Jorge Cervera

Signal Processing and Data Transmission using a FPGA

SIGNAL PROCESSING AND

DATA TRANSMISSION USING
A FPGA

Autor: Jorge Cervera

Tutors: Baran Crkl.
University: Mlardalens Hogskola (Vsters, SWEDEN).
Home University: Escuela Universitaria de La Almunia de Doa Godina (Zaragoza,
Espaa)
Date: Spring Semester 2008.
1

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

ABSTRACT:
This thesis work has three main parts: the first one is to implement the hardware design
of a FPGA capable of reading in parallel 64 digital signals and process them. The
second part is about implementing a component that allows a PC machine to interact
with a FPGA. In the third part, both components will be assembled in the final design.
First of all, an overview of the FPGA technologies, and its main developers and
suppliers is given to provide a major understanding of how this technology works. Then
the two main HDL languages and their characteristics are explained. The design tools
used to develop the project are then showed before the explanation of the development
of this first part of the project;
After that, in the first part of the project a suitable design to process a digital signal
generated by a square array of 88 ITO electrodes photo lithographically defined in a
0.5_m thick photoresist layer previously filtered amplified and converted to digital
electronically is going to be implemented step by step.
In the second part is about an attempt to create an interface between the FPGA and a PC
machine. This interface should be able to order the FPGA to record data and then
send them to the PC.
As a final part, those two components will be assembled and tested to give the final
design, a design capable of reading, processing and sending the signals received from
the electrodes to a PC machine. Once this design is finished and tested, the
characteristics that the FPGA needs to have to perform all the needed functions will be
known.
The main purpose is to find a FPGA that can manage to process 64 signals in parallel so
all the tests and results can be taken in the same time lapse. Rather than focusing on
hardware programming, finding a FPGA fast enough and with the required amount of
memory and pins is the aim of the document. The results of a simple test run and
simulated in each device is shown, besides of the description of the implementation of
the interface.

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Acknowledgements:
First acknowledgements are for all my family, specially to my mother M Dolores,
because thanks to you and to your effort I am where I am and I am what I am. To my
sister Mar you made my childhood a wonderful time and for being my best friend now.
To my sisters Nuria and Idoya, mothers and sisters at one time, thank you for your
patience. To my niece Sara, because with less than one year you have taught me to
fight. To Nadia and Sheila, my crazy nieces. And to my father, because you tried to
show me what I didnt want to see, thank you for teaching me what is the bad way,
although you are not here.
I dont want to forget my friends: All my friends from La Salle, someday we will pass
through the adolescence and well be adults. To the people from the Ludoteque that
taught me that Sparta does not retreat. And to all my friends in Sweden, the love
brigade, the fellowship of the ring, the north troglodytes, the troll women and the girls
from Madrid, rebro, Vsters or Ibarrekolanda.
Thanks to all of you for supporting me and being there.

Agradecimientos:
Primero quiero agradecrselo a toda mi familia, en especial a mi madre, M Dolores,
porque gracias a ti y a tu esfuerzo estoy donde estoy y soy quien soy. A mi hermana
Mara por hacer de mi infancia una poca maravillosa y por ser ahora mi mejor amiga.
A mis hermanas Nuria e Idoya, que fuisteis madres y hermanas a un tiempo, gracias por
vuestra paciencia. A mi sobrina Sara, porque con menos de un ao tu me has enseado
lo que es luchar. A Nadia y Sheila, mis sobrinas que estn aun mas locas que yo.Y a mi
padre, por intentar hacerme ver lo que yo no quera, gracias por ensearme cual es el
mal camino, aunque ya no ests.
Tampoco quiero olvidar a mis amigos: todos los del colegio la Salle, algn da
superaremos la pubertad y nos haremos adultos. A la gente de la Ludoteque, que me
ense que Esparta no se retira. Y a todos mis amigos en Suecia, la brigada del amor, la
comunidad del Anillo, los trogloditas del norte, las mujeres troll y sin olvidarme de las
chicas de Madrid, rebro, Vsteras o Ibarrekolanda.
Gracias a todos por apoyarme y estar ah.

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Index
1. Background ...8
2. FPGAS ......8
2.1. Historical Overview ...8
2.2. Basic Functioning ......9
2.3. Developers ...10
2.3.1. Altera .....10
2.3.2 Xilinx .....11
3. Hardware Description Languages ......11
3.1. VHDL ......12
3.1.1.Program Structure ......13
3.2. Verilog .....14
4. VHDL/Verilog Compared & Contrasted ....15
5. Hardware Design and Simulating Tools .....15
5.1. Alteras Quartus II 7.2sp3 Web Edition ......15
5.2. Xilinxs ISE Design Suite 10.1. ...16
5.3. ModelSim PE Student Edition 6.3.c ....16
5.4. Choosing a Design Tool ..16
6. Developing of the design ....16
6.1. First Part: Design of a system capable of receiving, processing and sending
signals in a FPGA ...17
6.1.1. First Design ...17
6.1.2. Second Design ......19
6.1.3. Third Design .21
6.1.4. Fourth Design 23
6.1.5. Fifth Design ......27
6.2. Second Part: Design of a system capable of interact with a PC machine29
4

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

6.2.1. Hardware ..30

6.2.2. Transmitter Device ...31
6.2.3. Receiver Device 36
6.3. Third Part: Assembling of the three components 41
6.4. Final Test .46
7. Conclussions ...48
7.1. Advantages ...49
7.2. Drawbacks ...49
7.3. Future Work .49
8. References ..50
Appendix A: The util package 53
Appendix B: Time Planning ...54

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

List of Figures
Figure 1: Basic Overview .................................................................................................8
Figure 2: Top FPGA 2007 suppliers ................................................................................9
Figure 3: Logic Cell .........................................................................................................9
Figure 4: Interconnection of Logic Blocks .....................................................................10
Figure 5: VHDL entities and architectures ....................................................................13
Figure 6: Verilog modules ..............................................................................................14
Figure 7: First Design ....................................................................................................18
Figure 8: First Design Code ..........................................................................................18
Figure 9: First Design Test ............................................................................................19
Figure 10: Second Design ..............................................................................................19
Figure 11: Second Design Code 1st Part .......................................................................20
Figure 12: Second Design Code 2nd Part ......................................................................20
Figure 13: Second Design Test ......................................................................................21
Figure 14: Third Design .................................................................................................22
Figure 15: Third Design Code .......................................................................................22
Figure 16: Third Design Test .........................................................................................23
Figure 17: Fourth Design................................................................................................23
Figure 18: Fourth Design Code Part 1...........................................................................24
Figure 19: Fourth Design Table of Signals ...................................................................24
Figure 20: Fourth Design Code Part 2 ..........................................................................25
Figure 21: Fourth Design Code Part 3 ..........................................................................26
Figure 22: Fourth Design Test .......................................................................................26
Figure 23: Fifth Design ..................................................................................................27
Figure 24: Fifth Design Code Part 1 .............................................................................27
Figure 25: Fifth Design Code Part 2 .............................................................................28
Figure 26: Fifth Design Test ..........................................................................................29
Figure 27: Male RS-232 plug .........................................................................................30
Figure 28: Female to Female DB-9 Cable .....................................................................30
Figure 29: Transmitter entity .........................................................................................31
Figure 30: Transmitter entity, constants and signals .....................................................32
Figure 31: Transmitter Table of Signals ........................................................................32
Figure 32: Transmitter Baud Generator ........................................................................33
Figure 33: Serializer State Machine ..............................................................................34
Figure 34: Multiplexor ...................................................................................................35
Figure 35: Transmitter Test ...........................................................................................36
Figure 36: Receiver Entity .............................................................................................37
Figure 37: Receiver Entity, Constants and Signals ........................................................37
Figure 38: Receiver Table of Signals .............................................................................37
Figure 39: Receiver Baud Generator .............................................................................38
Figure 40: D-Flip-Flop ..................................................................................................39
Figure 41: Received Data Filter ....................................................................................40
Figure 42: Receiver State Machine ................................................................................40
Figure 43: Receiver Shift Register .................................................................................40
Figure 44: Receiver Test ................................................................................................41
Figure 45: Final Design .................................................................................................42
Figure 46: Processor Entity ...........................................................................................42
Figure 47: Component Transmitter ................................................................................43
6

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Figure 48: Component Receiver .....................................................................................43

Figure 49: Port Mapping ...............................................................................................43
Figure 50: Counter Signal with PC Start Order ............................................................44
Figure 51: Processing signals with PC start order ........................................................44
Figure 52: Stop Signal PC Order ...................................................................................44
Figure 53: Final Test, Record Code Received ...............................................................46
Figure 54: Final Test, Signal not Valid ..........................................................................47
Figure 55: Final Test, Valid Signal Recorded ...............................................................47
Figure 56: Final Test, Transmission Finished ...............................................................48
Figure 57: Package Util User Defined Types ................................................................51
Figure 58: Package Util Code .......................................................................................52
Figure 59: Time Planning...............................................................................................53

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

1. BACKGROUND:
Reading signals from neural cultures is a very complicated and delicate work. The
signals that must be analyzed are really small, and the time that these signals can be
recorded is really small if it is compared with the time that takes to prepare the culture.
It takes about 6 months to prepare the culture and it can be analyzed during less than 5
minutes before the culture becomes useless. So a really accurate and fast prototype
needs to be developed to perform all the analysis required in the minimum amount of
time possible. Some other devices are being used to perform this tests, but they are quite
expensive and they dont perform all the required tests, wasting a lot of time in further
analysis than has to be done to achieve the desired results.
With this thesis work, the aim is to give the basics to develop a new prototype based on
the FPGA technology that can perform all the required tests and can be easily controlled
by a simple PC machine. These basic functions are as simpe as receiving one signal and
evaluate the time that it lasts, and if it is a valid signal, convert and send it to the PC
machine to perform further analysis.

FPGA

PC

Figure 1: Basic Overview

2. FPGAs:
2.1. Historical Overview:
The FPGA are the result of the convergence of two different technologies, the logical
programmable devices (PLDs [Programmable Logic Devices]) and the integrated
circuits of specific application (ASIC). The history of the PLDs began with the first
devices PROM (Programmable Read-Only Memory) and they versatility was added by
the PAL (Programmable Array Logic) that allowed a major number of income and the
incorporation of records. These devices have continued growing in size and power
while, the ASIC always have been powerful devices, but its use has needed traditionally
8

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

a considerable investment of time and money. Attempts of reducing this load have come
from the modularization of the elements of the circuits, as the ASIC based on cells, and
from the standardization of the masks, as Ferranti was pioneering with the ULA
(Uncommitted Logic Array). The final step was to combine both strategies with a
mechanism of interconnection that could be programmed using fuses, antifuses or RAM
cells, as the innovative devices Xilinx of middle of the 80. The resultant circuits are
similar in capacity and applications to the biggest PLDs, though there are punctual
differences that aim to different origins. Besides computation reconfigurable, the
FPGAs are in use in controllers, coders/decoders, VLSI circuits prototyping.
The first manufacturer of these devices was Xilinx and Xilinx's devices are still the
most popular devices. Other FPGA suppliers are Atmel, Actel, Altera and Lattice
Semiconductor.

Figure 2: Top FPGA 2007 suppliers

2.2. Basic Functioning:

To build a FPGA, a basic "logic-cell" is duplicated hundreds or thousands of times.
Basically this cells is a small lookup table ("LUT"), a D-flipflop and a 2-to-1 mux (to
bypass the flipflop if desired).

Figure 3: Logic cell

9

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

A hierarchy of programmable interconnections allows the logical blocks of a FPGA to

be interconnected according to the need of the designer of the system, something similar
to a programmable breadboard. These logical blocks and interconnections can be
programmed after the process of manufacture by the user / designer, so the FPGA can
perform any logical necessary function.

Figure 4: Interconnection of logic blocks

A recent trend has been to combine the logical blocks and interconnections of the FPGA
with microprocessors and peripheral related to form a programmable System in a chip
". Example of such hybrid technologies they can be found in the devices Virtex-II PRO
and Virtex-4 of Xilinx, which include one or more processors PowerPC absorbed
together with the logic of the FPGA. Atmel's FPSLIC is another similar device, which
uses a processor AVR in combination with Atmel's logical programmable architecture.
Another alternative is use cores of implemented processors using the logic of the FPGA.
These cores include the processors MicroBlaze and PicoBlaze de Xlinx, Nios and Nios
II of Altera, and the processors of opened code LatticeMicro32 and LatticeMicro8.
Many modern FPGA support the partial reconfiguration of the system, allowing a part
of the design to be rescheduled, while other parts continue working. This one is the
principle of the computation reconfigurable computation idea.

2.3. Developers:
2.3.1. Altera:
Altera Corporation is the pioneer of programmable logic solutions. Today, Altera offers
FPGAs, CPLDs, and structured ASICs in combination with software tools, intellectual
property, and customer support to provide high-value programmable solutions. Altera
is headquartered in San Jose, California, USA.
10

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Alteras devices can be divided in two main families: The Cyclone family and the
Stratix family. The main difference between these two families is that Stratix FPGAs
uses ALUTs instead of normal LUTs. Altera estimates than an ALUT to be equivalent
to 2.5 LUTs. This makes Stratix FPGAs much better but more expensive that Cyclone
ones.
2.3.2. Xilinx:
One of the worlds largest developers of FPGAs is Xilinx, Inc. It was founded in 1984
and was based in Silicon Valley. Today, their HQ resides in San Jose, California (U.S.);
the European HQ is based in Dublin (Ireland) and the Asia Pacific/Japan HQ is based in
Singapore.
Xilinx FPGAs can be divided in two main families: The Virtex family and the Spartan 3
family. The Virtex family targets high performance applications that require higher
clock rates, extensive math operations, or significant use of single cycle timed loops.
The Spartan 3 family is intended for higher volume applications and as such has some
tradeoffs when it comes to the number of resources available on the chip as well as
over-all performance. The Spartan 3 resource mix works well for applications that
require more logic operations and rely less heavily on specialty resources

3. Hardware Description Languages:

In previous decades, most logic design was performed graphically, using block
diagrams and schematics. However, the rise of synthesizable HDLs coupled with
programmable logic devices and very-large-scale ASIC technology in the 1990s has
radically changed the way that typical digital designs are done.
In traditional software design, high-level programming languages like C, C++, and Java
have raised the level of abstraction so that programmers can design larger, more
complex systems with some sacrifice in performance compared to hand-tuned
assembly-language programs. The situation for hardware design is similar. The circuit
produced by a VHDL or Verilog synthesis tool may not be as small or fast as one
designed and tweaked by hand by an experienced designer, but in the right hands these
tools can support much larger system designs. This is, of course, a requirement if we are
ever to take advantage of the millions of gates offered by the most advanced CPLD,
FPGA, and ASIC technologies.
3.1. VHDL
VHDL was originally developed at the behest of the US Department of Defense in order
to document the behavior of the ASICs that supplier companies were including in
equipment. That is to say, VHDL was developed as an alternative to huge, complex
manuals which were subject to implementation-specific details.
11

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

The idea of being able to simulate this documentation was so obviously attractive that
logic simulators were developed that could read the VHDL files. The next step was the
development of logic synthesis tools that read the VHDL, and output a definition of the
physical implementation of the circuit. Modern synthesis tools can extract RAM,
counter, and arithmetic blocks out of the code, and implement them according to what
the user specifies. Thus, the same VHDL code could be synthesized differently for
lowest cost, highest power efficiency, highest speed, or other requirements.
VHDL borrows heavily from the Ada programming language in both concepts (for
example, the slice notation for indexing part of a one-dimensional array) and syntax.
VHDL has constructs to handle the parallelism inherent in hardware designs, but these
processes differ in syntax from the parallel tasks in Ada. Like Ada, VHDL is stronglytyped and is case insensitive. There are many features of VHDL which are not found in
Ada, such as an extended set of Boolean operators including nand and nor, in order to
represent directly operations which are common in hardware. VHDL also allows arrays
to be indexed in either direction (ascending or descending) because both conventions
are used in hardware, whereas Ada (like most programming languages) provides
ascending indexing only. The reason for the similarity between the two languages is that
the Department of Defense required as much as possible of the syntax to be based on
Ada, in order to avoid re-inventing concepts that had already been thoroughly tested in
the development of Ada.
VHDL is a fairly general-purpose language, although it requires a simulator on which to
run the code. It can read and write files on the host computer, so a VHDL program can
be written that generates another VHDL program to be incorporated in the design being
developed. Because of this general-purpose nature, it is possible to use VHDL to write a
testbench that verifies the functionality of the design using files on the host computer to
define stimuli, interacts with the user, and compares results with those expected. The
key advantage of VHDL when used for systems design is that it allows the behavior of
the required system to be described and simulated before synthesis tools translates the
design into real hardware.
Another benefit is that VHDL allows the description of a concurrent system (many
parts, each with its own sub-behavior, working together at the same time). VHDL is a
dataflow language, unlike procedural computing languages such as BASIC, C, and
assembly code, which all run sequentially, one instruction at a time.
A final point is that when a VHDL model is translated into the "gates and wires" that
are mapped onto a programmable logic device such as a CPLD or FPGA, and then it is
the actual hardware being configured, rather than the VHDL code being "executed" as if
on some form of a processor chip.

3.1.1. Program Structure:

VHDL was designed with principles of structured programming in mind, borrowing
ideas from the Pascal and Ada software programming languages. A key idea is to define
the interface of a hardware module while hiding its internal details. Thus, a VHDL
12

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

entity is simply a declaration of a modules inputs and outputs, while VHDL

architecture is a detailed description of the modules internal behavior or structure.
A VHDL entity declaration is a wrapper for the architecture, hiding the details of
whats inside while providing the hooks for other modules to use it. This forms the
basis for hierarchical system design-the architecture of a top-level entity may use, or
instantiate, other entities, while hiding the architectural details of lower-level entities
from the higher-level ones.

Figure 5: VHDL entities and architectures

3.2. Verilog
The designers of Verilog wanted a language with syntax similar to the C programming
language so that it would be familiar to engineers and readily accepted. The language is
case-sensitive, has a preprocessor like C, and the major control flow keywords, such as
"if" and "while", are similar. The formatting mechanism in the printing routines and
language operators and their precedence are also similar.
The language differs in some fundamental ways. Verilog uses Begin/End instead of
curly braces to define a block of code. The definition of constants in Verilog require a
bit width along with their base, consequently these differ. Verilog 95 and 2001 don't
have structures, pointers, or recursive subroutines; however SystemVerilog now
includes these capabilities. Finally, the concept of time so important to a HDL
won't be found in C.

13

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

The language differs from a conventional programming language in that the execution
of statements is not strictly sequential. A Verilog design consists of a hierarchy of
modules. Modules are defined with a set of input, output, and bidirectional ports.
Internally, a module contains a list of wires and registers. Concurrent and sequential
statements define the behavior of the module by defining the relationships between the
ports, wires, and registers. Sequential statements are placed inside a begin/end block
and executed in sequential order within the block. But all concurrent statements and all
begin/end blocks in the design are executed in parallel, qualifying Verilog as a Dataflow
language. A module can also contain one or more instances of another module to define
sub-behavior.
A subset of statements in the language is synthesizable. If the modules in a design
contain only synthesizable statements, software can be used to transform or synthesize
the design into a netlist that describes the basic components and connections to be
implemented in hardware. The netlist may then be transformed into, for example, a form
describing the standard cells of an integrated circuit (e.g. an ASIC) or a bitstream for a
programmable logic device (e.g. a FPGA).
3.2.1 Program Structure:
The basic unit of design and programming in Verilog is a module a text file containing
declarations and statements. A typical Verilog module corresponds to a single piece of
hardware, in much the same sense as a module in traditional hardware design. A
Verilog module has declarations that describe the names and types of the modules
inputs and outputs, as well as local signals, variables, constants, and functions that are
used strictly internally to the module, and are not visible outside. The rest of the module
contains statements that specify the operation of the modules outputs and internal
signals.

Figure 6: Verilog Modules

14

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

4. VHDL/Verilog Compared & contrasted

VHDL is a strongly typed language, and as a result is considered by some to be superior

to Verilog. VHDL proves to be specially useful in the first stages of the design, at the
moment of choosing architectures, partitions of the circuit, types of logic, etc and for the
accomplishment of simulations to functional level.
Also it allows the documentation of the project a level that is independent from the
technology and even from the type of logic, which allows its re-utilization in other
designs.
The project needs to be implemented and tested in a wide known programming
language, as this is only the beginning and there will be further development of it.
All software development tools that are going to be used to develop and test the
program , ModelSim, Alteras Quartus II and Xilinxs ISE, supports VHDL, and the
same code can be compiled and used in the three applications.

5. Hardware design and simulating tools:

Many design tools are available to implement the project. Here an overview of the three
main applications that can be used to develop the project is going to be given. In the
final part of the section, the reasons to choose ModelSim as the main developing tool
are given.
5.1. Alteras Quartus II 7.2sp3 Web Edition:
The Quartus II development software provides a complete design environment for
system-on-a-programmable-chip design.
The Quartus II software offers a graphical user interface complemented with an easy-touse online help system. The Quartus II system comprises an integrated design
environment that includes every step from design entry to device programming.
The user interface to the Quartus II software allows working with multiple files at the
same time, editing multiple design files to transfer information between them, while
simultaneously compiling or simulating another project. It is possible to view an entire
hierarchy of design files and move from one hierarchical level to another.
The Quartus II Compiler lies at the heart of the system, providing powerful design
processing that is customizable to achieve the best possible silicon implementation of
the project. Automatic error location and extensive documentation on error and warning
messages make design modifications as simple as possible.

15

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

5.2. Xilinxs ISE Design Suite 10.1:

ISE software offers a complete EDA flow: synthesis from Verilog or VHDL, place-androute (PAR), and creation of the bit files that are used to configure the chip. Xilinx's
Embedded Developer's Kit supports the embedded PowerPC 405 core, and the
Microblaze. Xilinx's System Generator for DSP simulates and implements highperformance DSP designs on Xilinx's FPGAs.
ISE Design Suite include PlanAhead Lite. PlanAhead Lite provides the I/O pin
planning capabilities of the PinAhead technology. It also includes design analysis and
floorplanning capabilities as well as implementation control with the ExploreAhead
environment.
5.3. ModelSim PE Student Edition 6.3.c.:
ModelSim PE offers VHDL, Verilog, or mixed-language simulation.
Model Technology enables transparent mixing of VHDL and Verilog in one design.
ModelSim architecture allows platform independent compile with the outstanding
performance of native compiled code.
An easy-to-use graphical user interface enables to quickly identify and debug problems,
aided by dynamically updated windows. These cross linked ModelSim windows create
a powerful easy-to-use debug environment.
ModelSim PE fully supports the VHDL and Verilog language standards. Behavioral,
RTL, and gate-level code can be simulated separately or simultaneously. ModelSim PE
also supports all ASIC and FPGA libraries, ensuring accurate timing simulations.
ModelSim PE provides initial support for VHDL 2002.
5.4 Choosing a Design Tool:
Altera and Xilinx software are really complete applications, very usefull and with many
features. Both of them and ModelSim are completely free for academic use. However,
bot Quartus II and ISE have a handicap: They are designed to work with the specific
models of Altera and Xilinx respectively. As the project is a prototype, and only the
design will be implemented and simulated, a tool that could work with any FPGA is
needed. For this reason, ModelSim is the tool chosen. For further development, once the
system resources were well defined, Quartus II or ISE should be selected instead of
ModelSim.

6. Developing of the design

The project is divided in three parts. In the first part the development of a system
capable of receiving and processing the signal according to the time requirements will
be developed. In the second part, the interface between the FPGA and the PC machine
will be investigated and developed, and in the third part the entire project will be
assembled to finish the application.
16

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

6.1. First Part: Designing of a system capable of receiving, processing and sending
signals in a FPGA
Requirements:
The FPGA must be able to perform the following actions:
Receiving 64 digital signals (1 bit) in parallel
Process this signal in parallel:
1. The signals must have a duration of at least 100 s
2. The signals must be received in a maximum time of 3.2 ms
The components must be implemented in the best way to make the physical
components as cheap as possible
This is a first prototype, so all the implementation is aim to develop all the previous
functions but only receiving one digital signal.
Real Design:
It has not been possible for me to test the system connecting it to the real signal
generator system. Thus, the system is implemented with only two input ports, hoping
that in the future real tests can be made. However the system is easily implemented and
it is easy to understand and to implement all the capabilities to fulfill the project
requirements.

Design:
The first thing that I was thinking about was about the needed of storing data in the
FPGA device. As the device is connected to a PC machine that has in comparison a
huge storing capacity, I resolved that it was much better if the data were evaluated in the
FPGA device and then sent to the PC to be stored. But this drives to another problem:
Working with high values inside a FPGA machine in parallel requires a lot of the FPGA
pins, and this take me to have to choose really expensive devices to develop a simple
design. But after many tests and designs I found the solution to the problem as it is
going to be explained.
6.1.1. First Design
As a first approach to the design I thought that the easiest way to develop the project
was to implement a system with the desired number of bit input ports and the same
number of 32 bits output ports that could send the signal to the PC machine, plus
another input port for the clock signal that was set in 100s. This output port contained
two important data: the position inside the array aimed to the time, and the value of each
position showed the value of the signals. A process activated by the clock signal record
the value of the signal in the array.

17

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

1in0

Recorder

32out0
Clk
FPGA

Figure 7: First design

Watching at the code it can be seen that the most important parts of the design are:
-

Entity processor: contains all the input and output ports.

Process clk_counter: stores the value of the input signal received through the in0
port, and increases the value of the counter signal with each clock tick.

Figure 8: First design code

However, two problems come with this design: Some valid signals were missed as it is
showed in Fig.9, and the number of pins that should be used for the final
implementation were too high; one port for the clock signal, 64 bit ports that use 64 pins
18

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

and 64 output ports that use 32 pins each one, giving a final number of pins equal to
2113, a very complex hardware design for such a simple operation.

Figure 9: First design test

Signal lost appeared as the biggest problem with this design, as all the system was
completely useless if some signals were not registered. To solve it, a new design was
implemented, trying to send value of the initial time of the signal and the value of the
end of the signal if it last for at least 100s.
6.1.2. Second design
This design kept one input pin for each signal, one input port for the clock signal, an
integer output port for the initial value of the signal and another integer output port for
the final value of the signal. Two processes were implemented. The first one just
increases a counter signal with each clock tick.

1in032init
recorder

1clk32fin

Figure 10: Second design

19

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Figure 11: Second design code 1st part

Figure 12: Second design code 2nd part

The second one measures the time every time that the input signal changes, evaluates it
and if it lasts for at least 100s, the initial value and the final value of the time are sent
to the PC machine. Every clock tick was adjusted to take 1 s, getting a more accurate
design.

20

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Figure 13: Second design Test

As Fig. 13 shows, with this new design no signal is lost and both times are registered
and sent if the signal is correct.
However, the problem about the number of pins used in the FPGA was not solved, but
raised. In the first design, 34 pins were used only to receive and process one signal, so
2113 pins were needed to implement the final system, and without an interface with a
PC machine. This second design needs 66 pins to receive, process and sending the
signal. That means that for the final design 64 input bit ports, one clock input port and
128 output integer ports (64 for the initial time and 64 for the final time) are going to be
needed, giving a total number of 4161 pins, a number of pins totally unacceptable.

6.1.3. Third design

In the previous design 32 bits are used to give an output to each value, and so too many
pins are used. However, the highest integer value that is going to be represented is 3200,
a number that can be represented with 12 bits. So as a possible solution, a function to
transform an integer into a 12 bit array was implemented inside the package util, which
it is going to be explained in the Appendix A.
With this improvement, each integer is represented with 12 bits instead of 32 bits, a new
model based in the second model was implemented.

21

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

1in012init
Recorder

1clk12fin

FPGA

Figure 14: Third design

Figure 15: Third design code

As it can be seen in the test, the design works in the same way than the second design,
only the input ports have changed to reduce the number of pins using the
int_to_std_logic transform function, giving two output ports of 12 bits each.
22

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Figure 16: Third design test

This was an improvement to the design, but not a solution, because in this third design
26 pins are utilized with just one input signal, so if 64 signals were going to be
processed, 1601 pins would have to be utilized, still a very high number.
6.1.4. Fourth Design
After this third design I thought that it could not be improved, as two integer values had
to be sent to the PC machine, and after implementing the int_to_std_logic transform
function to make the integer bit representation as small as possible I thought that
nothing else could be done. But then I realized that according to the project
requirements, the input signals had to be received and processed in parallel, but then
they could be stored in a data matrix and after sent to the PC machine sequentially,
through a bit port. As the maximum time of recording is 3.2 ms, and a signal has to last
a minimum time lapse of 100 s, that gives a maximum of 32 measures for the initial
time of the signal and 32 measures for the final time of the signal in the worst case, one
valid signal every 100 s.

1in0

Recorder
12out0
1clock
FPGA

Figure 17: Fourth design

23

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

This is an optimal design, and besides it makes easier to develop the second part of the
project as sending one bit with each clock tick is much easier than sending an array of
bits.

Figure 18: Fourth design code part 1

As it can be seen, this new design has only 3 bit input ports and one output bit port,
giving a total number of 3 pins to receive and process one signal. In a final design of 64
input ports, only 66 pins are going to be used, a really small number of pins that all the
FPGA models support.
Figure 19 gives an explanation of the types and purposes of the signals:
Name

Type

Purpose

Counter

Integer

Increments its value each clock tick

Recording Boolean

True, signal is being recoded.

False, signal is not being recorded.

Clk_init

Integer

Saves the initial value of counter when the value of in0=1

24

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Aux

Matrix

Saves the value of the initial and ending times of a valid signal.
The value of the rows indicates the number of the port.

Column

Integer

Value of the columns of the matrix

Aux1

Bit
vector

Saves the value of the bit vector that is being sent to the PC
machine

Row

Integer

Value of the rows of the matrix

Figure 19: Fourth design Table of signals

Figure 20: Fourth design code part 2

The basic processes keep doing the same functions than in the previous design, but now
the values of the initial time and of the final time of a valid signal are stored in the
signal aux after being converted in an 11 bit vector.

25

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Figure 21: Fourth design code part 3

This new process is the one that makes possible to send all the information bit by bit. It
is activated with each clock tick after the maximum time of recording has been
exceeded, and it begins to read the matrix aux, keeping the value in the signal aux1 that
is sent bit by bit through the port out0, beginning from the msb to give the exact values
of initial and final time. These values should be processed in the PC machine to
reconvert them into integer values for further use.

Figure 22: Fourth designn test

26

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

6.1.5. Fifth design

The fifth and final design contains the implementation of the design using two input
ports instead of one. Some user-defined data are utilized; they are implemented in the
package util, which is widely explained in the Appendix A.
1in0

1in1

12out0
Recorder

1clock
FPGA

Figure 23: Fifth design

Figure 24: Fifth design code part 1

The basic functioning is the same than in the previous design, only one process changes
a little bit, including a second input signal in the sensitivity list, and some signals
changes to store the values corresponding to the second signal data, as it can be seen in
Figures 24 and 25.

27

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Figure 25: Fifth design code part 2

The behavior of this fifth design is the same than for the fourth one, to send the data the
process begins to send the data corresponding to the port number 0, and after sending all
data of this port, it begins to send the data of the port number 1. The system could be
easily improved to perform the same functions for as many input ports as desired, but it
has not been done because of the lack of real tests with this first prototype.

28

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Figure 16: Fifth design test

Figure 26 shows how two valid signals are recorded (one in each input signal port) and
then sent bit by bit through the same output port.

6.2. Second Part: Designing of a system capable of interact with a PC machine:

In this second part of the project, a design capable of interact with a PC machine is
developed. The system is based on a interface through the RS-232 port of the PC and
the machine. This part is also divided into three sub-parts, to make the developing
process easier and more understandable. First, a brief explanation of the hardware
needed will be given. In the second part, a transmitter design will be explained
implemented and tested. The device is capable of sending data bit by bit. In the third
part, a design capable of receiving data from the pc is implemented. This device is
needed because the PC must be able to give some simple orders to the FPGA, such
start-record, or stop-record. Both devices use an "asynchronous" protocol. That means
that no clock signal is transmitted along the data. The receiver has to have a way to
"time" itself to the incoming data bits. In the case of RS-232, that's done this way:
1. Both side of the cable agree in advance on the communication parameters
(speed, format...). That's done manually before communication starts.
2. The transmitter sends a "1" when and as long as the line is idle.
3. The transmitter sends a "start" (a "0") before each byte transmitted, so that the
receiver can figure out that data is coming.
4. After the "start", data comes in the agreed speed and format, so the receiver can
interpret it.
5. The transmitter sends a "stop" (a "1") after each data byte.
29

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

6.2.1. Hardware:
RS-232 (Recommended Standard 232) is a standard for serial binary data signals
connecting between a DTE (Data terminal equipment) and a DCE (Data Circuitterminating Equipment). It is commonly used in computer serial ports.
An RS-232 interface has the following characteristics:

Uses a 9 pins connector "DB-9".

Allows bidirectional full-duplex communication (the PC can send and receive
data at the same time).
Can communicate at a maximum speed of roughly 10KBytes/s.

Figure 27: Male RS-232 plug

It has 9 pins, but the 3 important ones are:

pin 2: RxD (receive data).

pin 3: TxD (transmit data).
pin 5: GND (ground).

Using just 3 wires, you can send and receive data.

To connect the FPGA with a regular PC a j is needed.

Figure 28: Female to female DB-9 cable

30

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

6.2.2. Transmitter device:

The design of the transmitter was developed separately from the processor device to
make it easy to develop and test. The device has a baud generator to generate the baud
tick signal to transmit the signal, a state machine to serialize the signal and a
multiplexor to send it to the PC.
As a transmitter device for 12 bit arrays, it has a std_logic input port for the clock
signal, TxD_start std_logic input port to indicate that the device is ready to transmit,
TxD_data 12 bits input port that has the data that should be transmitted, and two bit
output ports, TxD to sent the data to the PC machine and TxD_busy to give a signal that
means that the device is busy.

TxD
Se
12TxD_data1
ria
1TxD_start1busy

liz

er
clk
FPGA

Figure 29: Transmitter Entity

31

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Figure 30: Transmitter Entity,Constants and Signals

Name

Type

Purpose

BaudGeneratorACC

Std_logic_vector

State

Std_logic_vector

TxD_ready

Std_logic

TxD_dataReg

Std_logic_vector

BaudTick

Std_logic

Muxbit

Std_logic

Baud generator accuracy,

changes its value when the
device is busy
Points to the state of the
device
1 means the device is
ready to transmit, 0
means it is busy
It is used to store the data
that is being sent to the PC
Signal to point each baud
tick (speed)
Aims to the beginning of a
new data
32

Jorge Cervera
busy

Signal Processing and Data Transmission using a FPGA

Std_logic

1 the device is busy, 0

the device is ready to
transmit

Figure 31: Transmitter Table of signals

Baud generator: As a first step I thought about the speed that is desired for the
transmission. The speed is specified in baud, i.e. how many bits-per-seconds can be
sent. For example, 1000 bauds would mean 1000 bits-per-seconds, or that each bit lasts
one millisecond.
Common implementations of the RS-232 interface (like the one used in PCs) don't
allow just any speed to be used. One of the following standard speeds has to be
selected:

200 bauds.
9600 bauds.
38400 bauds.
115200 bauds.

The device must be as fast as possible, and as RS-232 are not very fast actually, for
instance, at 115200 bauds, each bit lasts (1/115200) = 8.7s. If you transmit 12-bits
data, that lasts 12 x 8.7s = 104.4s. But each byte requires an extra start and stop bit,
so you actually need 14 x 8.7s = 121.8s. So the highest value for the speed is selected
to provide an acceptable time transfer value.
Here's a design with a 25MHz clock and a 16 bits accumulator. The design is
parameterized, so easy to customize.

Figure 32: Transmitter Baud Generator

State machine: The first state machine takes a 12-bits data signal, and serializes it
starting when the TxD_start signal is asserted. The "busy" signal is asserted while a
transmission occurs. The "TxD_start" signal is ignored during that time. The RS-232
parameters used are fixed: 12-bits data, 2 stop bits, no-parity. The state machine is
activated every clock tick.

33

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Figure 33: Serializer state machine

Notice how the state machine starts right when "TxD_start" is asserted, but then only
advances when "BaudTick" is asserted.
Multiplexor: The main function of the multiplexor is to generate the "TxD" output and
the start and stop signals. It is done through the state signal, every time it changes the
state is checked and the right signal is sent to the PC machine. It begins sending the
MSB and finishing with the LSB.

34

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Figure 34: Multiplexor

With all these parts the transmitter device is completed, and as it can be seen in the
simulation it should possible to send 12 bits chunks to a PC machine.
The test shows how data is sent bit by bit to the PC machine every time that the baud
tick is equal to 1. The beginning of the transmission begins with a 1 and ends with
two 1, to signal the beginning and the end of the data chunk.

35

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Figure 35: Transmitter test

6.2.3. Receiver device:

The design of the receiver was developed separately from the processor device to make
it easy to develop and test. A baud generator like in the transmitter device is used.
First, the incoming "RxD" signal has no relationship with our clock. Two D-flipflops
are used to oversample it, and synchronize it to our clock. The data are filtered, so that
short spikes on the RxD line aren't mistaken with start bits. Then a state machine allows
to go through each bit received, once a "start" is detected, and finally a shift register
collects the data bits as they come.
As a receiver, the device has one input std_logic port for the clock signal, the RxD
iniput port that receives the data sent by the PC, RxD_data_ready that is activated when
the data are synchronized and ready to be read, and RxD_endofpacket that indicates
when a packet has been read.

36

Jorge Cervera

RxD

Signal Processing and Data Transmission using a FPGA

De

se
data
ria

li
clk1data_ready
zer

FPGA

Figure 36: Receiver Entity

Figure 37: Receiver Entity, Constants and Signals

Name

Type

Purpose

Baud8GeneratorAcc Std_logic_vector Baud generator accuracy, changes its value

when the device is busy
Baud8Tick
Std_logic
Signal to point each baud tick (speed)
RxD_sync_inv

Std_logic_vector Signal used to synchronize (Flip- Flop)

37

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

RxD_cnt_inv

Std_logic_vector Signal used to synchronize, it is used as a

counter

RxD_bit_inv

Std_logic_vector Signal that changes the state of the device

State
Bit_spacing

Std_logic_vector Points to the state of the device

Std_logic_vector Signal used for timing when receiving data.

Next_bit

Std_logic

Indicates when a next bit can be received

RxD_data_error

Std_logic

Indicates when a there is an error, and no

transmission is made

RxD_d

Std_logic_vector It is used to store the data that is being

received by the FPGA

Index

Integer

Points to the position when the data is going

to be saved

Figure 38: Receiver Table of Signals

Baud Generator: The baud generator follows the same process than for the transmitter
machine, only the constant BaudGeneratorInc has changed because the timing of
receiving signals is higher than for transmitting, and after reading some papers it has
been set that this is the most adapted value.

Figure 39: Receiver Baud Generator

D- Flip-Flops: The flip-flops are implemented by the RxD_sync_inv, signal. The value
of the LSB bit of the signal and set it as the MSB, and then the value of the RxD port is
added. RxD is inverted, so that the idle becomes "0", to prevent a phantom character to
be received at startup

38

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Figure 40: D-Flip-Flop

The data are filtered, so that short spikes on the RxD line aren't mistaken with start bits.

Figure 41: Received data filter

State machine: The state machine goes through different states to point to the position
of the array signal where the data must be writen, once a start is detected. Notice that a
"next_bit" signal is used to go from bit to bit.

39

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Figure 42: Receiver state machine

Shift register: The shift register evaluates which are the conditions to write, and if so, it
sends the information through the RxD_d signal and eventually it sends the complete 8
bit array to the output port. The signal is only registered if the input signal is a 0 or a
1.

Figure 43: Receiver Shift register

40

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

With all these parts the transmitter device is completed, and as it can be seen in the
simulation it should possible to receive 8 bit chunks from a PC machine.

Figure 44: Receiver Test

6.3. Third Part: Assembling of the three components:

Once every component has been developed and simulated, it is time to assemble the
three of them. VHDL allows a very easy way to link components, so it has not been
very difficult to do it. Just a few modifications have been needed to make all the
components to work together.

41

Jorge Cervera

RxD

In01

Signal Processing and Data Transmission using a FPGA

Receiver
18data12data1TxD
Transmitter
Processor
1ready1start1busy

Clkclk
Clk
FPGA

Figure 45: Final design

The processor is the main component, and the receiver and the transmitter work as
components of it, so the main difference with the previous version of the processor is
that some ports have been added to it to make possible to connect it with the other two
components. The port that took the output signal (out0) has been substituted by the
PC_TxD_data port, that has the same function.

Figure 46: Processor entity

The transmitter and the receiver are declared as components of the main entity
processor. The PC_TxD_data, TxD_start and the PC_TxD_endofpacket ports of the
processor are declared as inout, as the first work as output ports for the transmitter, and
42

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

as input ports for the processor and the opposite for the other two. In the case of
PC_TxD_data, the processor sends the data corresponding to the processed signals
through this port to the transmitter to be sent to the PC. In the case of the other two, the
transmitter activates a signal in this port to indicate that a chunk of information has been
sent, and that another one can be sent. So it is an output port for the transmitter, and an
input port for the processor.

Figure 47: Component transmitter

Figure 48: Component receiver

The component receiver is exactly the same than the one explained in section 6.2.3, and
it works in the same way.
Port mapping: The port mapping can be seen in Fig. 49. Each port of the two
components corresponds with one port of the main entity processor. The three
components share the same clock signal.

Figure 49: Port mapping

Receiving orders from the PC: As it has been described in section 6.2.3, the signal from
the PC is received and processed by a R2-232 serial interface. The receiver receives the
signal bit by bit, and then deserializes 8 bits to get an 8 bit array. Two orders have been
implemented, record and stop recording. The bit sequences that activate these orders
are:
43

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

10000001 Record.
10000010 Stop recording.
When the system receives the Record signal sequence, it begins to receive the signals
and processing them. The counter signal is initialized and the process that measures the
timing of the signals begins to work.

Figure 50: Counter signal with PC start order

Figure 51: Processing signals with PC start order

44

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

When the system receives the Stop recording signal sequence, it stops recording and
sends all the information to the PC, the same action that the system takes when a
complete recording sequence has been complete.

Figure 52: Stop signal PC order

Watching at the code, some differences between the code in the fifth design of the
processor and the final design can be seen, and it is not only the if clause
corresponding to the PC signal to stop. In the fifth design, one value of the aux signal
was sent with each clock tick. But this is too fast for our transmitter that works sending
one bit each baud tick, so the variable has to keep the same value until all the data bits
plus the two stop bits are sent to the PC. This is made through the
PC_TxD_endofpacket port, that keeps a value of 0 until the last stop bit has been sent,
taking the value of 1 and letting the process to take other data value to send to the PC
machine. Besides, the port PC_TxD_Start is set to 1 to make the transmission of data
begin.
The final design requires the following resources:
3444 logic elements
30 pins
0 memory bits

45

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

The number of pins is still very low, for a 64 input port device we will need 92 pins, a
really small number. The device needs 3444 logic elements. More logic elements will
be used with a more complex device, however it is not a really high number for the
FPGA devices that can be found in the market. For instance, a Cyclone device from
Altera has about 250 pins and 4.000 logic elements, and this is the most cheap and
simple device from Altera.

6.4. Final Test:

Now it is time to test the whole application, testing each one of the functions.
Figure 53 shows how the counter signal does not begin to increase until the code for
recording has been received by the FPGA, and no other processes are activated.

Figure 53: Final Test, Record code received

In Fig.54, we can see how an impulse has been received in the in0 port, but as it lasts
less than 100 s, it is rejected and the values of the initial and the final time of the signal
are not recorded in aux(0,0) and aux(0,1).

46

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Figure 54: Final test, signal not valid.

In Fig. 55 we can see how later in the system a valid signal is processed and recorded.
The value of the initial time is recorded in aux(0,0) and the final value of the time is
recorded in aux(0,1).

Figure 55: Final test, valid signal recorded

47

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Now in case of another valid signal, the values of the initial and final time of the signal
would be recorded in aux (0,2) and aux (0,3). The same process using the same
algorithm is used for the signals that come from the input port in1, so no further tests
are going to be done of this algorithm because it is clear that it is working.
Now the stop record signal is given, so the system must start to transmit the recorded
data. As it can be seen in Fig.56, as soon as the complete code for the order Stop
recording is received, the counter signal stops, the port pc_txd_start is activated and the
transmission starts. The complete sequence of transmitted bits can be seen in the pc_txd
port and in the pc_txd_data port.

Figure 56: Final test, Transmission finished.

Note than the end of the transmission of each chunk of data is denoted with a 1 in the
pc_txd_endofpacket port.

7. Conclusions
Finally a system capable of processing signals coming from an external source,
receiving simple orders from a PC machine and transmit the processed data has been
developed. The purpose of this work is to serve as a prototype for future development
and improvements. As a first approach it has some advantages and many drawbacks.

48

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

7.1. Advantages:
All the project requirements are fulfilled. The system is capable of receiving a digital
signal and processing it, finding if it is a valid signal or not. The algorithms used to
implement the processing function are quite easy to understand, and they can be easily
improved to receive as many signals as desired.
The receiver and transmitter components provide serial asynchronous transmission of
data from and to a PC machine.
7.2. Drawbacks:
The algorithm used to process the signal works and can be used to process more signals
simply doing some copy-paste job and adjusting three parameters. However this is not a
good way of programming, as it takes to a lot of rubbish code and to the consumption
of many unneeded resources.
The major drawback of this system comes with the transmiting and receiving modules.
Looking at Figures 53 and 56 we can see why. One of the main reasons to choose the
FPGA technology to develop the project was that FPGAs are capable of working in
parallel, one of the first requirements of the project. This capability makes FPGAs very
fast devices, another requirement of the project, as timing is really important. Looking
at Fig. 53 we can see that it takes approximately 525 s to receive and process the code
to start recording from the PC. If a valid signal has to last a minimum of 100s, that
result on the possible loose of 6 valid signals. Besides if more orders were sent from the
PC, with 6 signals sent no record would be possible. One similar thing happens with the
transmitter device. Looking at Fig. 56 we see that it takes approximately 450s to
transmit one integer value. In the worst case analysis:
64ports x 64values = 4096 values must be sent
4096values x 450s = 1843200 s = 1.843200 s
4096values x 12bits = 49152 bits
This value could not be seen as a very high value, but a transmition that last seconds for
this relatively small amount of data is unacceptable with the actual technology.
The time delays must be avoided in the measure of possible for this project. Besides not
real tests have been made, and with them probably some offsets due to interferences and
noise will increase time delays. So, these two drawbacks make the interface created to
interact with the PC machine totally useless with our system.

49

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

7.3 Future Work:

This section has only the purpose of giving a guide of what I would do if I could
continue working with the project, and to help other people that maybe will continue
with it.
As I said before, the algorithm to process the signal works, but it is not good
programming. One algorithm could manage all the processing work for all the ports,
not one algorithm per each port. This could be make through some kind of loop, but
taking count of the increased logic units that will be needed.
My work with the interface with the PC has proved to be completely useless due to time
delays. Serial and asynchronous communication is the easiest way to implement an
interface, and as I am not an expert in VHDL, I took this as a first option. I knew that it
would be slow, but I hoped that the time delays would not be significant for the system.
Unfortunately tests have proved that this kind of interface is completely useless for the
project. For future improvements, a synchronous and parallel interface should be
investigated and implemented if possible. Synchronous transmissions take time at the
beginning during synchronization, but then the transmission is faster than if it is
asynchronous. And of course parallel transmission are much faster that serial, and
FPGAs support this kind of transmission as well as PC machines.
And as a final recommendation, real tests with the FPGA should be made. It is easy to
simulate a design and to make it work virtually, but in reality things are different. You
will never know if your design works only simulating tests.

8. References
1. http://en.wikipedia.org/wiki/FPGA, last access 4th April 2008
2. http://en.wikipedia.org/wiki/VHDL, last access 7th April 2008
3. http://en.wikipedia.org/wiki/Verilog, last access 7th April 2008
4. http://www.altera.com, last access 15th April 2008
5. http://www.xilinx.com, last access 10th April 2008
6. http://www.fpga4fun.com last access 28th May 2008
7. http://www.model.com, last access 12th April 2008
8. Digital Design Principles and Practices4th Edition, John F. Wakerly, Prentice
Hall.
9. VHDL 3rd Edition, Douglas Perry, McGraw-Hill.

50

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

10. John V. Oldfield and Richard C. Dorf Field Programmable Gate ArraysReconfigurable Logic for Rapid Prototyping and Implementation of Digital
Systems, Wiley Interscience, 1995.
11. Lennart Lindh, Tommy Klevin. Progammerbara kretsar, Utveckling av
inbyggda system.
12. Richard A. Blum, An Electronic System For Extracellular Neural Stimulation
And Recording, Georgia Institute of Technology August 2007.
13. Tobias Nyberg, Akiyoshi Shimada, Keiichi Torimitsu, Ion conducting polymer
microelectrodes for interfacing with neural networks, NTT Basic Research
Laboratories, NTT Corporation, Atsugi, Kanagawa 243-0198, Japan, 9 August
2006.
14. Yasuhiko Jimbo, Nahoko Kasai, Keiichi Torimitsu, Takashi Tateno, and Hugh
P. C. Robinson, A System for MEA-Based Multisite Stimulation, IEEE
Transactions on Biomedical Engineering, Vol.50, No.2, February 2003.
15. Douglas E. Ott and Thomas J. Wilderotter, "A Designer's Guide to VHDL
Synthesis", Kluwer Academic Publishers.
16.
17.

David Pellerin and Douglas Taylor, "VHDL Made Easy", Prentice Hall.
Charles E. Roth, Jr., "Digital System Design Using VHDL" , PWS Publishing
Company.

51

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Appendix A: The util package:

The util package is a package that contains some user defined types and constants and
the function used to transform an integer into a 12 bit array. It is really simple and
useful. All the future conversion functions and user defined types should be placed in
this package.
The only constant defined is max_ports, that is equal to the maximum number of input
signal ports of the processor minus 1, due to 0.
The following are the used defined types:
Name

Type

Bool_vector array (max_ports downto 0) of boolean

Int_vector

array(max_ports downto 0)of integer range 0 to 3200

Bus_2

array(max_ports downto 0) of std_logic_vector(11 downto 0)

Bus_6

array (5 downto 0)of bit_vector(11 downto 0)

matrix

array (max_ports downto 0,64 downto 0)of std_logic_vector(11

downto 0);
Figure 57: Package util user defined types

52

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Figure 58: Package util code

Appendix B: Time Planning

I had a very clear idea of how to divide the project from the beginning. One was a
processor, the other one was about creating an interface with the PC, and the last part
was to assemble all the parts. I had to spent a lot of time learning VHDL after choosing
it as the programming language of the projcec, as it was a new programming language
for me.
After that, programming the three parts took most of my time, being the problems with
the resources when working with the processor the main problem, and the problems of
synchronizing were a problem too when assembling. During all the time of developing,
I wrote this paper, so no important details were forgotten.

53

Jorge Cervera

Signal Processing and Data Transmission using a FPGA

Figure 59: Time planning

54