Modeling and Characterization Challenges of

Modeling and Characterization Challenges of Heterogeneous Network Devices
(Invited paper)
Vijay Nair, Jerzy Kolinski and Vittal Kini
System Technology Labs, Corporate Technology Group
Intel Corporation, Hillsboro, OR, 97124 USA
Vijay.k.nair@intel.com
ABSTRACT
Technological advances in architecture design and semiconductor ICs are well underway to realize small
form factor multi-radio systems. Optimizing the functionality and performance of radios in small form factor
platforms requires the system architect to give special attention to many aspects of the design. System-level
modeling and accurate RF characterization of multiple radios are needed to fully understand the complex RF
interactions and to develop a consistent design methodology. RF interference mitigation and noise cancellation
techniques for multi-radio platform will have to be devised and incorporated for the proper operation of these
heterogeneous wireless systems. This paper will review the challenges in the design and modeling of multi-radio
communication systems. RF characterization approaches of mixed network handset will also be discussed. As a case
study the impact of platform noise and radio interference on GPS receiver performance, and Bluetooth -WLAN
radio co-existence in a small form factor proof concept device will be illustrated.
INTRODUCTION
It is becoming more and more evident that tomorrow’s network environment will be extremely
heterogeneous. Future communication devices should support key usage models such as simultaneous voice and
data transmissions, multi-media applications that scale across networks, seamless handoff, location-based services
etc. [1]. Handheld devices that operate under multiple wireless protocols are being developed to demonstrate
seamless transfer, interoperability, and radio coexistence of multiple wireless network devices. Such a
communication device necessitates high-level integration of digital, analog and RF circuitries within extremely
small size. Optimizing the functionality and performance of radios in small form-factor platforms requires the
system architect to give special attention to many aspects of the design. These include mechanical form-factor
design, tuning of multiple ICs for appropriate radio bandwidths, noise isolation, antenna characteristics, RF
interference and spatial proximity.
The characterization of interoperability and co-existence of multiple radios integrated on single platform are
very critical for successful implementation in a heterogeneous network. The usage model, wireless system standards
and IC integration complexity are important parameters to consider for multi-radio system design. System-level
modeling and accurate RF characterization of multiple radios are needed to fully understand the complex RF
interactions and to develop a consistent design methodology. RF interferences due to the placement of multiple
radiating elements in a small form factor device will have to be thoroughly investigated. Moreover, the convergence
of communication and computing devices will exacerbate these problems due to the presence of the high speed
digital circuits in close proximity to the high frequency RF circuits. This convergence would require high
performance computing capability, now available in laptop, to be made available in future small form factor
platform operating in multiple wireless protocols. Wi-Max radio, for example, will have very high date rate
transmission capability over a range of frequency bands that overlap other wireless system such as WiFi and UWB.
It will also have the capabilities of voice centric systems such as GSM, WCDMA etc. UWB radios have a wide
band of operation, must operate at very low transmit power and should avoid interference to and from radios
operating under WLAN and Bluetooth protocols. RF interference mitigation and noise cancellation techniques for
multi-radio platform will have to be devised and incorporated for the proper operation of these heterogeneous
wireless systems. Simulation tools capable of developing deterministic models of communication devices are
essential in the development of low cost wireless system.
Multi-Radio Platform
Fig. 1a shows the block diagram of a multi-radio platform designed and characterized to study the important
issues related to operation of small form factor devices. Fig. 1b shows the fabricated proof-of-concept (POC)
device. The application processor (Intel Bulverde processor) interfaces with the baseband processor (Intel Hermon
processor) which in turn is connected to the WCDMA module via an MSL bus. The WLAN module, GPS receiver
and Bluetooth module are directly connected to the application processor through standard interfaces.
Fig. 1a Multi-radio Platform Architecture
Fig. 1b Proof of Concept Device
W CDMA
GPS
B aseband
BT
P ro c e s s o r
W i-F i
Side 1
Side 2
Fig. 2. PCB layout of the POC device showing the location of radio modules
Fig. 2 shows the PCB layout of the multi-radio platform. Radio module and processor chipsets are clearly
identified. Special care was taken in the placement of components and radio modules to minimize the noise
propagation. Appropriate grounding techniques and shielding was used to minimize the conductive and radiative
noises.
Modeling and Characterization Challenges
Modeling and characterization of a highly integrated device like this is rather complex and considerations
should be given to several factors. Few of the important steps one should follow in the design of a heterogeneous
network platform are: definition of features and requirements, device electrical architecture and implementation,
functional validation, power characterization and battery sizing, industrial design, mechanical design and concept
validation. System level power management is another critical factor in increasing the battery life time. Fig. 3a
shows power consumption of different modules when MPEG4 runs at 400 MHz. Total system power consumption
was 2.4 watts. Fig 3b shows the power reduction achievable when the CPU is run at three different states (400 MHz
at 1.5V, 200MHz at 1.5V and 200 MHz at 1.25 V) resulting in CPU power saving of 46%. Also, from a platform
perspective the system power savings of up to 16% was achieved resulting in prolonged battery life [2]. Some of the
power saving approaches that could be incorporated includes processors with voltage and frequency scaling,
components with low active and minimal standby power, turning off of power planes as appropriate and
development of power policies to optimize overall battery performance.
System Power
MPEG4 MIB
3000
CPU core
320
513.4
2500
CPU I/O
93.4
109
LCD logic
2000
400@1.5
LCD BL
Audio controller
200@1.5
1500
200@1.25
Audio amp
199.4
Flash
1000
SDRAM
140.3
661
500
other 3.3V
251
44
other 5 V
Total power
14.7 91.6
0
CPU CPU I/O LCD BL LCD
core
logic
Fig. 3a Power consumption of various platform components
Audio
cntr
Audio
amp
Flash SDRAM other 5V other
3.3V
Total
Power
Fig. 3b CPU power control impact
In a multi-radio platform the noise mitigation and interference minimizations are very critical. Interference
tolerances from several concurrently operating radios are to be determined and mitigation techniques have to be
developed. Multi-radios in a single platform also require innovation in antenna technology. Innovative antenna
concepts like reconfigurable antenna, MIMO antenna have to be developed for small form factor devices.
Miniaturized filter technologies employing MEMS switches and bulk acoustic wave technology have to be
incorporated for filter miniaturization. Some of the innovative antenna concepts and receiver technologies will be
presented in another paper at this symposium [3]. The system level complexity of heterogeneous platform is
substantially higher due the increase in functional circuit modules, interfaces, power consumption and high speed
data transfer. Methods to handle the increasing complexity of the system include deterministic modeling,
standardized interfaces, system level power management, and flexible radio platforms. To achieve the best
performance in handheld devices adaptive solutions such as dynamic voltage and frequency scaling and
reconfigurable circuits and multi-band antennas have been employed [4].
Multi-radio noise interference can be categorized into three areas: a) platform-to-radio noise, b) radio-toradio noise, and c) radio-to-platform noise. The sources of noise and mitigation techniques in each of these
categories are different and need to be addressed individually. Platform-to-radio noise sources include the power
plane radiation, system clock harmonics, packaging etc. The mitigation techniques include frequency clock shifting,
ground and power plane separation, component placement, mechanical case design, RF shielding and software
mitigation. The radio-to-radio interference may originate due to the proximity of components, spurious pickup by
broadband antennas, transmit power leakage into receive band and isolation. The mitigation technique in this case
is comprised of component and antenna placement, RF shielding and new antenna technology. It is also possible to
have leakage from RF circuits affecting the other modules in the platform. Placement of the components and the
shielding can minimize most of the problem associated with this type of noise.
GPS receiver desensitization exhibited by the POC device is shown in fig. 4a. GPS sensitivity decreased
about 3 dB when the Bulverde processor was turned on. When the baseband processor was also turned on, an
additional 2 dB of desensitization occurred. The significant deterioration of the GPS receiver sensitivity was
observed when WCDMA was turned on. This deterioration is partially due to the physical proximity of the
WCDMA to the GPS receiver on the POC device. This graph shows that component placement and proper isolation
and shielding are important for small form factor multi-radios.
In fig. 4b the co-existence characterization of Bluetooth and wireless LAN is described. The Bluetooth and
WLAN radios were independently tested to confirm that they work with in specification. By interconnecting the
radios in different configurations the effect of coupled noise and radioactive noise were independently
characterized. When both radios are turned on simultaneously the packet error rate of Bluetooth increased
significantly.
Fig. 4a GPS Desensitization due to Interference
Fig. 4b Bluetooth and WLAN Interference
Acknowledgement
Authors would like to thank the members of the multi-radio research project team for their assistance in the
characterization of the proof-of-concept radio platform.
Reference
[1] V. Nair et al., “Universal Communicator: Next Generation Wireless Handset, in “Annual Review of
Communications”, Volume 57, International Electronic Consortium; Nov. 2004.
[2] J. Kolinski, “Design Challenges for Personal Mobile Clients,” Intel Development Forum Conference, Aug. 2005
[3] A. Fathy, S. El-Ghazaly, H. Chunna, S. Yang and V. Nair., “Reconfigurable Antennas and RF Front ends for
Wireless Receivers,” to be presented at this conference.
[4] P. Alinikula, “Multiradio Yields Challenge for Mobile Phones,” Microwave Journal, July 2005
Download PDF