University of Pennsylvania
Department of Electrical Engineering

ABEL-HDL Primer

ABEL Primer Contents

1 Introduction

ABEL (Advanced Boolean Equation Language) allows you to enter behavior-like descriptions of a logic circuit. ABEL is an industry-standard hardware description language (HDL) that was developed by Data I/O Corporation for programmable logic devices (PLD). There are other hardware description languages such as VHDL and Verilog. ABEL is a simpler language than VHDL which is capable of describing systems of larger complexity.

ABEL can be used to describe the behavior of a system in a variety of forms, including logic equations, truth tables, and state diagrams using C-like statements. The ABEL compiler allows designs to be simulated and implemented into PLDs such as PALs, CPLDs and FPGAs.

Following is a brief overview of some of the features and syntax of ABEL. It is not intended to be a complete discussion of all its features. This ABEL primer will get you started with writing ABEL code. In case you are familiar with ABEL, this write-up can serve as a quick reference of the most often used commands. For more advanced features, please consult an ABEL manual or the Xilinx on-line documentation.

2. Basic structure of an ABEL source file

An ABEL source file consists of the following elements.

Keywords (words recognized by ABEL such as commands, e.g. goto, if, then, module, etc.) are not case sensitive. User-supplied names and labels (identifier) can be uppercase, lowercase or mixed-case, but are case-sensitive (input1 is different from Input1).

A typical template is given below.

The following source file is an example of a half adder:

A brief explanation of the statements follow. For a more detailed discussion see the following sections or consult an ABEL-HDL manual. When using ABEL with the Xilinx CAD software, you can use the ABEL wizard which will give a template of the basic structure and insert some of the keywords. Also the Language Assistant in the ABEL editor provides on-line help. Go to the TOOLS ->LANGUAGE ASSISTANT menu. The Language templates give a description of most ABEL commands, syntax, hierarchy, etc, while the Synthesis template gives examples of typical circuits.

3. Declarations

Module: each source files starts with a module statement followed by a module name (identifier). Large source files often consist of multiple modules with their own title, equations, end statement, etc.

Title: is optional and can be used to identify the project. The title name must be between single quotes. The title line is ignored by the compiler but is handy for documentation.

String: is a series of ASCII characters enclosed by single quotes. Strings are used for TITLE, OPTIONS statements, and in pin, node and attribute declarations.

device: this declaration is optional and associates a device identifier with a specific programmable logic device. The device statement must end with a semicolon. When you are using the Xilinx CAD system to compile the design, it is better not to put the device statement in the source file to keep your design independent of the device. When you create a new project in Xilinx you will specify the device type (can also be changed in the Project Manager window using the Project Information button). The format is as follows:

comments: comments can be inserted anywhere in the file and begin with a double quote and end with another double quote or the end of the line, whatever comes first.

pin: pin declarations tell the compiler which symbolic names are associated with the devices external pins. Format:

One can specify more than one pin per line:

You do not need to specify the pin. Pin numbers can be specified later by using a "user constraint file " when doing the compilation using Xilinx CAD. This has the advantage that our design is more general and flexible. The ! indicates an active low (the signal will be inverted). The istype is an optional attribute assignment for a pin such as 'com' to indicate that the output is a combinational signal or 'reg' for a clocked signal (registered with a flip flop). This attribute is only for output pins.

node: node declarations have the same format as the pin declaration. Nodes are internal signals which are not connected to external pins.

other declarations allows one to define constants, sets, macros and expressions which can simplify the program. As an example a constant declaration has the following format:

The last two equations are equivalent. The use of ".." is handy to specify a range. The last example makes use of vector notations. Any time you use D in an equation, it will refer to the vector [D3, D2, D1. D0].

4. Numbers

Numbers can be entered in four different bases: binary, octal, decimal and hexadecimal. The default base is decimal. Use one of the following symbols (upper or lower case allowed) to specify the base. When no symbol is specified it is assumed to be in the decimal base.

BASE NAME BASE Symbol.
Binary 2 ^b
Octal 8 ^o
Decimal 10 ^d (default)
Hexadecimal 16 ^h

Examples:

    Specified in ABEL

 Decimal Value

    35

    35

    ^h35

    53

    ^b101

    5

5. Sets

A set is a collection of signals or constants used to reference a group of signals by one name. A set is very handy to simplify logic expressions. Any operation that is applied to a set is applied to each element.

A set is a list of constants or signals separated by commas or the range operator (..) put between square brackets (required).

    [D0,D1,D2,D4,D5]

    [D0..D6]

 " incrementing range

    [b6..b0]

 " decrementing range

    [D7..D15]

    [b1,b2,a0..a3]

 " range within a larger set

    [!S7..!S0]

 "decrementing range of active -low signals

However, the following is not allowed: [D0, X];

a. Indexing or accessing a set

Indexing allows you to access elements within a set. Use numerical values to indicate the set index. The number refers to the bit position in the set starting with 0 for the least significant bit of the set. Here are some examples.

To access one element in the set, use the following syntax:

Here a comparator operator (==) is used to convert the single-element (X[2]) into a bit value equivalent to X2. The comparator (==) gives a"1" or "0" depending if the comparison is True or False. Notice the difference between the assignment operator (=) and the equal operator (==). The assignment operator is used in equations rather than in expressions. Equations assign the value of an expression to the output signals.

b. Set operations

Most of the operations can be applied to a set and are performed on each element of the set according to the rules of Boolean algebra. Operations are performed according to the operator's priority; operators with the same priority are performed from left to right (unless one uses parentheses). Here are a couple of examples.


Example 1:

Example 2:

this is equivalent to two statements:

Example 3:

Example 4:

which is equivalent to

However consider the following expression

now the number "2" is first converted into a binary representation and padded with zeros (0010) if necessary. Thus the above equation is equivalent to:

Example 5:

Example 6:

The number "2" is converted into binary and padded with zeros (0010).

Example 7: Sets are also handy to specify logic equations. Suppose you need to specify the equation:

This can be done using sets. First define a constant Addr set.:

One can then use the following equation to specify the address:

which is equivalent to saying:

Indeed, if A7=0, A6=1 and A5=1, the expression Addr ==[0,1,1] is true (or 1) and thus Chip_Sel will be true (or 1). Another way to write the same equation is:

The above expressions are very helpful when working with a large number of variables (ex. a 16 bit address).

Example 8:

For the same constants as in the example above, the expression,

which is equivalent to

However, the following statement is different:

which is equivalent to:

However, the relational operator (==) gives only one bit, so that the rest of the equation evaluates also to one bit, and the "3" is truncated to "1":. Thus the above equation is equal to:

6. Operators

There are four basic types of operators: logical, arithmetic, relational and assignment..

a. Logical Operators

The table below gives the logical operators. They are performed bit by bit. With the @ALTERNATIVE directive, one can use the alternative set of operators as indicated in the table.

Operator (default) Description Alternate operator
! NOT (ones complement) /
& AND *
# OR +
$ XOR: exclusive or :+:
!$ XNOR: exclusive nor :*:

b. Arithmetic operators

The table below gives the arithmetic operators. Note that the last four operators are not allowed with sets. The minus sign can have different meanings: used between two operands it indicates subtraction (or adding the twos complement), while used with one operator it indicates the twos complement.

Operator Example Description
- -D1 Twos complement (negation)
- C1-C2 Subtraction
+ A+B Addition

The following operators are not used with sets:

* A*B Multiplication
/ A/B Unsigned integer division
% A%B Modulus: remainder of A/B
<< A<<B Shift A left by B bits
>> A>>B Shift B right by B bits

c. Relational operators

These operators produce a Boolean value of True (-1) or False (0). The logical true value of -1 in twos complement is represented by all ones (i.e.all bits will be ones: ex. for a 16 bit word all bits are one: -1 is represented by 1111 1111 1111 1111).

Operator Example Description
== A==B or 3==5 (false) Equal
!= A!=B or 3 != 5 (true) Not equal
< A<B or 3 < 5 (true) Less than
<= A<=B or 3 <= 5 (true) Less than or equal
> A>B or -1 > 5 (true) Greater than
>= A>=B or !0 >= 5 (true) Greater than or equal


Relational operators are unsigned. Be careful: !0 is the one complement of 0 or 11111111 (8 bits data) which is 255 in unsigned binary. Thus !0 > 9 is true. The expression -1>5 is true for the same reason.

A relational expression can be used whenever a number can be used. The -1 or 0 will be substituted depending on the logical result. As an example :

A will be equal to B if C is equal to D (true or 11111...; B XNOR 1 equals B), otherwise, A will be equal to the complement of B (if C is not equal to B (false or 0)).

d. Assignment operators

These operators are used in equations to assign the value of an expression to output signals. There are two types of assignment operators: combinational and registered. In a combinational operator the assignment occurs immediately without any delay. The registered assignment occurs at the next clock pulse associated with the output. As an example, one can define a flip-flop with the following statements:

The first statement defines the Q1 flip-flop by using the 'reg' as istype (registered output). The second statement tells that the output of the flip-flop will take the value of the D input at the next clock transition.

Operator Description
= Combinational assignment
:= Registered assignment

e. Operator priority

The priority of each operator is given in the following table, with priority 1 the highest and 4 the lowest. Operators with the same priority are performed from left to right.

Priority Operator Description
1 - Negation (twos complement)
1 ! NOT
2 & AND
2 << shift left
2 >> shift right
2 * multiply
2 / unsigned division
2 % modulus
3 + add
3 - subtract
3 # OR
3 $ XOR
3 !$ XNOR
4 == equal
4 != not equal
4 < less then
4 <= less then or equal
4 > greater than
4 >= greater than or equal

7. Logic description

A logic design can be described in the following way.

a. Equations

Use the keyword equations to begin the logic descriptions. Equations specify logic expressions using the operators described above, or "When-Then-Else" statement.

The "When-Then-Else" statement is used in equations to describe a logic function. (Note: "If -Then-Else" is used in the State-diagram section to describe state progression).

The format of the "When-Then-Else" statement is as follows:

Examples of equations:

One can use the braces { } to group sections together in blocks. The text in a block can be on one line or span many lines. Blocks are used in equations, state diagrams and directives.

b. Truth Tables

The keyword is truth-table and the syntax is

or

or

in which "->" is for combinational output and ":>" for registered output. The first line of a truth table (between parentheses) defines the inputs and the output signals. The following lines gives the values of the inputs and outputs. Each line must end with a semicolon. The inputs and outputs can be single signals or sets. When sets are used as inputs or outputs, use the normal set notation, i.e. signals surrounded by square brackets and separated by commans. A don't care is represented by a ".X.".

Example 1: half adder

However, if one defines a set IN = [A,B]; and OUT = [Sum, Carry_out]; the truth table becomes simpler:

Example 2: An excluse OR with two intputs and one enable (EN). This example illustrates the use of don't cares (.X.)

Example 3: (see Example in R. Katz, section 7.2.1 and table 7.14)

Truth tables can also be used to define sequential machines. Lets implement a three-bit up counter which counts from 000, 001, to 111 and back to 000. Lets call QA, QB and QC the outputs of the flip-flops. In addition, we will generate an output OUT whenever the counter reaches the state 111. We will also reset the counter to the state 000 when the reset signal is high.

For the use of .DOT extensions (.CLK and .AR) see section 7d.

c. State Description

The State_diagram section contains the state description for the logic design. This section uses the State_diagram syntax and the "If-Then-Else", "Goto", "Case" and "With" statements. Usually one declares symbolic state names in the Declaration section, which makes reading the program often easier.

State declaration (in the declaration section) syntax:

As an example: SREG = [Q1, Q2]; associates the state name SREG with the state defined by Q1 and Q2.

The syntax for State_diagram is as follows:

The keyword state_diagram indicates the beginning of a state machine description.

The STATE keyword and following statements describe one state of the state diagram and includes a state value or symbolic state name, state transition statement and an optional output equation. In the above syntax,

If-Then-Else statement:

This statement is used in the state_diagram section to describe the next state and to specify mutually exclusive transition conditions.

Syntax:

In which state-exp can be a logic expression or a symbolic state name. Note that the "IF-Then-Else" statement can only be used in the state_diagram section (use the "When-If-Then" to describe logic functions". The ELSE clause is optional. The IF-Then-Else statements can be nexted with Goto, Case and With statements.

Example (after R. Katz):

in the declaration section we define first the state registers:

"If-Then-Else" statements can be nested as in the following example (after Wakerly). We assume that one has defined the registers and states in the declaration section.

"with" statement:

Syntax:

in which trans_stmt can be "If-then-else", 'Goto" or a "Case" statement.
state_exp: is the next state, and equation is an equation for the machine outputs.

This statement can be used with the "If-Then-Else", "Goto" or "Case" statements in place of a simple state expression. The "With" statement allows the output equations to be written in terms of transitions.

Example 1:

In the above example, the output Z will be asserted as soon as the expression after the if statement evaluates to a logic 1 (or TRUE). The expression after the "With" keyword can be an equation that will be evaluated as soon as the if condition is true as in example 2:

Example 2:

The "With" statement is also useful to describe output behavior of registered outputs, since registered outputs would lag by one clock cycle. It allows one also for instance to specify that a registered output should have a specific value after a particular transition. As an example [1],

Example 3[1]:

state S1:

Notice that one can use curly braces to control a group of outputs and equations after the With keyword as in the example above.

Example 3:

You have to be aware of the timing when using the "With " statement with combinational or asynchronous outputs (as in a Mealy machine). A Mealy machine changes its outputs as soon as the input changes. This may cause the output to change too quickly resulting in glitches. The outputs of a Mealy machine will be valid at the end of a state time (i.e. just before the clock transition). In this respect a Moore output (with synhronous outputs) is less prone to timing errors. An example of a Mealy machine and a Moore machine is available.

Case statement

Syntax:

expression is any valid ABEL expression and state_exp is an expression that indicates the next state (optionally followed by WITH statement).

Example:

The case statement is used to list a sequence of mutually-exclusive transition conditions and corresponding next states. The CASE statement conditions must be mutually exclusive (no two transition conditions can be true at the same time) or the resulting next state is unpredictable.

d. Dot extensions

One can use dot extensions to more precisely describe the behavior of the circuit. The signal extensions are very handy and provide a means to refer specifically to internal signals and nodes associated with a primary signal.

The syntax is

Some of the dot extensions are given in the following table. Extensions are not case sensitive. Some dot extensions are general purpose (also called architecture independent or pin-to-pin) and can used with a variety of device architectures. Other dot extensions are used for specific classes of device architectures and are called architecture-dependent or detailed dot extensions. In general, you can use either dot extensions.

Dot extension Description
Architecture independent or pin-to-pin extensions
.ACLR Asynchronous register reset
.ASET Asynchronous register preset
.CLK Clock input to an edge-triggered flip-flop
.CLR Synchronous register reset
.COM Cominbational feedback from flip-flop data input
.FG Register feedback
.OE Output enable
.PIN Pin feedback
.SET Synchronous register preset
Device Specific extensions (architecture dependent)
.D Data input to a D Flip flop
.J J input to a JK flip-flop
.K K input to a JK flip-flop
.S S input to a SR flip-flop
.R R input to a SR flip-flop
.T T input to a T flip-flop
.Q Register feedback
.PR Register preset
.RE Register reset
.AP Asynchronous register preset
.AR Asynchronous register reset
.SP Synchronous register preset
.SR Synchronous register reset

The figure below illustrates some of the extensions.




Figure 1: Illustration of DOT extensions for: (a) an architecture independent (pin-to-pin) and (b) arhitecture dependent D-type (or T-type) Flip Flop Architecture

Example 1:

which accesses the tri state control signal of the output buffers of the signals S6..S0. When ACTIVE is high, the signals will be enabled, otherwise a high Z output is generated.

Example 2:

which resets to output of the registers (flip flops) to zero when reset is high.

8. Test vectors

Test vectors are optional and provide a means to verify the correct operation of a state machine. The vectors specify the expected logical operation of a logic device by explicitly giving the outputs as a function of the inputs.

Syntax:

Example:

One can also specify the values for the set with numeric constants as shown below.

Don't cares (.X.), clock inputs (.C.) as well as symbolic constants are allowed, as shown in the following example.


9. Property Statements

ABEL allows to give device specific statements using the property statement. This statement will be passed to the "Fitter" program during compilation. For the CPLD devices these properties include

10. Miscellaneous

a. Active-low declarations

Active low signals are defined with a "!" operator, as shown below,

When this signal is used in a subsequent design description, it will be automatically complemented. As an example consider the following description,

In this example, the signal OUT is an XOR of A and B, i.e. OUT will be "1" (High, or ON) when only one of the inputs is "1", otherwise OUT is "0". However, the output pin is defined as !OUT , i.e. as an active-low signal, which means that the pin will go low "0" (Active-low or ON) when only one of the two inputs are "1". One could have obtained the same result by inverting the signal in the equations and declaring the pin to be OUT, as is shown in the following example. This is called explicit pin-to-pin active-low (because one uses active-low signals in the equations).

Active low can be specified for a set as well. As an example lets define the sets A,B and C.

The last equation is equivalent to writing

References

Acknowledgement

The support of Mr. Jason Feinsmith and the Xilinx Corporation is acknowledged for providing the XILINX Foundation M1(TM) Software and FPGA Demoboards for educational purposes.

Back to ABEL Primer Contents | To to Common Mistakes list | Go to the EE Undergraduate Lab Homepage | Go to Xilinx Lab Tutorial Homepage | Go to the Foundation Tutorial page | Go to EE200 or EE200 Lab Homepage | .

Created by J. Van der Spiegel, <jan@ee.upenn.edu> Sept. 26, 1997; Updated Dec. 30, 1997