Unterschiede
Hier werden die Unterschiede zwischen zwei Versionen angezeigt.
Beide Seiten der vorigen Revision Vorhergehende Überarbeitung Nächste Überarbeitung | Vorhergehende Überarbeitung Nächste Überarbeitung Beide Seiten der Revision | ||
introduction_to_digital_systems:sequential_logic [2021/11/22 02:39] tfischer |
introduction_to_digital_systems:sequential_logic [2022/12/09 15:56] mexleadmin |
||
---|---|---|---|
Zeile 3: | Zeile 3: | ||
"I Know What You Did Last Cycle" | "I Know What You Did Last Cycle" | ||
- | ===== 6.1 State Diagram, State Transition Diagram | + | ===== 6.1 First Terminology |
- | ==== 6.1.1 Motivation ==== | + | The most important term for the upcoming topics is the word **state**. But what is a state? \\ It is a unique situation, where the possible next steps (= possible next states), the inner behavior or the outputs are distinguishable from other situations. |
+ | Here some practical examples: | ||
+ | * Being happy or being sad, are two different states, since the inner behavior is different (this least often also to a different output). | ||
+ | * Similarly, an empty memory (or harddrive) is in a different state compared to a filled one. | ||
+ | * A traffic light showing green has a output distinguishable from red, or yellow. | ||
+ | Sequential logic is used to describe logic cicruits which show internal states (" | ||
+ | |||
+ | The following terminology is used in the upcoming explanations: | ||
+ | * The **input vector** $\vec{X}$ represents the $k$ inputs $X_0 ... X_{k-1}$ | ||
+ | * The **output vector** $\vec{Y}$ represents the $l$ outputs $Y_0 ... Y_{l-1}$ | ||
+ | * The **state vector** $\vec{Z}$ represents the $m$ inputs $Z_0 ... Z_{m-1}$ | ||
+ | * The sign $(n)$ or $n$ marking the current point in time and therefore e.g. the current state $Z_0(n)$ | ||
+ | * The sign $(n+1)$ or $n+1$ marking the next upcomming point in time and therefore e.g. the next state $Z_0(n+1)$ | ||
+ | * Sequential logic circuits are also called **Finite State Machines** (FSM) or sometimes also shortened to "state machine" | ||
+ | |||
+ | The <imgref pic04> shows the different terms in an abstract diagram. The " | ||
+ | |||
+ | < | ||
+ | |||
+ | The principle interior of the blackbox in <imgref pic04> was already shown in one practical application in the [[: | ||
+ | |||
+ | < | ||
+ | |||
+ | <panel type=" | ||
+ | |||
+ | <imgref pic06> depicts a state machine. | ||
+ | * What happens, when $X$ is changed? On which edge the change is triggered? | ||
+ | * Write down how many components each vector $\vec{X}$ and $\vec{Y}$ has. | ||
+ | * How many bits (= flip flops) might the state vector $\vec{Z}$ need? | ||
+ | |||
+ | < | ||
+ | {{url> | ||
+ | |||
+ | </ | ||
+ | |||
+ | One simple example of a sequential logic is shown in <imgref pic09>. There, the combnatorial logic is explicitely shown. Depending on the input $X$ the output $\vec{Y}$ shows an up-counting 2-bit value counting $0 \rightarrow 1 \rightarrow 2 \rightarrow 3 \rightarrow 0 \rightarrow ...$. This is a simple state machine which will be used in the net chapters | ||
+ | |||
+ | < | ||
+ | {{url> | ||
+ | |||
+ | ===== 6.2 Classical State Machine Types ===== | ||
+ | |||
+ | The up-counter in the previous sub-chapter was able to count from $0$ to $3$. But what can we do in order to count differently, | ||
+ | |||
+ | ==== 6.2.1 Moore Machine ==== | ||
+ | |||
+ | The first idea might be to use what we already have: an up-counter, which faciliate 2 flip-flops in order to result into 2-bit output. | ||
+ | The wanted new state machine needs 3 bits for the output, since the binary representation of our outputs are $001_2$, | ||
+ | |||
+ | An simple idea is to take the 2-bit up-counter and add an combinatorial logic in behind. This logic shall convert the 2-bit up-counter output $00_2$ into $001_2$, the $01_2$ into $010_2$, $10_2$ into $011_2$ and $11_2$ into $101_2$. This can be logic can be created by: | ||
+ | * writing down the truth table | ||
+ | * putting the values into a Karnaugh map | ||
+ | * extracting the formula with view onto the implicants | ||
+ | * generating the circuit with gates | ||
+ | |||
+ | When this is done, the result looks like <imgref pic21> | ||
+ | |||
+ | < | ||
+ | {{url> | ||
+ | |||
+ | This resembles a so called **Moore Machine** | ||
+ | |||
+ | <WRAP column 100%> | ||
+ | <panel type=" | ||
+ | |||
+ | A state machine is a **Moore Machine**, when the output values $\vec{Y}$ depends only on the state values $\vec{Z}$. For this the moore machine uses two combinatorial circuits: | ||
+ | * The input circuit, which process the input values $\vec{X}(n+1)$ and the state values $\vec{Z}(n)$ (of the previous step) in such a way that the new states $\vec{Z}(n+1)$ are generated. | ||
+ | * The output circuit, which transform the state values $\vec{Z}(n)$ into the output values $\vec{Y}(n)$. | ||
+ | |||
+ | The properties of a moore machine are: | ||
+ | * The number of flip-flops is only given by number of states $m$. | ||
+ | * The output only changes when an edge on the clock input happen. The moore machine is a **synchronous state machine**. | ||
+ | * The moore machine usually need less logic gates. But this comes with the cost of optimizing two combinatorial logic circuits. | ||
+ | |||
+ | </ | ||
+ | </ | ||
+ | |||
+ | ==== 6.2.2 Mealy Machine ==== | ||
+ | |||
+ | When looking onto <imgref pic21> a bit more in detail, one can see, that the outputs $Y_0$ and $Y_1$ just equals output of the first combinatorial logic circuit. This is not surprising: the input logic circuit shows the $\vec{Z}(n+1)$ and this is for the counter always the stored value plus one, except when the maximum is reached. | ||
+ | |||
+ | With this information the state machine in <imgref pic21> can be simplified by using the outputs of the input circuit for $Y_0$ and $Y_1$. This is shown in <imgref pic22> | ||
+ | |||
+ | < | ||
+ | {{url> | ||
+ | |||
+ | <WRAP column 100%> | ||
+ | <panel type=" | ||
+ | |||
+ | A state machine is a **Mealy Machine**, when the output values $\vec{Y}$ depends not only on the state values $\vec{Z}$. For this the mealy machine uses two combinatorial circuits: | ||
+ | * The input circuit, which process the input values $\vec{X}(n+1)$ and the state values $\vec{Z}(n)$ (of the previous step) in such a way that the new states $\vec{Z}(n+1)$ are generated. | ||
+ | * The output circuit, which transform the state values $\vec{Z}(n)$ __and__ some inputs from ahead of the flip-flops into the output values $\vec{Y}(n)$. | ||
+ | |||
+ | The properties of a mealy machine are: | ||
+ | * The number of flip-flops is only given by number of states $m$. | ||
+ | * The output **not** only changes when an edge on the clock input happen: It is also dependent on the input $\vec{X}(n)$. The mealy machine is an **__asynchronous__ state machine**. | ||
+ | * The mealy machine usually need less logic gates. | ||
+ | * The mealy machine has to be designed properly in order not to get invalid outputs. | ||
+ | |||
+ | </ | ||
+ | </ | ||
+ | |||
+ | <panel type=" | ||
+ | |||
+ | The mealy machine in <imgref pic22> can show invalid outputs. Try to find these by the correct timing of the input $X=1$ or $X=0$. | ||
+ | * Which outputs can be created? | ||
+ | </ | ||
+ | |||
+ | ==== 6.2.3 Medvedev Machine ==== | ||
+ | |||
+ | < | ||
+ | {{url> | ||
+ | |||
+ | <WRAP column 100%> | ||
+ | <panel type=" | ||
+ | |||
+ | A state machine is a **Medvedev Machine**, when the output values $\vec{Y}$ is directly given by the state values $\vec{Z}$. For this the medvedev machine uses only one combinatorial circuit. This circuitprocess the input values $\vec{X}(n+1)$ and the state values $\vec{Z}(n)$ (of the previous step) in such a way that the new states $\vec{Z}(n+1)$ and the output values $\vec{Y}(n+1)= \vec{Z}(n+1)$ are generated. | ||
+ | |||
+ | The properties of a medvedev machine are: | ||
+ | * The number of flip-flops is given by number of outputs $l$. | ||
+ | * The output only changes when an edge on the clock input happens: The mealy machine is an **__synchronous__ state machine**. | ||
+ | * The medvedev machine usually need more logic gates. | ||
+ | |||
+ | </ | ||
+ | </ | ||
+ | |||
+ | |||
+ | The <imgref pic101> shows the principle differences in the architecture of the state machines. | ||
+ | |||
+ | < | ||
+ | |||
+ | \\ \\ | ||
+ | <WRAP column 100%> | ||
+ | <panel type=" | ||
+ | This chapter is only focussing on Moore machines. | ||
+ | </ | ||
+ | </ | ||
+ | |||
+ | |||
+ | ===== 6.3 State Diagram, State Transition Diagram ===== | ||
+ | |||
+ | ==== 6.3.1 Motivation ==== | ||
The diagrams of different states are well known from physics for example the state diagram (or better: phase diagram) of water, where its three states are: solid ice, liquid water and gaseous steam. The possible state transitions are due to temperature increase or decrease. | The diagrams of different states are well known from physics for example the state diagram (or better: phase diagram) of water, where its three states are: solid ice, liquid water and gaseous steam. The possible state transitions are due to temperature increase or decrease. | ||
In <imgref pic01> image (1) the states of water are shown on the temperature axis. When only the state transistions are relevant, the states are simplified to a circle, showing the state name and behaviour. The transitions are depict as arrows, where the needed condititon is written onto (See <imgref pic01> image (2) ). This diagram is called **state transition diagram**. | In <imgref pic01> image (1) the states of water are shown on the temperature axis. When only the state transistions are relevant, the states are simplified to a circle, showing the state name and behaviour. The transitions are depict as arrows, where the needed condititon is written onto (See <imgref pic01> image (2) ). This diagram is called **state transition diagram**. | ||
- | < | + | < |
For matter not only the dimension " | For matter not only the dimension " | ||
By this, another variable is available and more transistions. These can be drawn into the state transition diagram (<imgref pic02> image (2)). | By this, another variable is available and more transistions. These can be drawn into the state transition diagram (<imgref pic02> image (2)). | ||
- | < | + | < |
- | ==== 6.1.2 Simple logic Example ==== | + | ==== 6.3.2 Simple logic Example ==== |
In German, often one has to pay for entering the toilet. An example of such a entrance control system is shown in <imgref pic03>. At this (artificial) example, one can pay either 50ct or 1€. \\ | In German, often one has to pay for entering the toilet. An example of such a entrance control system is shown in <imgref pic03>. At this (artificial) example, one can pay either 50ct or 1€. \\ | ||
Once paid, the turnstile will release and one can enter. Once the turnstile was pushed the entrance is closed again. | Once paid, the turnstile will release and one can enter. Once the turnstile was pushed the entrance is closed again. | ||
- | < | + | < |
The <imgref pic04> the state transition diagram is drawn. | The <imgref pic04> the state transition diagram is drawn. | ||
Zeile 30: | Zeile 171: | ||
Important here are some additional considerations: | Important here are some additional considerations: | ||
* For the state transition diagram one has to **look for all possible transitions**. So, also pushing a closed turnstile or inserting more coins have to taken into account. | * For the state transition diagram one has to **look for all possible transitions**. So, also pushing a closed turnstile or inserting more coins have to taken into account. | ||
- | * A state transition diagram is not complete without a legend and without an beginning/ | + | * A state transition diagram is not complete without a **legend** and without an **beginning/ |
- | < | + | < |
Out of this state transition diagram one can create a table-like representation, | Out of this state transition diagram one can create a table-like representation, | ||
- | < | + | < |
the inputs, outputs and states have to be encoded into binary, in order to investigate this table a bit more. How the binary value is connected to the outputs does not matter. We will choose the following coding: | the inputs, outputs and states have to be encoded into binary, in order to investigate this table a bit more. How the binary value is connected to the outputs does not matter. We will choose the following coding: | ||
- | * Encoding of the states: turnstile closed ≙ $Q=0$, turnstile opened ≙ $Q=1$, | + | * Encoding of the states: turnstile closed ≙ $Z=0$, turnstile opened ≙ $Z=1$, |
* Encoding of the inputs: no coin inserted ≙ $Xc=0$, coin inserted ≙ $Xc=1$, turnstile not pushed ≙ $Xp=0$, turnstile pushed ≙ $Xp=1$, | * Encoding of the inputs: no coin inserted ≙ $Xc=0$, coin inserted ≙ $Xc=1$, turnstile not pushed ≙ $Xp=0$, turnstile pushed ≙ $Xp=1$, | ||
* Encoding of the outputs: disallow entrance ≙ $Y=0$, allow entrance ≙ $Y=1$, | * Encoding of the outputs: disallow entrance ≙ $Y=0$, allow entrance ≙ $Y=1$, | ||
- | This table is shown in < | + | This table is shown in < |
- | < | + | < |
+ | {{drawio> | ||
Interestingly, | Interestingly, | ||
- | When looking deeper onto the table in < | + | When looking deeper onto the table in < |
+ | |||
+ | ==== 6.3.3 First Adaption for Up-Counter==== | ||
+ | |||
+ | In the previous sub-chapter (6.2) we had a look onto different implementations of an up counter and on this chapter the way to represent a state machine via a state transition diagram. So one question is: how does these up counter look like in the state transition diagram? | ||
+ | |||
+ | The Answer ist quite simple: there are 4 states, so 4 " | ||
+ | With a deeper look onto is: This is therefore in particular a Medvedev machine! | ||
+ | |||
+ | In <imgref pic15> both types of state machines are shown. | ||
+ | * In the Moore Machine the already seen " | ||
+ | * In the Medvedev Machine only one value is written in the circle, since here the state vector is also the output vector. | ||
+ | |||
+ | < | ||
+ | {{drawio> | ||
+ | |||
+ | <panel type=" | ||
+ | The shown state transistion diagram can easily be created in the tool Digital: | ||
+ | - Click on the menu '' | ||
+ | - Here one either can add new states with a right click, or - more easier - go to the menu '' | ||
+ | - A (up) counter will get created | ||
+ | |||
+ | Tasks: | ||
+ | - Make the state machine get alive: | ||
+ | - Click to the menu '' | ||
+ | - Here, click on the menu '' | ||
+ | - Now, the circuit can be started via the start button. The present state in the state transition diagram will also get highlighted. | ||
+ | |||
+ | </ | ||
+ | |||
+ | The next step is to transform this into a up counter from $1...4$. For this simply the output has to be changed. This means the numbers in the "lower half of the circles" | ||
+ | |||
+ | < | ||
+ | {{drawio> | ||
+ | |||
+ | <wrap # | ||
+ | <panel type=" | ||
+ | This exercise directly attached onto Exercise 6.3.3.2 | ||
+ | |||
+ | The counter shall now be changed in such a way, that it counts $1 - 2 - 3 - 4$. For this: right click on each present states beginning with state 0 and add for '' | ||
+ | |||
+ | {{drawio> | ||
+ | |||
+ | Tasks: | ||
+ | - Make this state machine again get alive | ||
+ | - Look at the circuit: Is it a Moore, Mealy or a Medvedev machine? | ||
+ | |||
+ | </ | ||
+ | |||
+ | Looking on what we got, there is one part missing: the state machine does not have any input.. So the input vectors have to be added onto the transition (arrows). For an activateable counter, there only have to be one input $X$, which acts as an enable input. | ||
+ | |||
+ | < | ||
+ | {{drawio> | ||
+ | |||
+ | <panel type=" | ||
+ | This exercise directly attached onto Exercise 6.3.3.2 | ||
+ | |||
+ | The last step is to add the transitions: | ||
+ | |||
+ | {{drawio> | ||
+ | |||
+ | Tasks: | ||
+ | - Make this state machine again get alive | ||
+ | </ | ||
+ | |||
+ | <panel type=" | ||
+ | This exercise directly attached onto Exercise 6.3.3.3 | ||
+ | |||
+ | With the state machine in 6.3.3.3 it is also possible to create a state machine which can resemble a traffic light state machine. This shall transit from red, to red-yellow, to green, to yellow and then back to red. | ||
+ | |||
+ | Use the following addition to the resulting citcuit in order to get the light running: | ||
+ | {{drawio> | ||
+ | |||
+ | |||
+ | Tasks: | ||
+ | - Think about the right way to change the outputs in the state machine. | ||
+ | - Implement the needed correction and test the state machine. | ||
+ | </ | ||
+ | |||
+ | ====== Exercises ====== | ||
+ | |||
+ | <panel type=" | ||
+ | |||
+ | The following state transistion diagram shall be given: | ||
+ | {{drawio> | ||
+ | |||
+ | |||
+ | * There are two transitions marked with $A$ and $B$. What values does the inputs need to have in order show all transistions explicitely? | ||
+ | * How many flipflops are necessary for such a Moore Machine? | ||
+ | * Fill in the missing cells in the following state transition table: | ||
+ | |||
+ | {{drawio> | ||
+ | |||
+ | </ | ||
+ | |||
+ | <panel type=" | ||
+ | |||
+ | Design a state machine, which result in an output of the following (decimal) numbers: $4 - 5 - 2 - 1 - 6 - 7 - 0 - 3 - 4 - ...$ | ||
+ | * The numbers shall repeat cyclic and without any other input than the clock $CLK$ | ||
+ | * Draw the state transition diagram and the state transition table | ||
+ | * Use alternatively the structure of a 3bit up-counter | ||
+ | * Draw the structure of this moore machine with the up-counter as a blackbox, the input and output values and - if necessary - further blackboxes | ||
+ | * draw and fill in the additional table. | ||
+ | |||
+ | </ | ||
+ | |||
+ | <panel type=" | ||
+ | |||
+ | Design a state transistion diagram for a automatic milling machine as Moore machine, under the following conditions: | ||
+ | - The milling machine has 4 states: Initialisation $I$, Component Change $C$, Running $R$, Error $E$ | ||
+ | - The milling machine starts at the state $I$. | ||
+ | - Once a limit switch $L$ is read as activated, the state $E$ shall be entered, independent which state the machine was in. | ||
+ | - Only in the state $E$ an alarm shall ring $A=1$. | ||
+ | - $E$ can only be exited with a reset. | ||
+ | - In order to change from Component Change $C$ to Running $R$, the user has to activate a Key $K=1$. | ||
+ | - Only when the machine is running $R$ and the Fixed Condition is reached $F=1$, the state Component Change $C$ shall be entered. | ||
+ | - Initialisation $I$ only changes into Component Change $C$, when no limit switch $L$ is activated and the Key is deactivated (input $K=0$) and the Fixed Condition is reached $F=1$. | ||
+ | |||
+ | Explicitely draw all possible transitions. | ||
+ | |||
+ | {{drawio> | ||
+ | |||
+ | </ | ||
+ | |||
+ | |||
+ | <panel type=" | ||
+ | |||
+ | Develop a sequential circuit, which creates the following output | ||
+ | |||
+ | {{drawio> | ||
+ | |||
+ | |||
+ | * Draw the state transition diagram of the moore machine of the synchronous sequential circuit. | ||
+ | * Create the digital ciruit. | ||
+ | |||
+ | </ | ||
+ | |||
+ | |||
+ | <panel type=" | ||
+ | |||
+ | Develop a sequential circuit, which allows driving the following LED sequence | ||
+ | |||
+ | {{drawio> | ||
+ | |||
+ | * Draw the state transition diagram of the moore machine of the synchronous sequential circuit. | ||
+ | * Create the digital ciruit. | ||
+ | |||
+ | </ | ||
+ | |||
+ | |||
+ | <panel type=" | ||
+ | |||
+ | Develop a sequential circuit, which generate a clock driven up-counter from $1..6$. The not reqired states shall lead after one clock cycle to the state of number $1$. | ||
+ | |||
+ | * Draw the state transition diagram of the moore machine of the synchronous sequential circuit. | ||
+ | * Create the digital ciruit. | ||
+ | |||
+ | </ |