X04408122125

Lakshmy Sukumaran et al Int. Journal of Engineering Research and Applications www.ijera.com
ISSN : 2248-9622, Vol. 4, Issue 4( Version 8), April 2014, pp.122-125
www.ijera.com 122 | P a g e
Bit Error Rate Testers-A Study
Lakshmy Sukumaran, Dharani. K.G
Student, MVJ College of Engineering Bangalore
Assistant Professor, MVJ College of Engineering Bangalore
Abstract
This paper is a comparative study of the various BERT (bit error rate testers), which are used widely for
measuring the performance of any digital communication systems.
Keywords—BERT; Monte Carlo (MC); IS (importance sampling); software based BERT; hardware based
BERT; FPGA based BERT;
I. INTRODUCTION
Any digital transmission system which
transmits a series of bits over a communication
channel is likely to introduce some errors due to
various factors like noise, interference etc. We need
to ensure that these errors are reduced to a minimum
so as not to interfere with the working of the
communication system and maintain its data
integrity. To ensure such behaviour it is essential that
we carry out specific measurements of BER on the
test system. BER is a measure of how effective the
system is and indicates the number of bit error
introduced in the received message. BER Testers are
used to measure this value and this paper provides a
comparative study of the various types of BER
Testers available.
II. WHAT IS BER?
With the increasing demand for short design
cycles in the rapidly expanding field of wireless
digital communications, developing a reliable system
scalable to various emerging new standards and
technology is quite a challenge. One way to
accomplish this is to reduce the testing time and to
get the testing done at an earlier stage of design so as
to minimize the amount of reworks that would be
required later. So we go for testing at the earliest
possible stage, PHY layer or baseband layer. One
such widely used performance metric at the physical
layer which effectively identifies the reliability of a
wireless communication link (right from bits input to
the bits out through the entire antenna sub system,
signal path and channel) is Bit Error Rate. In simple
terms, it is defined as the number of bits in error to
the total number of transmitted bits, measured over
the fixed interval of time. A strong signal with an
undisturbed path will result in a very negligible value
of BER. Thus BER becomes significant when the
signal traverses through a rapidly changing disturbing
channel. The BER analysis is usually done
theoretically using the mathematical analysis.[1],[2].
Although the concept of determining the bit
error rate-sending a known sequence through the
channel and comparing the received signal against the
known sample to keep a count of the errors- is
simple, the execution suffers a setback. One method
is the pseudorandom method where we use the
pseudorandom data sequence for transmission. We
call this pseudo since it is not exactly possible to
model a random data transmission over a channel,
accurately and completely using the mathematical
models. Another one is the practical approach
wherein we need to install the complete
communication systems we intend to test, including
the antenna, entire radio, in the actual or laboratory
created environment. It is a highly impractical and
time consuming costly solution.[2]
There are mainly two methods employed to
measure the bit error rate. One is the software based
simulations like Monte Carlo simulations (MC) or
Importance sampling (IS).The other one is the
hardware based prototyping. [3]
III. SOFTWARE BASED
SIMULATIONS
Although in theory it is possible to obtain a
closed form expression for the bit error rate using the
various available mathematical analyses, it is
practically impossible. A more pragmatic approach
needs to be taken which can be applied even when the
analytical description of the system is unavailable.
A) Monte Carlo(MC)
One such practical approach is the software
based Monte Carlo simulations. Monte Carlo
simulations ,detecting an error is treated a success.
For a range of input sequences defined by probability
distribution functions, MC estimator has to estimate
the probability of error. [3]
RESEARCH ARTICLE OPEN ACCESS

B) The Importance Sampling (IS):
C) The IS approach is an improvement over the
MC process in such a way to alter the
properties of statistical processes involved in
the MC technique to the point of improving the
chances of success. It employs the variance
reduction technique by the use of a unbiased
estimator which modifies the original
distribution and hence reduce the variance of
the BER estimate calculated for a range of
inputs.[3]
The various problems encountered with the
software based simulations are given in the following
section.
 If the noise introduced by the channel is
extremely less, the errors introduced are less and
hence it can be considered as a reliable
transmission. So a larger time must be required
for the simulations in order to receive a desired
level of BER set. That indicates the time for the
simulations depends on the set BER that we need
to verify implying that the time needed for the
BER estimate will be a function of BER itself.
For example, the MC simulation requires at least
100/Pe experiments (where Pe is the desired
BER) to yield an estimator with relative
precision of 10% This makes the MC simulation
method highly inappropriate for systems with
very low BER.[4]
 The performance of a wireless system depends a
great deal on the characteristics of the radio
channel .In the laboratory test environment, these
channels are replicated using fading channel
simulators. The disadvantages of this equipment
is that it is very costly in terms of half a million
dollars or so.It cannot however be scaled up to
provide the requirements of emerging new
wireless systems.The commercially available
fading channel simulators along with some of
their chracteristics have been given in table1
below.[1],[5]
TABLE I. COMMERCIALLY AVAILABLE
FADING CHANNEL SIMULATORS
 For complex communication systems with
computationally intensive systems like the
MIMO, the software based MC simulations of
baseband layer is extremely time consuming and
costly. As an example, a wireless system should
perform over various radio channel characteristics
like AWGN, fading channel, indoor or outdoor
environments etc. Also all the available operating
modes should also be tested like the 300odd
modulation schemes and coding available in
802.11 standards etc. Thus a wide variety of
possible configurations need to be explored
before arriving at the optimum one, for systems
like MIMO. This makes the software simulations
a computationally daunting process, thus
providing a hindrance to timely design
verification.[1]
 The BER measurement of wireless systems over
different channels differs. It is a relatively
simpler process to do MC based simulation for
BER analysis on a AWGN channel compared to
the fading channels, where the error performance
needs to be averaged not only over independent
instances of noise and data (like we do for a
AWGN channel )but over fading channel
samples as well, over a longer duration.[14]
 The bit error rate of any reliable communication
link should be very low. For example in the
IEEE 802.3 10 GB/sec Ethernet , it should be of
Fading
channel
simulators
No of
chann
els
No of
paths
per
channe
l
Maxi
mum
Dopp
ler
(HZ)
Maxim
um
Delay
spread(
msec)
Baseb
and
bandw
idth
(Mhz)
ARC
Smartsim
[6]
64 36 2777 - 20
ACE 400
NB[7]
32 18 6 0.0012 20
PROP sim
C8[8]
16 48 2200
0
6.4 50
N5115A
[9]
2 48 2400 10 15
R&S
ABFS
[10]
2 12 1600 1.6 5
NJB-
1600B
[11]
2 12 800 0.2 10
SR_550
[12]
2 24 2000 2 13
Sofi05
[13]
2 24 4000 25 8

the order of 10^(-12) for it to be termed as
reliable. This indicates that a very large number
of symbols need to be processed for software
based simulators, over different SNR ratios to
obtain a reliable performance measurement .Thus
the simulation time can get extended to months
.The time required to estimate the BER reliably
is directly proportional to the target minimum
number of error at each operating point and
inversely proportional to erfc (10 ^ (Eb /N0)).[15]
T= [2Ne]/[ Rb* erfc (10^ {Eb/20 *No}] …….(1)
Where
Ne is the target minimum number of error at an
operating point.
Rb-number of samples per second, i.e. the bit rate
erfc- complementary Gaussian error function integral
(Eb /N0)- signal to noise ratio in db.
No - The one sided spectral density.
Eb- Energy per received bit
From the above equation (1), it can be
inferred that the simulation time increases
exponentially with magnitude of the signal to noise
ratio.[15][16]
IV. HARDWARE BASED BERT
The accurate BER measurements can be
obtained by observing the performance in the test
field .This calls for a very extensive and expensive
test set up. Also carrying out the over the air
experiments with a large number of possible
configurations makes it a very difficult approach. So
we go for a more practical approach, hardware based
prototyping at the baseband.[1][16]
The hardware based simulators accelerate
the performance compared to software simulators.
But the available commercial test equipment has the
following drawbacks
 They are not scalable to keep pace with the
emerging new standards and technology, as new
configurations may need to be tested.
 The available BERT can be used only at a later
stage, typically at the RF stage after the
fabrication of the system.
 New custom interfaces may have to be designed
to connect the device under test to the BERT,
which could add to the design burden.
The available BER testers in the market has
been compiled and presented in the below sections.
 Tektronix offers a wide range of Bit Error Rate
Testers for bit error rate analysis of the device
under test thus reducing the product development
cycle.[17]
TABLE II. COMMERCIALLY
AVAILABLE BERT FROM
TEKTRONIX[17]
 Another bit error rate tester from DCB (Data
Communication for Business Inc) is the 120V ac
powered BT-1 RS-232 Bit Error Rate Test Set
BERT. The BT-1 is an excellent bench test or
field testing tool. Applications involves in
locating problems in polling systems, multiplexer
systems, carrier communications, wireless
systems, and more.[18]
 Picosecond Pulse Labs, which has been now
acquired by the Tektronix, has the new and
improved 32Gb/s PAM-4 BERT system.[19]
 Absolute Analysis Investigator™ Bit Error Rate
Tester (BERT) supports multiple speeds from
1Gbps to 4Gbps, and generates test patterns
for fiber channel, Ethernet, FPDP, and many
other physical layers.It can detect Bit error rates
upto 10^(-12) and supports 32 ports for
testing.[20]
 Another one is from LUCEO Technologies
providing the X-BERT and Parallel X and 14g
parallel BERT systems.[21]
Most of the published hardware based BERT
uses the FPGA (Field Programmable Gate Array) and
tools like SIMULINK to integrate various FEC
(forward error correction) blocks onto FPGA. The
logic of FPGA is held either in look up tables,
registers, multiplexers, distributed or blocks memory.
In addition to these FPGA have dedicated circuitry
for speedy calculations, high speed input output
circuitry and embedded processors. The memory
bandwidth of n FPGA is also highly commendable.
These two features of FPGA-high speed parallel ALU
and memory bandwidth- enables FPGA as a potential
candidate for implementing the various data path
algorithms. The FPGA usage limited by the lack of
familiarity with the hardware designs, FPGA and
languages like Verilog, HDL.[17]
Even though using the system tools like
SIMULINK eliminates the need for having a good
knowledge of the hardware, it poses some
disadvantages [15]
 The simulation library may have only limited
resources like various communication blocks. It
may not have blocks which could cater to new
emerging technology standards.
Product Series Maximum
bit
rate(Gb/sec)
Channels
BitAnalyser BA series 1.5-1.6 1
BERTscope BSA
series
8.5-28.6 1
Patternpro@PPGseries 12.5-40 1,2,4
Patternpro@PEDseries 32-40 1,2

 The available designs do not have the flexibility
in architecture for adding, removing a module.
 The Gaussian noise can be generated using
analog components, which lacks accuracy and
could be used only for specific circumstances.
The other method is the generation of
psuedonoise using the digital components, which
is more flexible and widely used. Most of the
published BERT utilizes an inaccurate model for
GNG (Gaussian Noise Generator), reducing the
reliability of BER measurements
So we go for a more complete solution offered
by implementing all of the digital communication
processing modules like the encoder, modulator,
decoder, demodulator along with the fading channel
and the AWGN channel onto a single FPGA,
providing a repeatable test environment. It has been
observed that the FPGA based BERT increases cost
effectiveness by reducing the testing cost than the
one done using commercial test equipment and
channel simulators. Also such FPGA based BERT
increases productivity as time required for these,
when compared with software based simulations, is
very less. It also provides flexibility.[15]
V. CONCLUSION
A comparative study of the various BER
Testers was performed and it is deferred that FPGA
based BERT is a step ahead than its contemporaries,
in terms of speed and flexibility. Hardware-
accelerated validation is essential to speed up the
characterization of computationally intensive and
rapidly evolving modern wireless communication
systems.
REFERENCES
[1] A. Alimohammad, S. F. Fard, and B. F.
Cockburn, “FPGA-based accelerator for the
verification of leading-edge wireless systems,” in
Proc. IEEE Int. DAC, Jul. 2009, pp. 844–847.
[2] Breed, G. (2003, January). Bit Error Rate:
Fundamental concepts and measurement issues.
Retrieved April 07, 2014, from high frequency
electronics: http://www.high
frequencyelectronics.com/Archives/Jan03/HFE01
03_Tutorial.pdf[
[3] Bohdanowicz, A. (n.d.). On Efficient BER
Evaluation of Digital Communication Systems
via Importance Sampling. Retrieved April 07,
2014, from CiteSeerx:
http://citeseerx.ist.psu.edu/viewdoc/download?do
i=10.1.1.118.2785&rep=rep1&type=pdf
[4] Peter J. Smith, Mansoor Sha_ and Hongsheng
Gao, Quick sim-ulation: A review of importance
sampling techniques in communication systems,"
IEEE Journal on Selected Areas in
Communications, vol. 15, pp. 597{613, May
1997.
Cockburn, “FPGA-accelerated baseband design
and verification of broadband MIMO wireless
systems,”in Proc. IEEE 1st Int. Conf. Adv. Syst.
Testing Validation, Sep. 2009,pp. 135–140.
[6] ARC Seibersdorf Research GmbH, Vienna,
Austria.MIMO Channel Simulator, 2007.
[7] Azimuth Systems Inc., Acton, MA, USA. ACE
400NB MIMO Channel Emulator, 2006.
[8] Elektrobit Inc., Oulunsalo, Finland. Multi-
Channel Emulator, EB Propsim C8, 2008.
[9] Technical Overview, N5115A, Agilent
Technologies Inc. Baseband Studio for Fading,
2005.
[10] Data sheet, Rohde & Schwarz.
[11] Baseband FadingSimulator ABFS, Saving costs
through baseband simulation, 2004.Product
Manual, NJZ-1600B, Japan Radio Co.Multipath
Fading Simulator, 2005.
[12] Technical Document, SR5500, Spirent
Communications. Wireless Channel Emulator,
2006.
[13] Sofimation, Helsinki, Finland. SOFI05 Radio
Channel Simulator, 2006.
[14] A. Alimohammad, S. F. Fard,. “FPGA-Based Bit
Error Rate Performance Measurement of Wireless
Systems”in. IEEE Transactions on VLSI systems.
DOI10.1109/TVLSI.2013.2276010
Cockburn, “Hardware-based error rate testing of
digital baseband communication systems,” in
Proc.IEEE ITC, Oct. 2008, pp. 1–10.
[16] V. Singh, A. Root, E. Hemphill, N. Shirazi, and J.
Hwang, “Accelerating bit error rate testing using
a system level design tool,” in Proc. 11th
Annu.
IEEE Symp. FCCM, Apr. 2003, pp. 62–68.
[17] Inc, T. (n.d.). Bit error rate tester. Retrieved April
07, 2014, from tektronix inc:
http://in.tek.com/bit-error-rate-tester
[18] CB. (n.d.). Datasheet. Retrieved April 07, 2014,
from DCB:
http://www.dcbnet.com/datasheet/bt1.html?gclid
=CKKw-4jdxr0CFcwipQodgWIA_g
[19] Picosecond Pulse Labs, I. (2013, August).
PICOSECOND PULSE LABS ANNOUNCES
NEW 32GB/S PAM-4 BERT SYSTEM.
Retrieved April 07, 2014, from Picosecond Pulse
Labs:
http://www.picosecond.com/about/content.asp?sp
id=133
[20] Absolute Analysis, I. (n.d.). Bit error rate tester.
Retrieved April 7, 2014, from Absolute analysis:
ttp://www.absoluteanalysis.com/solutions/by-
products/bit-error-rate-tester.html
[21] Inc, L. (n.d.). 14G single channel and parallel
testing for 16GFC, Infiniband FDR, QSFP, AOC,
CFP, CXP and more. Retrieved April 07, 2014,
from Luceotec: http://www.luceotec.com/

X04408122125

More Related Content

What's hot

Viewers also liked

Similar to X04408122125

Recently uploaded

X04408122125