Cook Chapter6

Download as doc, pdf, or txt
Download as doc, pdf, or txt
You are on page 1of 14

Low Energy Wireless Communication

I. Intro/Abstract
In this chapter, we will explore the energetic requirements of RF wireless
communication from both a theoretical and practical standpoint. We focus on energy per
transferred bit rather than continuous power consumption because it is more closely tied
to the battery life of a wireless device. We begin with a look at the fundamental lower
limit on energy per received bit imposed by the celebrated channel capacity theorem set
forth by Claude Shannon. Based on this lower bound, we derive an energy efficiency
metric for evaluating practical RF systems. By examining power-performance tradeoffs
in RF system design, we begin to understand why and by how much will practical
systems exceed this fundamental energy bound. From the discussion of system tradeoffs
emerge a handful of low energy design techniques allowing systems to move closer to the
fundamental energy bound. Finally, we present theory and measurements of a low energy
2.4GHz transceiver implemented in a 130nm RF CMOS process and discuss its energy
saving architecture.
II. Fundamental Energy Requirements of Wireless Communication
Consider the task of properly detecting a signal with information rate R (in bits
per second), and with continuous power P0. The energy per bit in the signal is simply:
P
Eb 0
R
In this section, we use Shannons channel capacity theorem to determine the
minimum value of Eb that will allow successful detection of the signal and relate this to
other important system parameters. Shannons theorem establishes an upper bound on R
for communication over a noisy channel. This bound is called the channel capacity C
in bits per second.
C B log 2 1 SNR
B is the signal bandwidth and SNR is the ratio of signal power to noise power. If
we assume the signal is corrupted by additive white Gaussian noise (AWGN), then may
be rewritten:

P
E R
C B log 2 1 0 B log 2 1 b
N 0 B
N0 B

N0 is the noise power spectral density in Watts/Hz. P0 is the signal power at the
input of the receiver. If the channel is thermal noise limited, then N0 is equal to the
product kT, where T is temperature and k is Boltzmanns constant.
The ratio Eb/N0 is referred to as the SNR-per-bit and the ratio R/B is a measure of
spectral efficiency in bps/Hz. Both quantities are important metrics for comparing digital
modulation schemes. It is important to distinguish between SNR and Eb/N0. SNR is a
ratio of powers, while Eb/N0 is a ratio of energies. For the purposes of evaluating a given
schemes energy per bit performance, Eb/N0 is more meaningful than SNR. For instance,
if scheme A requires ten times greater Eb/N0 for demodulation than scheme B, then
scheme A will require ten times more energy to deliver a given data payload than B.

From , the capacity of a Gaussian channel increases logarithmically with signal


power P0. A cursory glance at would suggest that C increases linearly with B, but the
capacity-bandwidth relationship is actually more subtle due to the dependence of SNR on
B. It turns out that C does increase monotonically with B, but only approaches an
asymptotic value. Thus, for a given signal power P0 and noise power density N0, the
channel capacity reaches its maximum value as B approaches infinity.

P
P
P
Cmax lim B log 2 1 0
log 2 e 0 1.44 0

B
N 0 B
N0
N0

() and ()offer two different perspectives on Shannons theorem. In (), the channel
capacity is plotted versus signal bandwidth while P0 and N0 are held constant and in (),
the maximum spectral efficiency (i.e. when R = C) is plotted against Eb/N0 [1]. The
minimum achievable Eb/N0 follows from by setting the information rate (R) equal to Cmax.
E
P0
min b
ln 2 1.6 dB
N 0 Cmax
N 0
This powerful result tells us that error-free communication can be achieved so
long as the noise power density is no more than 1.6 dB greater than the energy per bit in
the signal. In a thermal noise limited channel (i.e. N0 = kT), the lower limit for Minimum
Detectable Signal energy per bit (Eb-MDS) at the receiver input becomes:
Joules
min Eb MDS kT ln 2 3 10
21
bit
Unfortunately, the theorem does not describe any modulation scheme that
reaches the limit, and most popular schemes require far greater Eb/N0 than -1.6 dB. For a
given modulation scheme (i.e. binary-PSK, OOK, etc.), the spectral efficiency R/B and
minimum Eb/N0 required for demodulation, call it (Eb/N0)min, are fixed values, independent
of transmission rate. The R/B and (Eb/N0)min values of the system may change, however,
if coding is applied to the modulation.
Spread spectrum systems employ pseudo-noise (PN) codes that reduce R/B by
increasing signal bandwidth (thus achieving processing gain), sometimes by several
orders of magnitude, enabling reliable communication with SNR (power, not energy!)
well below -1.6 dB. However, PN codes do not bring a system closer to the Eb/N0 limit
from , because the reduction in required SNR is compensated by the requirement of
sending many chips per bit, so that overall energy per bit actually remains constant. The
purpose of PN codes is to spread the signal over a wider bandwidth, which is useful for:
mitigation of multi-path fading, improved localization accuracy (i.e. GPS), multiple user
access, interference avoidance, and more [1, 2]. Error correcting codes, on the other
hand, can offer substantial reduction of energy per bit at the expense of system latency
and computational power overhead.
a. Theoretical System Energy Limits
To this point, we have only considered the energy per bit at the input of a receiver.
The goal is to find a lower bound on energy consumed by the system (including receiver
and transmitter) per bit (Eb-Sys):
P PRX
Eb Sys TX
R

PTX and PRX are the power consumed by the transmitter and receiver, respectively.
In the best possible case, with a 100% efficient transmitter and zero power receiver, all
the energy consumed by the system would go into the transmitted signal. Therefore, the
fundamental lower bounds on Eb-Sys and transmitted energy per bit (Eb-TX) are the same.
min Eb Sys min Eb TX
To find the lower bound on Eb-Sys, we now consider the minimum transmitted
energy per bit (Eb-TX). Eb-TX must exceed Eb-MDS to compensate for attenuation of the signal
as it propagates from transmitter to receiver, or path loss. Path loss for a given link is a
function of the link distance, the frequency of the signal, the environment through which
the signal is propagating, and other variables. Accurate modeling of path loss is beyond
the scope of this chapter, but a review of some popular models is offered in [3]. The ratio
by which Eb-TX exceeds Eb-MDS is known as link margin (M) and is usually expressed in dB.
For a reliable link, the system must have more link margin than path loss. In a thermal
noise limited channel, the fundamental lower bound on Eb-TX, and thus Eb-Sys, required to
achieve a link margin M is:
min Eb Sys min EbTX M kT ln 2
To achieve link margin M while only consuming MkTln2 Joules per bit, a system
must meet the following criteria:
- the receiver adds no noise
- the modulation scheme achieves the Shannon limit of -1.6dB for Eb/N0
- the transmitter is 100% efficient
- the receiver consumes zero energy per bit
Clearly, such a system is impossible to design. In real systems, transmitters are
far from 100% efficient, the modulation scheme requires more Eb/N0 than the limit, and
the receivers are noisy and may consume a large portion of the total system energy. It is
not uncommon for a system, especially a low-energy system, to consume 10,000 times
more energy per bit than this lower limit. For instance, radios targeting sensor network
applications have reported link margin of 88-120dB [4-12], resulting in a theoretical
minimum energy per bit of 1.9-3000 pJ, but the actual energy consumed by these
systems per bit ranges from about 4.4-1320 nJ.
Since the lower bound on Eb-Sys scales with M, and M may vary over several orders
of magnitude from system to system, a simple comparison of Eb-Sys is not really fair. To
let us compare apples to apples, we define an energy efficiency figure of merit for
communication systems with an ideal value of 1:
ideal energy/bit M kT ln 2

actual energy/bit
Eb Sys
The values for the sensor network radios previously mentioned are shown in
tableXX. We have mentioned several factors contributing to low energy efficiency in
wireless systems. Now our goal is to capture the relative impact of said factors by
incorporating them into an expression for . We begin by redefining link margin.
E
Eb TX
M b TX
Eb MDS F kT
Eb N 0 min

(Eb/N0)min is the minimum SNR-per-bit required for demodulation and F is called


the receiver noise factor. F is a non-ideality factor (F 1) for the noise performance of a
receiver and is discussed in greater detail in section III. In the ideal case, F = 1 and Eb/N0
= ln2. Equation allows us to express in a much more intuitive form.
E 1
ln 2
b TX
E F E N

b
0 min
b Sys

Each of the three terms in may assume values from 0 to 1 and has an ideal value
of 1. The first term tells us what portion of the total energy consumed by the overall
system gets radiated as RF signal energy in the transmitter. The second term describes
how much the link margin is degraded due to noise added by the receiver. The third term
quantifies the non-ideality of the systems modulation/demodulation strategy as
compared to the minimum achievable Eb/N0 from .
Wireless systems with very high output power tend to have higher because
transmitter overhead power and receiver power do not scale up with transmitted power; a
larger proportion of the overall power budget will burned in the PA. This is evident in
figXX where the 2 highest values come from the systems with highest output power.
For this reason, it is most useful to compare for systems with similar values for Eb-sys.
Equation provides a good starting point for further exploration of low energy
system design, but it is not a perfect metric and there are a few caveats attached with its
use. First of all, we have not considered dynamic effects such as the startup energy
spent as the transceiver tunes to the proper frequency. Nor have we included network
synchronization or the overhead bits due to training sequences, packet addressing,
encryption, etc. Rather than attempt to capture all the initialization effects that lead to
radios being on with no useful data flowing, we have narrowed our scope by assuming
the transmitter and receiver are already time synchronized and their typical data payload
per transmission is large enough that startup energy is negligible. At this point, we shift
our focus to design of low-energy wireless communication systems and discuss
techniques that can improve .
III. Low Energy Transceiver Design
In the discussion that follows, we examine the impact of modulation scheme on
system energy consumption and transceiver architecture and then discuss general design
techniques for boosting transmitter efficiency and building low noise, low power
receivers.
A. Modulation Scheme
Modulation scheme directly impacts a communication systems bandwidth
efficiency (R/B) and minimum achievable energy per bit (Eb/N0). A reasonable question
to ask is: which has the potential for lowest energy per bit, a complex modulation
scheme that packs many bits of data into each signal transition, or a simple binary
scheme? The answer is not obvious because there is a tradeoff; more complex schemes
achieve higher information rates but typically also require higher SNR to demodulate.
() provides a comparison of several popular modulation schemes with respect to
the Shannon limit, plotting R/B versus the Eb/N0 required for reliable demodulation. If
system link margin is held constant, then the best modulation strategy will largely be

determined which resource is more precious, bandwidth or energy. Schemes with lower
Eb/N0 will deliver more data for a fixed amount of energy, while those with higher R/B
will deliver highest transmission rate for a fixed amount of bandwidth.
As an example, the 802.11g standard employs 64-QAM (OFDM on 48 subcarriers) to achieve 54Mbps in the crowded 2.4GHz ISM band while only occupying
about 11MHz of bandwidth. In the case of 64-QAM, high bandwidth efficiency comes at
the cost of poor energy efficiency as evidenced by its high Eb/N0 requirement. On the
other hand, 802.11g specifies a 6Mbps mode which uses BPSK (OFDM on 48 subcarriers) also occupying 11MHz and having the same coding rate as the 54Mbps mode.
Using BPSK, the data rate only decreases by a factor of 9 but the 802.11 spec requires a
60X receiver sensitivity improvement over the 54Mbps mode, owing to the lower
(Eb/N0)min of BPSK versus 64-QAM. provides sensitivity, link margin, and data from
an 802.11G chipset using these modulation methods.
Since we are most concerned with minimizing energy consumption, we would
tend to favor a modulation scheme with as small an (Eb/N0)min requirement as possible.
Furthermore, given the relatively low data throughput and short range of the systems of
interest, some sacrifice of bandwidth efficiency is justifiable if it affords an energy
benefit. In theory, the lowest energy uncoded modulation scheme would be M-ary FSK
with M approaching infinity [1]. This strategy is not popular because (Eb/N0)min only
decreases incrementally at large M, while the occupied bandwidth and system complexity
grow steadily.
In practical systems targeting low energy, 2,4-PSK, 2-FSK, and OOK are the most
common modulation methods representing a compromise between energy efficiency
and simplicity of implementation. Radios designed for sensor network applications have
used either PSK [8, 12], binary FSK [4, 6, 7, 9-11], or OOK [4, 5]. The original 802.15.1
standard (Bluetooth) uses Gaussian 2-FSK and the 802.15.4 standard uses a form of
QPSK (i.e. 4-PSK) that can be implemented as 2-FSK. Newer versions of Bluetooth
adopt 8-DPSK as the modulation technique to extend data rate to 3Mbps, but the energy
efficiency of these systems will most likely drop somewhat (Eb/N0)min for 8-DPSK is
substantially higher than the original GFSK format.
2. System Architecture Considerations
When choosing a modulation scheme for low-energy, (Eb/N0)min does not tell the
complete story. Even if (Eb/N0)min is low, the overall system can still be inefficient if the
power needed to generate, modulate, and demodulate the signal is comparable to or larger
than the transmitted power. For applications requiring relatively small link margin (i.e.
low transmit power), such as WPAN and sensor networks, it becomes particularly
important to choose a modulation scheme that requires little power to implement so that
the system may remain efficient even with low power output. An ideal modulation
scheme would maximize link margin or capacity for a given signal power (i.e. smallest
(Eb/N0)min) without requiring complex, high-power circuits.
802.11g in its highest data rate represents a good example of what not to do if
energy conservation is the goal because 64-QAM has a high (Eb/N0)min and its
implementation is generally power hungry and quite complex. The receivers are high
power because demodulation requires a fast, high-precision ADC, substantial digital
signal processing, and linear amplification along the entire receive chain. The 802.11g

transmitters tend to be power hungry because generating the 64-QAM signals requires a
linear PA and a fast, low-noise PLL and VCO. Since the transistor devices constituting
the amplifiers (and all blocks) in a transceiver are inherently nonlinear, achieving linear
amplification in the receive chain and PA comes at the cost of increased power and/or
complexity.
In contrast to QAM and PAM, FSK and PSK have a common trait that only one
nonzero signal amplitude must be generated. This has important consequences for
system efficiency. First of all, the PA can be a nonlinear amplifier making much higher
efficiency possible. Secondly, since information is only carried in the phase (or
frequency) of the signal, the receive chain need not remain linear after channel selection,
so demodulation can be accomplished with a 1-bit quantized waveform. Finally, with
FSK (and some forms of PSK) it is possible to generate the necessary frequency shifts by
directly modulating the frequency of the VCO, thereby eliminating the transmit mixer
and saving power.
The potential power savings of direct VCO modulation depend strongly on the
phase accuracy required of the transmitter. If moderate frequency or phase errors are
tolerable, the VCO can simply be tuned directly to the channel with a digital FLL and
modulated open-loop [6] resulting in a simple, low power implementation. For phaseerror intolerant specs such as GSM, a variant of direct VCO modulation known as the 2point method is often used. In the simplest version of the 2-point method, a continuous
time (fractional-N) PLL with relatively low bandwidth attempts to hold the VCO
frequency steady while an external input modulates the VCO frequency. A high precision
DAC feeds forward a signal to cancel the error perceived by the PLL due to the
modulation [13]. Though the 2-point method eliminates the need for a transmit mixer,
the power consumed by the DAC and PLL curtail the potential power savings somewhat.
This method has been verified for 802.15.4 [8], Bluetooth [14], GSM [15] and other
standards.
3. Error Correcting Codes (ECC)
With respect to modulation scheme, a tradeoff between spectral efficiency and
energy efficiency has emerged from both theoretical and practical perspectives. First of
all, Shannons capacity theorem shows that the minimum achievable energy per bit for
any communication system is logarithmically related to spectral efficiency and several
popular (uncoded) modulation schemes, though not approaching the Shannon limit, do
exhibit a strong positive relationship between R/B and Eb/N0. Further, from a practical
perspective, the schemes with highest R/B, such as m-PAM or m-QAM with large m,
require complex and high power hardware to implement. The confluence of these factors
suggest that simpler schemes, such as 2-FSK, OOK, and 2,4-PSK, will offer the best
tradeoff when minimizing energy is the goal.
Even with an optimal demodulator, 2,4-PSK, 2-FSK, and OOK still require at
least 10 times higher (Eb/N0)min than the Shannon limit to achieve reasonably low
probability of error (i.e. BER = 10-4). The capacity equation tells us that, to approach the
Shannon limit and reclaim some of this wasted energy, the bandwidth efficiency R/B will
have to be reduced. Error correcting codes (ECC), such as Hamming, Reed-Solomon,
Turbo Codes, etc., can reduce (Eb/N0)min significantly, but also incur substantial
computational power overhead that could increase Eb-Sys enough to outweigh the (Eb/N0)min

reduction, particularly in low power systems. In [16], the (Eb/N0)min reduction (or coding
gain) and digital computation energy of several ECCs were evaluated for a 0.18m
CMOS process with 1.8V supply (). Though ECCs have traditionally found use in
higher power systems, these estimates would suggest that digital computation energy is
now low enough that ECCs are an effective option. ECCs will only become more
favorable as supply voltages and digital process features continue to scale.
B. Minimizing Overhead Power
Fig XX shows a simplified block diagram of a direct conversion or low-IF
transceiver the de facto standard topologies in todays commercial radios. The basic
functions of the transmitter are: generate a stable RF signal, modulate the frequency,
phase and/or amplitude of the RF signal according to information to be transmitted, and
drive the modulated signal onto the antenna with a PA. In a sense, energy consumed by
the modulation and signal generation circuitry constitutes overhead because it does not
contribute directly to the systems link margin. This overhead power (POH,TX) is, to firstorder, independent of transmitter output power. Efficient transmitter designs will spend
proportionally small amounts of energy generating and modulating the RF signal, with
the greatest share of energy consumed by the PA.
The receiver functions can be summarized as: low-noise, linear amplification,
selection of communication channel, and demodulation. The low noise amplifier (LNA)
boosts the incoming signal amplitude to overcome the noise of subsequent stages while
adding as little noise and distortion as possible. The excess noise contributed by an LNA
is inversely related to its power consumption; increasing power in the LNA directly
increases link margin. In a low-IF or direct conversion architecture, channel selection
and demodulation are accomplished with a VCO, mixers, low frequency filters, and other
circuits. A certain amount of power (POH,RX) must be spent in these blocks for the receiver
to function, but increasing their power beyond that point does not have as direct an
impact on link margin as increasing LNA power. A first order model of the power
performance tradeoffs in a generic transceiver is illustrated in () [17].
As mentioned in section X, the overhead power (POH,RX and POH,TX ) spent
generating and demodulating the RF signal is strongly dependent on the hardware
requirements of the modulation scheme employed. From a hardware standpoint, the
modulation schemes with lowest overhead are OOK and 2-FSK (with large frequency
deviations) because they require only a single non-zero signal amplitude and are tolerant
of moderate phase/frequency errors. These relaxed specifications permit simpler, lower
power modulation and demodulation circuits so that a larger proportion of the overall
power can be burned in the PA and LNA. However, even the most barebones low-IF or
direct conversion transceivers still require an RF VCO to operate. Thus, in the limit of
system simplicity, overhead power is VCO power.
1. Overhead Power in the VCO
A VCO is an autonomous circuit with either feedback or negative resistance
designed to cause periodic oscillation at one frequency; that frequency is set by an RC,
RL, or resonant LC network (). The vast majority of VCOs designed for communication
systems use a parallel LC resonator (or LC tank) to select the frequency of oscillation

because of its potential for superior noise performance. The power requirements and
noise performance of an LC VCO are largely determined by the impedance at resonance
(RT) and quality factor (Qtank) of this resonant LC tank.
Integrated circuit processes are inherently better suited to making capacitors than
inductors and, for frequencies below about 10GHz, the value of Qtank is usually limited by
the losses in the inductor. The inductor quality factor (QL) is:
L
QL o
Qtank
RL
For the parallel LC tank in (, left), the approximate magnitude of the tank
impedance at resonance (RT) is given by:
RT o L
QL

Some popular LC VCO topologies are shown in () [18]. A certain amount of


current is needed for oscillation to begin, but the current required to meet output swing
requirements is usually much greater. Typically, Vo must be at least a few hundred
milliVolts. Vo can be expressed as a constant times the product of ISS and RT for both
VCOs in (). Hence, RT must be maximized to minimize current, making high value, highQ inductors critical to reducing power in the VCO.
The choice of VCO topology is also an important consideration for minimizing
power. For instance, Vo as a function of ISS and RT for the NMOS only circuit is [18]:
2
Vo I SS RT

Whereas, Vo for the complementary (CMOS) VCO is:


4
Vo I SS RT

The CMOS VCO will deliver twice the output swing for a given current, but can
only generate half the maximum swing of the NMOS only device, which swings about
the supply rail. Thus, the CMOS would be the preferred choice unless it cant generate
sufficient swing. For a given bias current, the CMOS VCO provides twice the voltage
swing because the commutating current ISS flows through a parallel impedance of 2RT,
whereas the impedance seen by ISS in the NMOS VCO is only RT. The CMOS VCO can
also be seen as a vertical stack of two VCOs sharing the same bias current. As well see
below, stacking RF circuits to reuse bias current is a powerful tool for improving system
efficiency.
2. Voltage Headroom and RF Circuit Stacking
Even with the most barebones transceiver architecture significant power may still
be wasted if the available voltage headroom is not used optimally. Many mobile systems
use a 3.3V lithium supply, but the voltage swing required by the PA, VCO, or LNA may
be much lower. For instance, if a VCO is powered by a 3.3V supply but only needs to
generate a 300mV0-pk signal to drive mixers, buffers, or frequency dividers, there will be
substantial waste because the VCO swing spec could be met with a much lower supply
voltage.
Since supply voltage is typically not a flexible design variable, circuit techniques
are needed to optimize use of headroom when supply voltage is high. One way to reduce
wasted power is by stacking RF circuits [6]. Stacking is accomplished by placing two RF

blocks in series with respect to bias currents flowing from the supply. Thus, the current
used in one block is reused by another block. To avoid signal crosstalk between the two
blocks, they are isolated from each other with a large decoupling capacitor that provides a
low impedance node at high-frequencies. For integrated transceivers, stacking is only
feasible for high-frequency circuits where effective isolation can be implemented with
on-chip decoupling capacitors.
A few different stacked configurations are shown in (). The effect of stacking two
small-signal LNAs is to either double the transconductance gm (if the inputs and outputs
are coupled in parallel), or to increase the voltage gain Av (if signals traverse the LNAs in
series). Stacking two PAs doubles the output current, provided the halved voltage
headroom is still sufficient. PA stacking techniques are discussed in more detail in
section XX. In [6], the VCO was stacked with the LNA in the receiver and with the PA in
the transmitter. In this design, the current available to the PA and LNA was set by the
VCOs current requirements.
C. Receiver Noise Factor and Passive Voltage Gain
In this section, we will see how high impedances and passive voltage gain allow a
receiver to achieve good noise performance with reduced power. Noise performance of
RF receivers is most often reported using the noise factor (F) defined as the ratio of the
SNR at the receiver input to the SNR at the output. From a system perspective, F is the
factor by which link margin is degraded by the receivers own internal noise generators.
To maintain a given link margin, an increase in F must be compensated by an equivalent
increase in transmitted power.
In the absence of an input signal, F can be expressed as the ratio of the systems
total output noise to the output noise due to the source resistance. Referring to stage S1
with voltage gain Av in (, left), the squared voltage noise at the output is the sum of the
source noise times |Av|2 and the noise added by S1. Thus, F can be expressed:
2
2
2
SNRin Vn , S 1 Vn, src Av
FS1

SNRout
Vn2, src Av 2
Without loss of generality, we have chosen to sum noise contributions using
voltage gains and squared voltage noise rather than power gain and noise power.
Summing noise voltage is more convenient when the impedances between stages within
the receiver are not specified which is typically the case in integrated transceivers. We
add rms voltage noise because we assume the noise sources are uncorrelated.
Alternatively, we can represent the noise added by S1 with an equivalent input voltage
source that produces the same total output noise (, right).
Vn2,S 1
2
Vni , S 1 2
Av1
2
V ni is called the input referred noise voltage of S1. Referring noise to the input is
useful for determining minimum detectable signal levels because it gives a direct measure
of how large an input signal must be to overcome the noise contributed by the system and
source noise. From , we can express the noise factor of S1 in terms of its input referred
voltage noise.

2
2
Vni2, S 1
SNRin Vni , S1 Vn , src
FS1

1 2
SNRout
Vn2, src
Vn, src

A receiver is a cascade of stages, each having a different voltage gain and noise
contribution (). Each stage amplifies the signal and noise at its input and adds its own
noise. In general, the noise added by each stage is uncorrelated with the signal at its
input. If Avk and V2n,k represent the gain and output noise of the kth stage, respectively,
then the noise factor the cascaded system can be expressed.
2

Vni2,3
Vni2,casc
1 2 Vni ,2

Fcasc 1 2 Vni ,1 2 2 2 ... 1 2


Av1 Av1 Av 2
Vn , src
Vn, src

shows that the impact of noise added by a given stage is reduced by the square of
the total voltage gain preceding it. Typically, the first active stage in a receiver is a lownoise amplifier (LNA) achieving roughly 15-25dB of voltage gain. Thus, the following
stages can have much greater input referred noise than the LNA and still only a minor
effect on the cascaded system noise factor.
The noise contribution of an LNA depends on the current consumption, device
technology, circuit topology and other factors. However, LNA noise is usually dominated
by the input transconductor, consisting of one or more transistors biased for small signal
amplification. The input referred voltage noise of a CMOS transconductor (or Bipolar
device) can be related to current consumption directly:
Vni2
vdsat

4kT
2kT
f
gm
Id
Vdsat is called the saturation voltage and the right side of holds (roughly) for Vdsat
100mV. Though just represents the input noise of a single MOS transistor, its basic form
is common to most LNA topologies. Furthermore, the input referred noise of mixers,
low-frequency filters, and other stages following the LNA will generally be inversely
related to current consumption by a similar relation. Hence, from and , it is clear that
voltage gain at the front of the receiver chain reduces the current required to meet a given
noise spec.
1. Passive Voltage Gain with Resonant LC Networks
It is possible for a stage to achieve voltage gain without increasing the power in
the signal. This is only possible when the impedance at the output is larger than at the
input. For instance, if a given block is lossless and has an output impedance 100 times
greater than the input, then the output voltage will be 10 times larger than the input
(because power = V2/R will remain constant), but the signal current will decrease by a
factor of 10.
Passive transformers, resonant LC circuits, or even resonant electromechanical
devices, can achieve voltage gain while consuming zero power. Passive voltage gain is a
powerful tool for reducing receiver power consumption and is particularly well suited to
CMOS because CMOS transistors accept voltage as input and have a capacitive input
impedance that can be incorporated into a resonant network without contributing much
loss.

() is a tapped-capacitor resonant transformer, an LC network capable of delivering


passive voltage gain. In this circuit, RS is the source resistance and RL models resistive
loss in the inductor. The quality factor of an inductor (QL) is a common metric that
quantifies how close an inductor is to ideal. An ideal inductor would have infinite QL.
L
QL o
RL
The source driving the RF port has magnitude 2Vi to account for the voltage
dropped across RS. If the input impedance of the network is matched to RS, then Va = Vi.
Note that in reality there will be some capacitance in parallel with the inductor due to
both finite inductor self resonance frequency (SRF) and the input capacitance of any
devices connected to the output port. If the value of this parasitic capacitance is small
relative to C1, then we can safely neglect its impact.
From the output port, the matching network appears as a simple parallel LC tank
with a lossy capacitor and inductor. The lossy capacitor consists of elements C1, C2, and
source resistance RS and its effective quality factor (QC) is set by RS and the ratio of
capacitors C2 and C1. The overall network Q at the parallel resonance is a parallel
combination of QL and QC. QC may assume a wide range of values depending on the
values of C1 and C2, permitting design flexibility. QC is defined:
C
1
QC RS 2 C2 C1
C1
RS C1
The output impedance of the network at resonance is real and its magnitude is:
QQ
Ro o L C L
o
QC QL
Noise at the output is contributed by both RS and RL. If QC is very large, then the
overall Q is limited by the inductor and most noise at the output will come from RL,
leading to high noise factor. On the other hand, the network has the lowest noise factor
when QC is much smaller than QL because losses and output noise are dominated by RS.
To quantify the relationships between L, QL, and QC we first determine the voltage gain
from noise voltage sources in series with both RL and RS to the output, denoted AVL and
AVS, respectively:
2QC QL
AVL (o )
QC QL
2QC QL
o L
QC QL RS QC
Therefore, the noise factor (F) of the network at resonance becomes:
AVS (o )

Q
R A ( )
1 L VL o
1 C

RS AVS (o )
QL

The maximum gain is achieved when the source impedance is perfectly matched
to RL. This is an intuitive result because all power delivered to the network must be
dissipated in RL and the output voltage is largest when the current through RL is
maximum. Matching occurs when QL and QC are equal. Thus, from , the noise factor is 2
(NF=3dB) when matched. The voltage gain of the network when matched is:

2o L
RS RL
The noise factor, gain and S11 of a tapped-capacitor network are plotted versus C2
in (). In this example, the resonant frequency is 2.45GHz, inductance is 10nH and QL is
18. The inductance and QL are roughly based on values achieved with integrated
inductors in a current 130nm RF CMOS process. Though the network increases the
voltage amplitude of the signal, it actually decreases the signal power by a factor of F-1.
If used in a receiver front-end, as in [19], this network places a lower limit on the
achievable system noise factor, but it consumes no power, remains perfectly linear and
allows for substantial power reduction in subsequent stages due to its voltage gain.
AVS (o ) max

D. Efficient PAs with Low Power Output


As output power creeps below 1mW or so, designing an efficient transmitter
becomes increasingly difficult. First of all, there are numerous system blocks whose
power consumption does not necessarily scale down with transmitted power resulting in
a proportionally large power overhead. Secondly, with typical supply voltages of 1-3V
and an antenna impedance of roughly 50, standard PA topologies will be inherently
inefficient when putting out such little power. We have seen how power overhead can be
minimized by choosing the right modulation scheme, VCO topology, and, if necessary,
RF circuit stacking. We will now address a couple techniques for increasing the
efficiency of low power PAs.
When transmitting a constant envelope signal, a nonlinear PA can be employed to
increase efficiency. To avoid wasting power, the active element(s) in the PA should
switch on and off completely and have close to 0V across them when strongly conducting
implying the PA should be driven at or near its maximum possible voltage swing.
Hence, the most efficient output power for the PA is determined by the real part of its
load impedance RLoad and available zero-to-peak voltage swing vo,max.
v2
Pmax ff o ,max
2 RLoad
TodesignanefficientPAwithverylowpoweroutput,then,weneedasmallVmax
andlargeRLoad.
1.PATopologyandVmax
Generallyspeaking,supplyvoltageisfixedbyotherdesignconstraints,sovo,max
canonlybereducedbychangingthePAtopology.(),illustratesafewdifferentPA
topologieswithdifferentvaluesforvo,max.Atthefarright,twoidenticalpushpullPAs
areeffectivelystackedontopofoneanothertocutvo,maxbyafactorof4.Eachelectronin
theoutputcurrentflowsthroughtheloadfourtimes.Thus,foragivenaveragesupply
currentIdc,thisPAcandeliver4timesasmuchcurrenttotheloadasthePAatthefar
left.Thestackedpushpulltopologywaspresentedin[6],demonstrating40%efficiency
with250Woutputpowerinthe900MHzISMband.
ii.BoostingLoadImpedancewithResonantNetworks

AnotherkeytoincreasingefficiencyatlowpoweroutputistoincreaseRLoad.In
narrowbandsystems,RLoadcanbevariedoveraverywiderangebyusingaresonantLC
matchingnetworktotransformtheoriginalantennaimpedanceRA.Theratioofthe
transformedRLtotheoriginalRAtypicallyscaleswiththesquareoftheoverallnetwork
qualityfactor(Qnet).Hence,transformingimpedancebylargeratiosisonlyusefulfor
narrowbandsystemswheremoderatevaluesofQareacceptable.Theparasiticseries
resistanceofrealinductorsandcapacitorswillplaceanupperlimitonQnetand,therefore,
themaximumachievableimpedancetransformationratio.Thenetworkwilltendtohave
poorefficiencyasitapproachesthemaximumtransformationratio.
Asanexample,wecomputetheoutputimpedanceandnetworkefficiencyforthe
tappedcapacitordescribedinsectionXasafunctionofRAandQL.To simplify the
calculations, we assume the inductor is much lower Q than the capacitor Q (seesection
X)andignoretheseriesresistanceofthecapacitors.Theoutputimpedanceofthe
networkisexpressedinandtheefficiencyisshownbelow,withQC,effasdefinedin
sectionX.
QL

QC QL
ThehighestefficiencyoccursatC2=0,whereQC,effisminimized.Notethatthis
isactuallytheinverseof,whichrepresentsthenetworknoisefactor.Astheimpedance
ratioisincreased,alargerproportionofthesignalpowerisdissipatedinRL,resultingin
lowerefficiency.
IV. A Low Energy 2.4GHz Transceiver
In this section, we examine a low energy transceiver with respect to the energy
saving techniques discussed thus far. A much more detailed, circuit focused analysis of
the system is carried out in [19]. This 2.4GHz transceiver, implemented in a 0.13m RF
CMOS process, achieves 1nJ per received bit and 3nJ per transmitted bit with 300W
transmit power and 7dB receiver noise figure. The resulting energy efficiency figure of
merit is -30dB, and is actually dominated by the high Eb/N0 required to demodulate 2FSK noncoherently.
The transceiver block diagram is shown in (). A 400mV supply was chosen for
this system to accommodate a single solar cell as the power source. In sunlight the entire
transceiver could operate continuously from a 2.6mmx2.6mm silicon solar cell [20].
Because of the reduced supply voltage, all circuits are made differential to increase
available swing. This transceiver uses 2-FSK with a relatively large tone separation,
effectively trading spectral efficiency for a simplified low-power architecture. This
tradeoff is particularly favorable for sensor network applications, wherein data rates
below 1Mbps are the norm and 85MHz of unlicensed spectrum is available in the 2.4GHz
ISM band.
In the transmitter, a cross-coupled LC VCO directly drives an efficient nonlinear
PA taking advantage of the relaxed phase accuracy requirements to eliminate
upconversion mixers and transmit buffers by using direct VCO modulation. Furthermore,
the PA input capacitance is incorporated into the resonant LC tank of the VCO to
minimize the current consumed driving the PA. Since the supply voltage is so low, the

NMOS-only VCO architecture was chosen to achieve maximize swing. The differential
PA drives a tapped-capacitor resonator to boost its load impedance from 50 to about
1k. The PA achieves 45% efficiency at 300W output power and the overall transmitter
efficiency is 30%. A plot of transmitted power versus total power consumed is shown in
().
The tapped capacitor network at the PA output is also used in the receiver frontend to achieve impedance matching and substantial voltage gain effectively supplanting
an RF LNA. This network interfaces directly to highly linear CMOS passive mixers
designed to present a high impedance to the LC network to avoid reducing its voltage
gain. A reconfigurable front-end was devised to limit capacitive loading on the LC
network and the VCO by reducing transistor count in the front-end. In essence, a single
quad of transistors can be configured as a PA or mixer, depending on bias voltages and
the states of a couple switches. () illustrates this reconfigurable topology.
The passive mixers downconvert the desired signal to baseband and attenuate
wideband interference with a 1MHz first order low-pass filter at the output. The overall
voltage gain from the balanced 50 receiver input to the mixer output is about 17dB.
The mixer outputs drive a sequence of bandpass filters with enough gain to convert the
incoming signal to a 1-bit quantized waveform for simple demodulation. The cascaded
receiver noise factor versus power consumption is shown in (), and a summary of the
measured performance is in ().
V. Summary and Conclusions
We began this chapter using Shannons capacity theorem to determine the
fundamental energy requirements of wireless communication from the most general,
theoretical standpoint. As a natural consequence of this discussion, we proposed a figure
of merit for evaluating the energy efficiency of real wireless systems relative to
fundamental limits. Next, we separated the major contributors to low efficiency by
expressing this figure of merit in terms of a practical systems (Eb/N0)min, global transmit
efficiency, and receiver noise factor.
Examining (Eb/N0)min figures and architectural implications of some popular
modulation schemes revealed that modulation is extremely important in determining a
systems energy efficiency and suggested that simple schemes such as OOK and 2-FSK
would be well suited to low-energy systems. Furthermore, error correcting codes (ECC)
with substantial coding gain are becoming feasible even for low power systems, owing to
the continued scaling down of digital CMOS [16]. With the assumption of an OOK, FSK,
or PSK system, several circuit techniques were presented aimed at boosting PA
efficiency, reducing power overhead, or achieving low noise factor with a low power
receiver. Finally, we looked at a 2.4GHz transceiver incorporating many of these
techniques to achieve high energy efficiency at very low transmit power.

You might also like