Received February 5, 2019, accepted February 22, 2019, date of publication February 28, 2019, date of current version

March 20, 2019.

Digital Object Identifier 10.1109/ACCESS.2019.2902365

Pulse Characteristics of a Direct Antenna

Modulation Transmitter
KURT SCHAB 1 , (Member, IEEE), DANYANG HUANG2 ,
AND JACOB J. ADAMS 2 , (Senior Member, IEEE)
1 Department of Electrical Engineering, Santa Clara University, Santa Clara, CA 95053, USA
2 Antennas and Electromagnetics Laboratory, Department of Electrical and Computer Engineering, North Carolina State University, Raleigh, NC 27695, USA
Corresponding author: Kurt Schab (kschab@scu.edu)
This work was supported by the DARPA under Grant D16AP00033. The work of K. Schab was supported by the Intelligence Community
Postdoctoral Research Fellowship Program.

ABSTRACT Direct antenna modulation (DAM) methods using time-varying matching networks offer the
possibility to reduce the detrimental effects on broadband signal transmission from a narrowband antenna.
We propose time- and frequency-domain models for the effects of a DAM scheme for the transmission of
ON–OFF-keyed RF signals and test the validity of the proposed models through over-the-air measurements at
27.12 MHz. Random bit sequences are transmitted separately using both a switched-mode DAM transmitter
and a conventional transmitter and received in the far field. Comparison of measured data with our frequency-
domain models shows that the DAM transmission method partially removes the effects of narrowband
filtering due to the transmit antenna. Models and measurements also demonstrate that the removal of this
filter improves the signal quality as observed in eye diagrams of the demodulated received signals. Finally,
a simple time-domain model of the DAM transmitter is shown to predict the radiated transient signal under
varying synchronization conditions.

INDEX TERMS Antenna measurements, electrically small antennas, narrowband antennas, time varying
circuits.

I. INTRODUCTION In fact, techniques for circumventing the conventional

The performance of an antenna having dimensions much
smaller than its operating wavelength is limited by sev-
eral well-known bounds. Perhaps most referenced is Chu’s
limit on Q-factor (Q or quality factor), which implies that
small antennas are constrained by a lower bound on Q-factor
that rises with decreasing electrical size [1]. Because of the
inverse relationship between fractional impedance bandwidth
and Q-factor [2], this lower bound on Q-factor has often been
interpreted as an upper bound on impedance bandwidth for a
particular electrical size.
However, the relation between Q-factor and bandwidth
assumes that the antenna and matching network consists
of linear time-invariant (LTI) components, and there is no
immediate reason to suspect its applicability to systems that
are non-linear or time-varying. This implies that electri-
cally small antennas, while obeying all standard notions of
energy conservation and bounds on Q-factor, may not have
strictly bounded effective transmit or receive bandwidths
when interfaced with non-LTI components (e.g., switches,
dynamic tuners).
The associate editor coordinating the review of this manuscript and
approving it for publication was Yejun He.
2169-3536
2019 IEEE. Translations and content mining are permitted for academic research only.
VOLUME 7, 2019 Personal use is also permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
bandwidth limitations of electrically small antennas using
time-varying matching networks (i.e., direct antenna modu-
lation, DAM) were first proposed for use in the VLF regime
in the 1950s-1970s [3]–[8]. More recently, these have been
studied using modern numerical techniques [9]–[12] and
implemented at radio frequencies using solid-state switching
devices [9], [10]. While the aforementioned approaches focus
on manipulating the oscillations of a resonant narrowband
system, other techniques employ switched DC excitations for
the direct impression of signals at the terminals of an antenna
[13], [14]. Note that, in contrast with the methods described
in [15]–[17], all of the aforementioned schemes for DAM
require synchronization of time-varying elements relative to
the RF carrier cycle in order to manipulate energy storage
and radiation. As such, we refer to these schemes as energy-
synchronous DAM.
The purpose of this paper is to propose time- and
frequency-domain models for an energy-synchronous DAM
technique used to transmit broadband on-off-keyed (OOK)
signals from an electrically small monopole antenna and to
compare these models to measured far fields produced by a
DAM transmitter. Robust models are essential to predict the
potential performance of an OOK-DAM transmitter and to
30213

provide insight into the non-ideal factors that drive engineer-
ing requirements. However, prior studies on OOK-DAM lack
systematic and detailed comparisons between measured far-
field data and theoretical models. In this work, we measure
radiated bit sequences transmitted using DAM and compare
their spectral and time domain signal characteristics to those
predicted by simple models. Furthermore, we directly com-
pare these to signals produced by the same antenna operated
in its conventional (passive or non-switched) mode to assess
the resulting differences in measured and modeled signal
characteristics.
The direct modulation technique under study is briefly
described in Sec. II. In Sec. III, we propose a system block
frequency-domain model to test the ability of DAM to elim-
inate the impedance mismatch effects of the narrowband
transmit antenna. In Sec. IV a time-domain model is used to
study the effects of varying switch timing and other non-ideal
switching effects. Finally, these models are tested experimen-
tally using an over-the-air communications link in Sec. V.
II. ENERGY-SYNCHRONOUS ON-OFF-KEYING
This work focuses on the analysis of an energy-synchronous
DAM scheme for transmitting broadband OOK signals
using electrically small dipole antennas, originally pro-
posed in [4]. Here, we simply denote this scheme as DAM
OOK and provide a brief explanation of the concept. In
principle, this approach is applicable to any electrically
small antenna that presents an open-circuit input impedance
at DC.
The schematic in Figure 1 shows a variation of the original
circuit used in [4], where a small dipole antenna (represented
by an RC circuit) is tuned to resonance at frequency fc by an
inductor Lm . While this model is ideal in several ways (ideal
switch, inductor and simple RC antenna model), we use it
here to explain the underlying DAM concept.
In the case of a conventional transmitter configuration,
an OOK modulated source would be attached to the tuned,
narrowband antenna with the switched placed permanently
in the ON position. The narrow bandwidth of the antenna
naturally results in distortion of the transmitted signal. For
the DAM transmitter configuration, the tuned antenna is
connected to an OOK modulated or continuous-wave (CW)
source and the switch is used to generate the modulation
sequence by controlling the storage and release of energy in
the antenna as described below.
Suppose that for times t < 0, the switch is open and
the antenna terminals are charged with a DC voltage V0
matching the steady-state amplitude of the voltage across the
antenna terminals when driven at frequency fc . Because the
switch is open at times t < 0, the antenna terminal current,
ia is zero. These conditions imply that the stored energy on
the antenna’s capacitance is CV02 /2 and the energy in the
inductor is zero. These energy states precisely match those
of the system when driven at resonance during the instant
in the cycle when the antenna terminal voltage is maximum,
assuming high Q-factor. If, as shown in Figure 1 at time t = 0,
30214
K. Schab et al.: Pulse Characteristics of a DAM Transmitter
FIGURE 1. On-off-keying scheme using direct antenna modulation
(following [4]). A small dipole is modeled by an RC network and driven
using a switched matching network (top). Antenna terminal current ia and
voltage va are depicted under ideal synchronization (middle) as well as
the approximate radiated field Eff (bottom).
the switch is closed at the instant the driving source voltage is
maximum, the initial conditions match the steady state oscil-
lation conditions and the antenna instantaneously ‘‘resumes’’
oscillating as though at it had been in steady state oscillation
for all time. Similarly, if the switch is opened again when the
antenna terminal voltage va takes on the maximal value of V0 ,
energy is trapped on the antenna and the system state becomes
exactly that of times t < 0. The source may be turned off
while the switch is open or an asymmetric absorbing switch
may be used to prevent damage to the generator caused by
reflection.
Assuming short-dipole behavior, the antenna termi-
nal current ia is linearly related to an equiphase cur-
rent over the antenna which generates a radiated field
Eff ∼ F(θ, φ)∂ia /∂t [18], also shown in Figure 1. Here
F(θ, φ) is a spatial term depending only on the dipole’s
orientation. Second-order transient effects manifest in more
complex models of small radiators [11]–[13], however these
transients are specific to an antenna’s geometry, are rela-
tively low amplitude, and typically occur at frequencies much
higher than the tuned carrier frequency. As such, we do not
treat these transients here.
In the above description we have explicitly required proper
synchronization of the carrier frequency and switch actu-
ation. For this OOK scheme, the ideal switching instant
always coincides with the peak antenna terminal voltage va .
Deviation from this ideal condition can be described by the
time τ , as drawn in Fig. 2. Desynchronization of this kind is
studied in detail in Sec. IV.
VOLUME 7, 2019
K. Schab et al.: Pulse Characteristics of a DAM Transmitter
FIGURE 2. Desynchronization time τ as the difference between the ideal
switching instant (tideal , peak antenna voltage va ) and a non-ideal,
desynchronized switch time tdesynch .
III. FREQUENCY-DOMAIN MODEL
The direct modulation technique described in [4], [9],
[11], [19], and [20] implies that an electrically small
antenna’s narrowband filtering due to impedance mis-
match can be effectively removed, i.e., DAM OOK
allows for broadband signals to be directly impressed
at a transmitting antenna’s terminals through time-
varying impedance modulation. To study this effect,
we isolate the filtering behavior of receiving and transmit-
ting antennas’ impedance mismatch responses via a transfer
function relating the voltage Vtx at the transmitting generator
to the voltage Vrx appearing on a receiver load. In terms of
antenna effective heights h, antenna reflection coefficients Γ ,
and propagation tensor P, this transfer function for a conven-
tional link is [21]
jη0 βhrx · P · htx
 
Conv.
Vrx = (1 − Γrx ) (1 − Γtx )Vtx
4R0
= (Hrx Hch Htx ) Vtx . (1)
Here, β is the free space wavenumber, η0 is the impedance
of free space, R0 is the generator impedance, and the ten-
sor P contains all propagation effects, reducing to the scalar
P0 e−jβR /4π R for free-space links. Having decomposed the
transfer function into components that depend only on the
mismatch between the antennas and their feeds (Htx = 1−Γtx
and Hrx = 1 − Γrx ), we collect all propagation and radiation
pattern effects into a third term Hch . Should the above relation
be cast in terms of power transfer, the usual forms of (1−|Γ |2 )
follow.
If DAM OOK performs as proposed in [4], [9], [11], [19],
and [20], then the transmit antenna’s impedance mismatch
filter is eliminated using an ideal DAM transmitter. Thus,
we can propose a modified model for the received signal
VrxDAM ≈ Hch Hrx Vtx . (2)
The expression in (2) provides a physically-motivated,
frequency-domain model for the effect of an ideal DAM
transmitter in a communications link. In Sec. V, we test
the model for DAM proposed in (2) using over-the-air
VOLUME 7, 2019
measurements while using the model of the conventional
transmitter in (1) as an experimental control.
IV. TIME-DOMAIN MODELS
The frequency-domain model in (2) assumes the ideal switch
behavior and control synchronization described in Sec. II.
If these conditions are not met, the total bypassing of the
transmitter mismatch filter Htx = 1 − Γtx is not possible.
Deviations from the ideal case thus imply that the system can
no longer be represented by an equivalent transfer function
and must be studied directly in the time domain. In this
section, we derive a time-domain expression for the radiated
waveform under imperfect synchronization in order to assess
how desynchronization impacts DAM transmissions.
The method shown in Fig. 1 relies on precise switch
control to store energy on the antenna during non-radiating
states to facilitate fast transitions to the radiating state with-
out the characteristic rise time observed in a conventional
narrowband transmitter. The following analysis denotes the
desynchronization time τ (see Fig. 2) in terms of the relative
phase of the carrier at the switching instant φ = 2π fc τ .
As the switching phase progresses from ideal (φ = 0◦ -
when maximum voltage appears at the antenna terminals),
to worst-case (φ = ±90◦ - when zero voltage appears at
the antenna terminals), a diminishing amount of energy is
retained by the antenna during OFF symbols. This results
in the need to partially recharge the system and realign it
to the source during each radiating symbol. Note that while
desynchronization affects the rise time and amplitude of the
radiated field, the fall time of the transmitted pulse remains
abrupt regardless of switch timing [17].
If the simple RC model of a short dipole antenna in Fig. 1
is adopted, the form of a desynchronized radiated pulse can
be written analytically as
Eff (t) = b cos φ cos(2π fc t + φ)e−α(t−t0 )
h i
+ cos(2π fc t) 1 − e−α(t−t0 ) , t ∈ t0 , t0 + Tsym
 
(3)
where the steady state (t → ∞) value is normalized to unity,
the pulse begins at time t = t0 , and α ∼ Q−1 is a time
constant depending on the antenna’s loaded Q-factor. The
factor b ∈ [0, 1] models non-idealities that may result in the
incomplete energy capture and release when moving between
states. This effect is independent of switch control synchro-
nization and is due to several factors including finite transi-
tion times between a switch’s high and low impedance states
and non-linear behavior of the switch under high voltages.
Using measurements in Sec. V, we estimate the parameter b
and test the ability of (3) to predict radiated fields on RF
timescales.
V. EXPERIMENTAL VALIDATION
Here we report on over-the air measurements conducted
outdoors on North Carolina State University campus. All
experiments were conducted in the far field using a carrier
30215

frequency of 27.12 MHz. Data were collected to test the
following hypotheses:
1: The relative behavior of a conventional and synchro-
nized DAM OOK system is represented by the models
in (1) and (2), respectively.
2: The model in (3) accurately predicts the time-domain
waveforms produced by a DAM transmitter under vary-
ing degrees of desynchronization.
A. MEASUREMENT SETUP
The measurement setup is drawn schematically in Fig. 3.
The transmitting and receiving antennas are wire monopoles
mechanically supported by freestanding, non-conducting
masts. The antennas are fed at ground level and radials
are used to reduce ground losses. The transmit antenna
measures 0.09λ (0.94 m) tall and is tuned to resonance
at 27.12 MHz using a series inductance of approximately
1600 nH. A CMOS-based reflective switch (Analog Devices
ADG902) is used to implement DAM OOK. A 3 : 1 trans-
former is used to match the tuned antenna to an arbitrary
waveform generator (AWG, Tektronix AWG70002A). To
facilitate synchronization control, we configure the AWG to
produce both RF and baseband switch control signals.
FIGURE 3. Schematic diagram of over-the-air experiment and measured
transmitter and receiver reflection coefficient, as represented by S11 .
The receiving antenna measures 0.22λ (2.43 m) tall and is
well matched at the selected carrier frequency. The reflection
coefficient Γ and −10 dB fractional bandwidth FBW−10 dB
of both antennas are shown in Fig. 3. Time-domain signals at
30216
K. Schab et al.: Pulse Characteristics of a DAM Transmitter
the receiver are recorded using a digital oscilloscope (DPO,
Tektronix DPO70404C) and a +30 dB low-noise amplifier
(LNA, Minicircuits LNA-530).
The two configurations use identical hardware in their
signal paths. By including the switch in both conventional and
DAM modes, we mitigate potential biasing of comparison
results due to resistive broadbanding effects from switch
losses. Efficiency reduction due to the switch is discussed
in Sec. V-E.
Data were recorded at varying data rates using both con-
ventional and DAM OOK configurations. In conventional
mode, the transmitter switch is fixed in the closed position
and the AWG produces an OOK modulated 64-bit pseudo-
random bit sequence (PRBS). In DAM OOK mode the AWG
outputs a CW signal with the same voltage level and controls
the switch according to the scheme described in Section II.
In both modes of operation, a startup time of many ‘‘1’’
bits is used to charge the system to steady state before the
bit sequence begins. Likewise, a cooldown sequence of ‘‘0’’
bits is used to fully discharge the system between sequence
repetitions.
Along with measured antenna impedances and the model
in (1), a channel sounding sequence is used to obtain an
estimate of the channel Hch . The sounding signal consists
of a linear chirp spanning a symmetric fractional bandwidth
of 75% around the carrier frequency, large enough to con-
tain the broadband signals used in the experiments. Each
measurement consists of transmitting and recording 50 rep-
etitions of both chirp and data sequences. No averaging
is used during data collection. Parsing and averaging of
the repeated sequences is used to raise signal-to-noise ratio
in post-processing. Independent GPS-disciplined oscillators
supply synchronized triggering and clock references to the
transmitter and receiver.
B. POST-PROCESSING SETUP
The time-domain data are parsed and averaged before esti-
mating the channel Hch and applying any RF or baseband pro-
cessing. Measured data were coherently demodulated numer-
ically and processed using the block diagram in Figure 4 to
produce the in-phase envelope yi . The initial bandpass filter is
centered at the carrier frequency and has a rectangular profile
in frequency with fractional bandwidth of 75%. Final low-
pass filtering is carried out using a rectangular profile filter
with a fractional bandwidth of 75% relative to the carrier fre-
quency. The width of both filters is set to retain components
of the signal within the spectral width of the channel sounding
chirp.
FIGURE 4. I/Q demodulation chain used to extract in-phase envelope yi
from a measured or modeled signal Vrx .
VOLUME 7, 2019
K. Schab et al.: Pulse Characteristics of a DAM Transmitter

C. EXPERIMENT 1: REMOVAL OF TX FILTER VIA DAM OOK
Sequences of OOK signals were transmitted at three data
rates characterized by the number of carrier cycles N per
symbol. The values of N = 10, 5, 3 correspond to nominal
fractional bandwidths B = 20%, 40%, 66%, respectively.
For each measurement, the channel Hch was estimated using
the channel sounding sequence and modeled signals were
calculated according to (1) or (2), depending on the mode
of operation. For clearer interpretation, both modeled and
measured data are presented after channel inversion (i.e., mul-
−1
tiplication by Hch ). To avoid amplification of noise near the
edges of the receiving antenna’s bandwidth, the filter Hrx is
not removed in this way. Proper synchronization for the DAM
switching control is established experimentally.
Figure 5 shows the RF spectra of measured (light orange)
and modeled (dark purple) data for conventional (Conv.) and
DAM modes. No normalization is applied. Bars display the
99% power bandwidth (occupied bandwidth) of each spec-
trum. The bandwidth of the conventional system is limited
by the transmit antenna’s frequency response and does not
increase significantly with increasing data rate. Conversely,
the bandwidth of the DAM transmission does expand as
the bandwidth of the data sequence increases. In all cases,
good agreement is observed between measured and modeled
data, implying that the DAM transmitter effectively removes
the impedance filter of the narrowband transmit antenna as
predicted in (2).
it more prone to noise-induced errors. Note that we do not
quantify specific performance metrics here such as rise and
fall time, bit error rate, etc. as these are heavily influenced by
the receiver design, which is beyond the scope of the present
study.
Comparison of measured and modeled data in DAM mode
show general agreement, with much more open eye diagrams
at all data rates. One noticeable difference between the DAM
model and measurements is the asymmetric rise and fall times
in the measured data. This is attributed to imperfect switch
transition characteristics reducing the ability of the system to
store energy between each pulse. From Fig. 6, we estimate the
switch factor in (3) as b = 0.65 from the relative immediate
height of rising edges.
The data in Fig. 6 are obtained by averaging the 50 mea-
sured repetitions of a single transmission. The mean data for
the N = 3 case is shown again as a solid line in Fig. 7 as a
function of time (note this trace is identical to the measured
data in the N = 3 column of Fig. 6, only not collapsed into
eye diagram format). In Fig. 7 we also plot a shaded region
illustrating the interval of width 2σ (t), where σ (t) is the stan-
dard deviation of the 50 measured trials at each time sample.

FIGURE 6. Demodulated in-phase envelopes produced from

channel-inverted measured (light orange) and modeled (dark purple)
received signals in conventional (top) and DAM (bottom) modes.

−1 V (ω)| from measured (light orange) and

FIGURE 5. Spectra |Hch rx
modeled (dark purple) received signals in conventional (top) and
DAM (bottom) modes.

The measured data and models in (1) and (2) were

also studied in the time-domain to assess their compara-
tive signal characteristics. Figure 6 contains demodulated
data from measurement and models in eye-diagram format.
As in Fig. 5, no normalization is used. As an experimental
control, the measured and modeled data in the conventional
mode of operation show excellent agreement, with character-
istic rise and fall times that partially or completely close the
eye as the data rate is increased (N is decreased). Depending
on specific demodulation and decision parameters, this high
FIGURE 7. Mean (solid line) and uncertainty (shaded region, width equal
level of intersymbol interference may hinder the communi- to two standard deviations σ ) of measured N = 3 signals after
cations performance of the conventional transmitter, making demodulation.

VOLUME 7, 2019 30217


The function σ (t) is nearly constant and identical between
the conventional and DAM measurements, suggesting that
the noise is an ergodic process. This analysis supports the
claim that the differences observed in Figs. 5 and 6 are due
to deterministic changes in the way the transmitting antenna
produces a radiated field and are not generated by chance due
to a random process.
D. EXPERIMENT 2: DESYNCHRONIZATION MODEL
To test the validity of the time-domain desynchronization
model in (3), measured RF waveforms are studied under
varying degrees of desynchronization. As in the previous
experiment, channel inversion is applied to the measured data.
Corresponding model data is generated according to (3) along
with convolution with the measured receive filter Hrx . The
switch factor b = 0.65 is used according to estimates made
from the rising edge of symbols in Fig. 6. Modeled and mea-
sured RF waveforms are plotted in Fig. 8 showing excellent
agreement. In all cases, the DAM waveforms exhibit much
shorter fall times than the conventional system. The rise times
vary with synchronization φ, with a fast rise to 65% of the
maximum envelope value with φ = 0◦ and a slow rise time
identical to the conventional system with φ = 90◦ .
FIGURE 8. Measured RF waveforms in conventional and DAM modes for
varying degrees of desynchronization phase φ. Modeled results with
switch factor b = 0.65 are overlaid with measured results for four bits
(‘‘0110’’) near tfc /N = 20. For consistent comparison, the measured
results are channel inverted and the modeled results are passed through
the measured receive filter Hrx .
In addition to validating the time-domain model in (3),
these results demonstrate the effect of desynchronization
within a realized DAM system. Interestingly, the degrada-
tion of rise time is slow with respect to increasing desyn-
chronization, owing to its cos φ dependence. This suggests
some resilience to imperfect timing in the implementation
of such a transmitter. For example the immediate envelope
signal height at t = t0 lowers only to 64% of its maximum
when φ = 50◦ .
E. DISCUSSION OF EFFICIENCY TRADEOFFS
In the reported measurements, the transmitter hardware was
kept identical between conventional and DAM trials. Based
on relative input impedance and bandwidth measurements,
30218
K. Schab et al.: Pulse Characteristics of a DAM Transmitter
inclusion of the DAM circuitry lowers the transmitter effi-
ciency by a multiplicative factor of 0.81, i.e., a reduction in
gain of 0.91 dB. This drop in efficiency corresponds to an
increase in conventional impedance bandwidth by a factor
of 1.23. While bandwidth in a conventional system can con-
tinue to be increased via resistive loading, it comes at the cost
of lower efficiency and system gain. The use of DAM presents
an alternative scheme wherein a fixed cost in efficiency
incurred from the switching elements allows for significant
improvements in effective bandwidth. Because the concept of
impedance bandwidth is not well-defined when using time-
varying matching networks, the effective bandwidths of a
conventional and DAM transmitter must be examined at the
signal level.
VI. CONCLUSIONS
Time- and frequency-domain models for a direct antenna
modulation (DAM) scheme for on-off-keying are formulated
and validated through physical measurements. A conven-
tional transmitter and its corresponding model are used as
an experimental control. Measured data show good agree-
ment with the proposed models for the received spectra and
envelopes of DAM OOK transmissions that assume ideal
switching conditions. However, the measured data show a
longer rise time for the DAM OOK transmissions than the
model due to non-ideal factors in realization. A separate
model for the time-domain behavior of the DAM OOK sys-
tem under desynchronized operating conditions accurately
predicts measured data at the RF level.
The models presented here improve the understanding of
the energy synchronous DAM OOK scheme and provide
concrete comparisons to conventional transmitters for the
assessment and experimental validation of this scheme. Sim-
ilar models may be developed and applied to other forms of
direct antenna modulation.
ACKNOWLEDGMENT
All statements of fact, opinion, or analysis expressed are those
of the author and do not reflect the official positions or views
of the Intelligence Community or any other U.S. Government
agency. Nothing in the contents should be construed as
asserting or implying U.S. Government authentication of
information or Intelligence Community endorsement of the
author’s views. Approved for public release; distribution is
unlimited.
REFERENCES
[1] L. J. Chu, ‘‘Physical limitations of omni-directional antennas,’’ J. Appl.
Phys., vol. 19, no. 12, pp. 1163–1175, Dec. 1948.
[2] A. D. Yaghjian and S. R. Best, ‘‘Impedance, bandwidth, and
Q of antennas,’’ IEEE Trans. Antennas Propag., vol. 53, no. 4,
pp. 1298–1324, Apr. 2005.
[3] H. Wolff, ‘‘High-speed frequency-shift keying of LF and VLF radio cir-
cuits,’’ IRE Trans. Commun. Syst., vol. 5, no. 3, pp. 29–42, Dec. 1957.
[4] J. Galejs, ‘‘Switching of reactive elements in high-Q antennas,’’ IEEE
Trans. Commun. Syst., vol. 11, no. 2, pp. 254–255, Jun. 1963.
VOLUME 7, 2019
K. Schab et al.: Pulse Characteristics of a DAM Transmitter
[5] P. Johannessen, ‘‘Automatic tuning of high-Q antenna for VLF FSK
transmission,’’ IEEE Trans. Commun. Syst., vol. 12, no. 1, pp. 110–115,
Mar. 1964.
[6] H. Hartley, ‘‘Electronic broad banding of VLF/LF antennas for FSK
radio communication,’’ IEEE Trans. Commun. Technol., vol. 19, no. 4,
pp. 555–561, Aug. 1971.
[7] L. Vallese, ‘‘VLF-LF wide shift FSK antenna feed networks,’’ Electro-
physics Corp., Fairfield, NJ, USA, Tech. Rep. AD-746119, 1972.
[8] J. T. Gamble, ‘‘Wideband coherent communication at VLF with the exper-
imental transmitting antenna modulator (etam),’’ Rome Air Development
Center Griffiss AFB, Rome, NY, USA, Tech. Rep. AD 916687, 1973.
[9] X. Xu, H. C. Jing, and Y. E. Wang, ‘‘High speed pulse radiation from
switched electrically small antennas,’’ in Proc. IEEE Antennas Propag.
Soc. Int. Symp., Jul. 2006, pp. 167–170.
[10] M. Salehi, M. Manteghi, S.-Y. Suh, S. Sajuyigbe, and
H. G. Skinner, ‘‘A wideband frequency-shift keying modulation technique
using transient state of a small antenna,’’ Prog. Electromagn. Res.,
vol. 143, pp. 421–445, 2013.
[11] R. Janaswamy, ‘‘Time-varying antennas for enhanced bandwidths,’’ in
Proc. IEEE Antennas Propag. Soc. Int. Symp. (APSURSI), Jul. 2014,
pp. 621–622.
[12] K. Schab and J. Adams, ‘‘Calculation of radiation transients in direct
antenna modulation systems,’’ in Proc. IEEE Int. Symp. Antennas Propag.,
Jul. 2017, pp. 521–522.
[13] J. T. Merenda, ‘‘Digital wideband small antenna systems,’’ BAE Systems,
Greenlawn, NY, USA, 2006.
[14] M. Manteghi, ‘‘A wideband electrically small transient-state antenna,’’
IEEE Trans. Antennas Propag., vol. 64, no. 4, pp. 1201–1208,
Apr. 2016.
[15] V. F. Fusco and Q. Chen, ‘‘Direct-signal modulation using a silicon
microstrip patch antenna,’’ IEEE Trans. Antennas Propag., vol. 47, no. 6,
pp. 1025–1028, Jun. 1999.
[16] S. D. Keller, W. D. Palmer, and W. T. Joines, ‘‘Switched antenna circuit
with increased information bandwidth,’’ IEEE Antennas Wireless Propag.
Lett., vol. 9, pp. 1045–1048, 2010.
[17] S. Srivastava and J. J. Adams, ‘‘Analysis of a direct antenna modulation
transmitter for wideband OOK with a narrowband antenna,’’ IEEE Trans.
Antennas Propag., vol. 65, no. 10, pp. 4971–4979, Oct. 2017.
[18] C. A. Balanis, Antena Theory. New York, NY, USA: Harper and Row, 1982.
[19] X. Xu and Y. E. Wang, ‘‘Wideband pulse transmission from switched elec-
trically small antennas,’’ in Proc. IEEE Radio Wireless Symp., Jan. 2007,
pp. 483–486.
[20] X. J. Xu and Y. E. Wang, ‘‘A direct antenna modulation (DAM) trans-
mitter with a switched electrically small antenna,’’ in Proc. Int. Workshop
Antenna Technol. (iWAT), 2010, pp. 1–4.
[21] D. Lamensdorf and L. Susman, ‘‘Baseband-pulse-antenna techniques,’’
IEEE Antennas Propag. Mag., vol. 36, no. 1, pp. 20–30,
Feb. 1994.
VOLUME 7, 2019
KURT SCHAB (S’09–M’16) received the B.S.
degree in electrical engineering and physics from
Portland State University, in 2011, and the M.S.
and Ph.D. degrees in electrical engineering from
the University of Illinois at Urbana–Champaign,
in 2013 and 2016, respectively. From 2016 to
2018, he was an Intelligence Community Postdoc-
toral Research Scholar with North Carolina State
University, Raleigh, NA, USA. He is currently an
Assistant Professor of electrical engineering with
Santa Clara University, Santa Clara, CA, USA. His researches focus on
the intersection of numerical methods, electromagnetic theory, and antenna
design.
DANYANG HUANG was born in Guangzhou,
China, in 1990. He received the B.S. degree in
physics from the Inner Mongolia University of
Technology, Huhhot, China, in 2013, and the M.S.
degree in electrical engineering from Minnesota
State University, Mankato, MN, USA, in 2016.
He is currently pursuing the Ph.D. degree with
North Carolina State University, with a focus on
switching transients and parasitic effects in direct
antenna modulation. His research interests include
substrate-integrated waveguide circuits and direct antenna modulation.
JACOB J. ADAMS (S’05–M’11–SM’17) received
the B.S. and M.S. degrees in electrical engineering
from The Ohio State University, in 2005 and 2007,
respectively, and the Ph.D. degree from the Univer-
sity of Illinois at Urbana–Champaign, in 2011.
From 2011 to 2012, he was an Intelligence
Community Postdoctoral Research Associate with
the University of Illinois at Urbana–Champaign.
Since 2013, he has been an Assistant Professor
with the Department of Electrical and Computer
Engineering, North Carolina State University. His research interests include
electrically small and reconfigurable antennas, characteristic mode theory,
and novel materials and fabrication methods for microwave devices.
Dr. Adams is a member of Tau Beta Pi and Eta Kappa Nu. He has
received several national awards, including the DARPA Young Faculty
Award, the U.S. Army Research Office Young Investigator Award, and the
NSF Graduate Research Fellowship.
30219