78