MXLED.exe

MXLED is a simulation of the LED matrix that can be used with a microcontroller simulator on Micro .exe. With MXLED, the system design of LED matrix, commonly referred to as Running Text, or many also call it Moving Sign, be more easily implemented.

MXLED on this version provides the size up to 200 columns x 80 lines. With this size, we can perform simulations for the LED matrix which suffice to needs that more real.

Enlarge the size of the matrix on this version is done by reducing the size of the LED to be only of 10 x 10 pixels. However, if we use monitors with size 1360 x 768, then size that can be displayed only about 135 columns x 64 lines.

MXLED

Arrange LEDs as matrix

MXLED is a simulator of the circuit matrix of LEDs. With this simulator, we can try to program the LED matrix controllers even without the hardware. MXLED made LED matrix by arranging the LEDs on the vertical lines and horizontal lines. We must provide the number of rows of horizontal lines (we make eight lines). Then, we also make as many vertical lines of columns. Arrangement of vertical lines and horizontal images are as follows:

Vertical and Horizontal Line

Lines of vertical and horizontal are not connected. Then, at each meeting point between the vertical and horizontal lines, attach a LED by connecting its anode to the horizontal lines and its cathode to the vertical lines. Installation of these LEDs are as shown below:

Installation of LEDs

By installing LEDs as above, the light up LED is the LED where the anode is connected to the horizontal lines which is high (1) and the cathode is connected to the vertical lines which is low (0). There is only one low vertical line at a time, while other line must be kept high. This low vertical line we refer to as the active column. Unlike the vertical lines, horizontal lines consisting of eight lines may has high or low value without having to pay attention to the other horizontal lines.

Horizontal Control

To give the voltage on this horizontal line, we can not directly connect it to port. This is caused by the need for large electrical currents. Therefore, horizontal lines are the powered using a PNP transistor as shown below:

The power supply for the horizontal lines

Each horizontal line given a transistor like the picture above. In this way, to make the horizontal lines can give the current to the LED, then the base must be low state (0). Meanwhile, if the base was given a high state, then the line could not provide current to the LED.

Vertikal Control

As horisontol lines, vertical lines can not be controlled directly using the port. This is caused by the large amount of current that must be sunk into the ground. Therefore, we can use the NPN transistor to sink current from these vertical lines. Way of installation is as shown below:

Vertical line current sink

Each vertical line is controlled using a transistor as shown above. If the base was given a high condition (1), then the line (column) will be an active line. Conversely, if the base were given the condition of low (0), then the line becomes inactive column.

As already mentioned above, there is only one column that should be active, while the other columns must be inactive. There are many ICs that output works like that. One of the cheapest is the 4017.

Columns control using IC 4017

Too bad that this IC only has 10 outputs. Thus, we can only make the controls to 10 columns

Relax, we can use other tricks to extend the control capabilities of this 4017. The trick is to make the block columns. The columns are grouped into block of columns. Each block columns consists of 10 columns. The columns in each block columns are numbered from 0 to 9. So, column 0 is the column 0 of block 0, column 10 is the column 0 of block 1, column 21 is column 1 of block 2, and so on.

Each base of the column with the same number are combined into one and is controlled by an output of 4017. For example, the base of the column 0 is connected to the base of the column 10, column 20, column 30, and so on. Then the bases which have become one, controlled by the Q0 of 4017.

Block Columns

From the picture above, we can see that although the base of column 0 and column 10 are controlled concurrently, but the column that can sink current is only the column with the active block. Seen that way of controlling the “column” or “block columns” is the same. If the number of block columns is only a few, then the controlling block of columns can be done directly using the port of the microcontroller. However, if the number of block columns is quite a lot, then the block columns can also be controlled using the other 4017. So, we develop a multilevel 4017, ie 4017 units (control column) and 4017 tens (the controlling block of columns). And if the number of block columns more than 10, then we can make the next level, ie 4017 level in the hundreds.

To cascade 4017 as above, is by connecting the carry output of 4017 units level as the input clock for the 4017 of tens level. And if there is 4017 of hundreds level, then the carry output of 4017 of tens level is used as the input clock for the 4017 of hundreds level. And so on. While the reset input of all of 4017 are combined into one. Thus, the overall control system only requires two lines for columns control, ie the input clock (clock input for 4017 of unit level) and the reset input (composite of all reset of 4017). Thus, the preparation of the LED matrix is exactly the same as the MXLED simulation.

Power Current Calculation

From the explanations above, we can see that there are three kinds of transistors viewed from the position. First is the row controller transistor, second is the column (unit) controller transistor, third is the column block (tens) controller transistor. Of course this calculation assuming that the number of columns are not more than 100 columns.

Calculation of the row controller transistor current

If each LED using a current of I, then the total maximum amount of current through the line control transistor is I x number of columns. For example, if each of our LEDs designed to use 5mA current and the number of columns are 30 columns, then the row controller transistor should be able to drain 5 x 30 = 150mA current.

The next question is, what is the value of collector resistor of the transistor?

To answer these questions, we should see the path of current from the power supply to get to the ground. First, the power supply current enter the PNP transistor through the emitter to the collector. Then passes through the collector resistor, then continues to enter into LED, enter into the collector of column controller, then enter the collector of the block columns controller.

If the power supply used is 5V, how much is the clamp voltage across the resistor? The voltage across the resistor is 5V – V on LED – VCE line controller tansistor – VCE of column controller transistor – VCE of column block controllers transistor. LED clamp voltage is typically about 1.7 V, but there are some types of LEDs which have a clamp voltage of up to 3V. While VCE transistors in saturation state is typically around 0.3 V. We assume that we use a regular LED with a clamp voltage of 1.7 V. Thus, the voltage across the resistor pin is 5 – 1.7 – (3 x 0.3) = 2.4 V.

Once we know the clamp voltage at the collector resistor of the line controller transistor, to calculate the magnitude of the resistor is R = V / I = 2.4 V / 150mA = 16 ohms. Too bad that we might be difficult to obtain this value of 16 ohms. So we can use a slightly lower value, eg 15 ohms, so that the LED current will be slightly higher, or use a little higher, for example 18 ohms, so that the LED current will be slightly lower.

Things should be kept on the LED matrix design like this is, that these calculations is the calculation of the current if the scanning is running. If scanning is not running, then a current of 150mA will be entered on one LED alone And currents of this magnitude would likely destroy the LED. So, we must take precautions so that the current should only flow if scanning process has been running. A little good news is, if we use the MCS-51 as a controller, so when the reset (the scanning process is not running) port is always in a state of high. So, if the base of line controller transistor is connected to the MCS-51 port, so it can be sure that the current will not flow to the LEDs on the reset state. However, you must ensure that the program will not hang. Because if the program hangs and stops the scanning process, then it means disaster for your LED matrix. And if you are not sure that your program can run smoothly, then you can reduce the LED current so that even if the scanning process is not running, current flows are still small enough to be able to be borne by one LED.

To calculate the base resistor value of the line controller transistor, we only consider the transistor to work as a transistor switch, which is so current that flows is the saturation current. With the calculation of the collector current of 150mA, then we can use the C9012 transistor that has a current gain of about 150. We can take a value of 100 to make it more secure. With the gain of 100 and the collector current of 150mA, then the base current should be about 1.5 mA. Clamp voltage at the base resistor is approximately 5V – VBE, where VBE is about 0.7 V. Thus, the clamp voltage at the base resistor is about 4.3 V. Thus, the base resistor value is 4.3 V / 1.5 mA = 2K8. And again that this value is not available values. So we could slightly reduce the value to the nearest number of available, namely 2k7.

Calculation of the column controller transistor current

Column controller will sink as much current number of rows in that column. And we have determined that the number is eight lines. If we have determined that the LED current is 5mA, then the maximum current to be sunk by the column controller transistor is 8 x 5mA, which is 40mA. With the value of this, we simply use the C9013 transistor to control the column. The Gain of C9013 transistor is also about 150. And as before, we take saver value to 100 only. With the gain of 100 and the collector current of 40mA, then the base current in the column controller transistor is 0.4 mA.

Clamp voltage at the base resistor of the column controller transistor is the output voltage of 4017 minus 2x the VBE voltage. If we use a 5V power supply, then the output voltage of 4017 is approximately 4.8 V. Thus, the clamp voltage at the base resitor is about 4.8 – 2 x 0.7 = 3.4 V. Thus, the base resistor value of the column controller transistor is 3.4 V / 0.4 mA = 8K5. And again, this value is not available values. So, just use the 8K2.

Calculation of the block columns controller transistor current

The block columns controller transistor will sink as much current in the column controller transistor x 10, which is 10 x 40mA = 400mA. From the datasheet, C9013 has a maximum collector current of 500mA. So, we can still use the C9013 to controll block columns.

Clamp voltage at the base resistor of the block columns controller transistor is the voltage used to control this transistor reduced by 1x VBE. If we also use 4017 to control the block column, then the clamp voltage across the base resistor of the block columns controller transistor is approximately 4.8 – 0.7 = 4.1 V. Thus, the value of the base resistor of the block columns controller transistor is 4.1 V / 4mA = 1K.

That’s it. Happy trying

SSMaker.exe

SSMaker.exe is a program to help make a lighting table of seven segment.

Click on the segment to be switched on or off. After that, the table to form the light will be instantly displayed. Lights are represented by the value bit 0. To make the light represented by the value bit 1, check the Active High.

Constants for the number 5 is 92H (light = 0) or 6DH (light = 1)

By default, bit 0 is for segment A, bit 1 is for the segment b, and so on. But you also can change the order. Of course this arrangement tailored to the hardware. To determine the bits to its segments, right click on the segment to be determined its bit, then from pupop menu that appears, choose the number of bits.

Right-click on a segment for selecting bits for the segment.

If we use the bits that are being used by another segment, the segment which bits are used will be made to not connected to any bit. Signs that the segment is not connected to the bit is the segment which is marked with the letter X.

This program is a package of Microcontroller Project

In addition to using assembly language, Microcontroller Project can also use c language for writing programs, using SDCC.

Microcontroller Project integrate editor that comes with code explorer, code hint, and auto-completion so that writing programs is easier. And with the simulator, you will be facilitated in understanding the course of the program and while tracking program logic error.

Microcontroller Project provides additional features from previous versions that will add convenience you make microcontroller projects. New features include:

• Code Explorer

  If we create a variable declaration, then automatically the variables will be listed and will appear in the left panel of each editor. Likewise, the label that we have created, a macro that we created and others.

  The list was made in the left pane are sorted alphabetically, not position. By double-clicking on an object, the cursor will be placed in the position where the objects is written.

• Auto Completion

  With the auto completion, when we write programs, we will be treated to a choice with the objects we have created. Thus, writing the program can be more quickly and does not easily occur wrongly written.

  

  Code completion appear automatically or can be invoked with the Ctrl+Spasi

  Code completion will automatically appear if we create a space, and is the right part to add the code. For example, when we make a space after writing acall, then that will appear are all labels that have been made. Likewise, if we write mov, then after adding a space, will soon be raised all the variables we have made.

  We can also invoke code completion by pressing the Ctrl+Spasi. If we write programs in C language, then the code completion will not appear automatically, we must call it to appear by pressing the Ctrl+Space.

• Code Hint

  If we create a macro or function on writing C programs, then sometimes we forget what the parameters are owned by the macro or function. When we write a macro or function call, it will automatically appear hint that indicates the parameters that are owned by the macro or function.

  

  Code Hint helps show the parameters in the macro or function. Appears automatically or can be invoked by pressing Ctrl+Shift+Space

  And just like code completion, if code hint does not appear automatically, then we can call it to appear by pressing the Ctrl+Shift+Spasi.

The most interesting of the Microcontroller Project is the integration with simulators that can be linked to various programs / other simulators. Simulator capability that can be linked to other simulators makes Microcontroller Project very easy for us in designing microcontroller-based systems.

Micro V7.0

In the example in Showing images on LED matrix using simulator, the note is displayed as an image. First we create a text image, then we create a constant from the text image. After that it displayed as an image.

Displaying text in this way is only appropriate if the text to be displayed is short. In addition, the displayed text will not be changed anymore. If the note to be displayed is a long text, or text that is displayed will be changed, then there is a better way than that way.

A better way is as follows: First, we create a character constant. In the example here, we make a character constant of 8 rows x 5 columns character size. The character constant was created using Karakter.bmp that you can edit to fit your desires.

Once we get a constant character, by compile the Karakter.bmp using ImgTable.exe, then every time we will display the text, we do it by reading the text character-by-character. Each character is read, then displayed by looking at the character table. In this way, we can display any article and can be changed while the program is running. Of course you have to create a procedure to change the text to be displayed if you want to replace it.

Unfortunately the program listing only writen in assembly language. So, if you prefer to use the C language, please do the conversion yourself

If you want to try the example program using MXLED, do the simulator settings as in Showing images on LED matrix using simulator

Good luck

What trick to use?

It must be admitted that designing microcontroller based system does require a fairly complex science base. So many people give up before trying. They were puzzled by a vicious cycle, where to start?

Importantly, you need lots of practice, both theory and practice. Therefore, you should first prepare the tools and materials, ie minimal microcontroller system (consisting of a microcontroller chip, resistors, capacitors, and crystals), downloader, and Integrated System Environment (IDE).

The components you can buy in electronics stores. Downloader you can make yourself or buy. IDE you can get here. Do not forget, have the Easy and Fun Learning Microcontroller book beside you

Indeed, to be an expert, it took quite a long time. But, to only to be able to use a microcontroller, one or two days is enough! Start with the simplest, gradually continue to rise to complex. Later, if you feel stuck facing a complicated problem, make the problem becomes simple. If you are facing a big problem, make the problem becomes small. If you encounter a difficult chapter, do not force yourself. Believe me, another time you can certainly understand it.

What microcontroller to use?

Once you master one type of microcontroller, it will be easy for you to move to other types of microcontrollers. So, no matter what microcontroller to use, but its programming techniques.

AT89S51 or AT89S2051 from MCS-51 family can be a good choice. This easy-cost microcontroller-rousing. You will not be too sad if you make it destroyed. Instead, you’ll be very happy if you can master the science of microcontroller with low-cost.

What language to use?

There are many choices of programming languages. There are suggestion to use high-level programming language first, such as Basic, Pascal, C, and others. However, the best programming language that will really make you understand the microcontroller is the assembly that this language requires you to make the instructions carefully.

Happy learning!

A book reader Mudah dan Menyenangkan Belajar Mikrokontroler asked as follows:

1. In the program listing of digital clock with button (p. 91), please explain more about the use and calculation of Timer 0 Mode 2 so as to produce an accurate timer of 1 second.
2. Please explain more about the workings of the listing of button data retrieval (p. 119).
3. How do I use a variable declaration in listing with *.M51 extension such: ~byte, ~bit, ~word, and ~array?
4. How to use the tool: Virtual LCD, Tombol, MXLED, in the simulator?
5. Well, the next question may not relate to the book, but still about a microcontroller. How to make one button so it has two functions? For example we put the button at P3.0. If pressed once it will execute subprogram A and if pressed twice it will execute subprogram B.
6. Mas, I tried to make the clock circuit on PCB and I filled it with the “debouncing 2.HEX” program. The result was running perfectly. But when the minutes or hours modifier button was pressed, the value on 7 segment did not changed directly but wait until the program completing one minutes. How to make changes instantly?

And here’s the answer:

1. Timer 1 sec accurate.

   Actually, the information in the book is clear enough. First, the crystal used was 11.0592 MHz. Thus the speed of the machine cycle is 921.6 KHz (11.0592 MHz/12). Second, TH0 filled #0, which means timer0 interrupt will be called every 256 cycles of the machine. Remember that TH=256-interval or interval=256-TH. Thus the calling interval is 256 machine cycles.

   From both the data above we can calculate the speed of the calling of timer0 interrupt is 3600Hz (921.6 KHz/256). Now, to obtain a second, we must divide this interval again by 3600. So we do this in two steps, which is divided by 225 (3600/225 = 16), then we can simply divide the result by 16 to obtain a frequency of 1Hz. However, because we need to set the frequency to 2Hz to make the blink of the point, then we do not divide by 16 but by 8.

   1Hz frequency is obtained with a flip-flop which is a 2 divider. Flip-flop is obtained by using the cpl command on a flag (Bendera1Detik). Hopefully quite clear.

   Back to top

2. Techniques to retrieve button data by debouncing.

   Pieces of the program will retrieve multiple keys concurrently and has been carried out debouncing. So we can use any key and do not have to worry about bounce. How to use it is just call the procedure, ie “AmbilTombol”.

   After a call to this procedure, Acc will contain the button data. For example, if the button mounted on the bit0 and bit1, then after calling this procedure, A will contain 0 if no keys are pressed, 1 if the button on bit0 is pressed, 2 if the button on bit1 is pressed, and 3 if both buttons are pressed.

   Of course, the resulting value depends on where the buttons are mounted. Just look at the value of bits. Bit0 is 1 and bit1 is 2. About how it happens, try to look at the included flowchart. Learn the flowchart carefully. Hopefully quite clear.

   Back to top

3. Memory allocation for variables with a technique that easy to make any modification.

   ~byte, ~bit, ~word, and ~array variable are the relatively easy way to set the location of memory compared with the set of memory locations in a way such as:

     Buffer  equ 8  ;4 byte
     Angka   equ 12
     Puluhan equ 13
     Satuan  equ 14

   It would be easier if we use:

     ~Array 4 Buffer
     ~byte Angka Puluhan Satuan

   With the second way, we will put the Buffer at location 8, the Angka at 12 locations, Puluhan on the location 13, and Satuan on the location 14. And automatically there will be a constant that holds the highest position of the memory location that is not used. In this position should be stored for stack initiation. The constant is named SaveStack. Thus, we can initiate the stack with:

     mov SP,#SaveStack

   and it would have the same meaning with:

     mov SP,#14

   because the highest location used is 14.

   If we use the first way and we want to change the memory requirement for the buffer to be only 2 bytes, we will set it to:

     Buffer  equ 8
     Angka   equ 10
     Puluhan equ 11
     Satuan  equ 12

   and the stack initiation becomes:

     mov SP,#12

   Well, we have to change all the numbers in its memory location setting. But if we use the second way, then we simply replace the number 4 to 2 in the ~array declaration such as:

     ~Array 2 Buffer
     ~byte Angka Puluhan Satuan

   or because the Buffer only requires 2 bytes, then we can replace it with the ~word type such as:

     ~word Buffer
     ~byte Angka Puluhan Satuan

   Note that we only change the Buffer declaration and no other parts that we change.

   While the variable ~bit is a 1-bit variable that is placed on location 20H.0 to 3FH.7.

   Placement of all these variables will always find a lowest location that is still unused starting from address 8 (the default) or can be moved to a specific position using ~basequ.

   And the explanation about this is already in the appendix of the book. Hopefully quite clear.

   Back to top

4. “Virtual LCD” (LCD simulation), “MXLED” (LED Matrix simulation), and “Tombol” (Button simulation using keyboard).

   • “Virtual LCD” can be used for experiments that require a lot of character display. This simulation can be linked with the real circuit using a serial connection. See Chapter 19 of “Easy and Fun Learning Microcontroller”. However, these simulations can also be used from the simulator with “Link Message”. To do this, select used Com to “Message”. If data about the keyboard keystrokes are sent to the simulator via the serial communication (received in SBUF), the “Message to send” must be filled by 1035. Whereas if the data desired to be received by the port, then the “Message to send” is 1034 and “lParam to send” filled with the port number, ie 0, 1, 2, or 3. Besides, “Handle” should point to the active simulator. We can use the “Capture Handle” then click on the window of the target simulator. In addition to settings on the “Virtual LCD”, simulator port must also be adjusted. If the simulator wants to send data to the “Virtual LCD” through a serial communication simulations, the Com should use the “Message” and set “Handle” to point to the desired “Virtual LCD” with “Message to send” in the form WM_USER (1024). Meanwhile, if the simulator was about to send its data through a port, then we have to set the port to use “Link Message” directed to the “Virtual LCD”.
   • “MXLED” is a LED matrix simulation. This simulation also receive data through the “Link Message” from the port settings on the simulator. Because the explanation rather long, so I put in a separate post that is in the Showing Images on LED Matrix using Simulator.
   • For “Tombol”, if the port settings on the simulator using the “Link Message”, then the keyboard keystrokes on the “Tombol” will be sent to the simulator. “Tombol” itself does not require setting, because automatically it will find the current simulator.

   Back to top

5. One button for many functions.

   Such buttons programming requires a fairly precission timing. The first is to check whether the button is pressed or not. If yes, then the process of button detection began.

   The detection of this button is basically the same as other debouncing techniques, ie wait until the button is released again, but if the button has been released does not directly assume that the button has been released. There is a minimum time so that the button has been released is really considered to has been released. If time is not reached, then the next keystrokes will still be considered as part of previous keystrokes.

   Now if the button is eligible to be considered to has been released, meaning that the minimum waiting time (timeout debouncing) exceeded. Program does not directly perform a certain action but simply increase the variable that stating the number of keystrokes, then wait the next keystroke again. This is done by creating a flag stating that the program is waiting for more key presses. This waiting period we now call the “next press timeout”. If the button is pressed again before the waiting time is exceeded, then the keystroke will only increase the calculation of the amount of keystrokes. The program will only process the button press data only when “next press timeout” is exceeded.

   You can download the example program which is a two-digit counter 0-99. Port0 used to transmit data segment, port1 to control digits, while Port3.0 used as a button. To increase the count, we must press the button once. As for decrease the count, we must press the button twice quickly. We can also directly reach a value of 99 by pressing the button four times quickly. As for the immediate return to 0, then we must press the button three times quickly.

   From this example, we can use one button for a lot of actions. But remember, if you create a program and there is a function executed by pressing the button quickly to ten times, can certainly be uncomfortable in their use

   Back to top

6. Revised example program of digital clock with buttons with debouncing.

   After I recheck, there are something missing and some were wrong. Should the MainLoop structure is like:

                 mov   R4,#NoMode     ;R4 =status mode
      MainLoop0: acall IsiBuffer
      MainLoop:  acall AmbilTombol

   And each finished making changes to the data either minutes or hours, then the jump is toward MainLoop0 and not to the MainLoop. In addition, the constants for key presses should be:

      S1ditekan   equ   00010000b ;original =00001000b
      S2ditekan   equ   00100000b ;original =00010000b
      S1S2ditekan equ   00110000b ;original =00011000b

   And the button filtering on “AmbilTombol” should:

      AmbilTombol_:
                  mov   A,P3
                  cpl   a            ;data inversion
                  anl   A,#00110000b ;original =00011000b
                  ret

   These errors resulted in a mistaken reading of the buttons because the buttons position are at P3.4 and P3.5 instead of P3.3 and P3.4. But I have made a revised version with a little addition so that when the status mode is on ModeJam, then the hours digit will be blinking. Similarly, when the status mode is on ModeMenit, then the minutes digit will also be blinking. Thus the user knows what will happen if the button is pressed. Revised program listing can be downloaded here.

   Back to top

Which can be downloaded from this post:

1. Examples of one button many function programs
2. “Debouncing 2″ program revision

There are six sizes provided by MXLED, ie 8 × 16, 8 × 32, 8 × 48, 16 × 16, 16 × 32, and 16 × 48. For each size there are two choices of orientation, ie landscape and portrait.

Control signals only use bit.0 and bit.1. Bit.0 used to reset the counter, while bit.1 used to increase the counter.

Data signal used to determine which LED where lit and LED where off. For each time, there were only eight LED that are controlled, which LED on the current column. To determine which columns is active, we use a counter. At the time counter is reset, then the column 0 is active. then if we provide the clock signal, the signal on bit.1, then the active column will move to column 1. Then if given the clock signal again, the active column will be column 2. and so on.

There are two choices of the clock signal, ie the L to H transition or the H to L transition.

The arrangement of the columns of LEDs depends on the size and orientation. For landscape orientation, then the left column is a lower column and the right column is a higher column. This applies to the size of 8x. As for the size of 16x, then the LED array is divided into two blocks of rows. The first column number of the second row block is the column number of the last column of the first row block plus 1.

More details are as follows:

16 x 16

Row 0..7 : 0  1  2  ......15

Row 8..15: 16 17 18 ......31

16 x 32

Row 0..7 : 0  1  2  ......31

Row 8..15: 32 33 34 ......63

16 x 48

Row 0..7 : 0  1  2  ......47

Row 8..15: 48 49 50 ......95

For each block row, bit.0 will control the LED at the top, while bit.7 will control the LED at the bottom.

For portrait orientation, we divide the LED into a column or a block of columns and rows. Counter will determine which row is active. Top row is row 0, and will be active if the reset signal is activated. The active row increase along with the clock signal acquisition.

For the 8x size, column 0 is the leftmost column and activated by bit.0. Whereas column 7 is the rightmost column and activated by bit.7

Rule for the 16x size could be analogous to the position of landscape orientation.

MXLED simulated on the system to work like a real matrix. If we managed to change the active row or column slowly, then it can be seen only eight moving lights jumping around. However, if the changes is fast enough, then the lit changes of the LED will be seen steady not blinking. So the simulation MXLED will feel like a real LED matrix.

You can see examples of the use of this MXLED in the example Showing Images on LED Matrix using Simulator

Controlling the LED matrix is actually almost the same as controlling seven segment. The difference is that the LEDs of seven segment are organized into seven sections which could form the figures. While the LED matrix, LEDs are arranged into a matrix that can be referenced by column and row. The columns on the LED matrix can be equated with the digits on the seven segment. While the lines on the LED matrix can be equated with LED a through LED g plus a decimal point LED on seven segment. Thus between the seven segment and LED matrix is basically the same, especially if the number of rows are only eight.

Then how to display images on the LED matrix? Yes basically the same as displaying the numbers on the seven segment. If you are already familiar with how to display data on the seven segment display using the buffer as is done in discussion about seven segment in the Easy and Fun Learning Microcontroller book, in which to display data on the seven segment, all we have to do is fill the display buffers, then to display image on the LED matrix, all we need to do is also just fill data on corresponding display buffers.

In the picture above, the size of the LED matrix that we use is 8 rows x 32 columns. This size is equivalent to seven segment which the number of digits are 32 digits. Thus basically we need a total of 32 bytes of display buffers. The LEDs on the top line we associate with bit 0 of port and successively LEDs on the lines below to bit 1, bit 2, and so on until bit 7. To make a LED lit on a particular line, then we need only fill the correlated bits with 0 or 1 depending on the configuration of the matrix that we use, active low or active high. For ease of understanding, we usually use active high so to lit the LED on the top line then we have to make the bit 0 to has the value of 1.

As an example, consider the logo below. The logo size is 32 x 32. Therefore, the logo may not be displayed on the LED matrix with the size of 32 x 8 at once. To display it, we have to split the logo into four parts. After that we have to show four parts one by one.

To the left of the logo there is a number from 0 to 7 and repeated up to four times. While above the logo are the numbers from 0 to 31. These figures will show us how to display the parts of the logo on the LED matrix. The figure above the logo shows the buffer index, while the figure at left of the logo shows the bit position. Thus, to show the first part of the logo (the top), then the buffer [0] must be filled with data 11111111b, buffer [1] = 00000001b, buffer [10] = 00011001b, and so on.

  0.        1.        2.        3.
  01234567890123456789012345678901
0 ©©©©©©©©©©©©©©©©©©©©©©©©©©©©©©©©
1 © .............................©
2 ©..........©©©©..©©©©..........©
3 ©........©©©©©©..©©©©©©........©
4 ©......©©©©©©©©..©©©©©©©©......©
5 ©.....©©©©...©©..©©...©©©©.....©
6 ©....©©©.....©©..©©.....©©©....©
7 ©...©©©......©©..©©......©©©...©
0 ©..©©©.......©©..©©.......©©©..©
1 ©..©©........©©..©©........©©..©
2 ©.©©©............©©............©
3 ©.©©.............©©............©
4 ©.©©©©©©©©©©©©©..©©©©©©©©©©©©©.©
5 ©.©©©©©©©©©©©©©..©©©©©©©©©©©©©.©
6 ©............©©.............©©.©
7 ©............©©............©©©.©
0 ©..©©........©©..©©........©©..©
1 ©..©©©.......©©..©©.......©©©..©
2 ©...©©©......©©..©©......©©©...©
3 ©....©©©.....©©..©©.....©©©....©
4 ©.....©©©©...©©..©©...©©©©.....©
5 ©......©©©©©©©©..©©©©©©©©......©
6 ©........©©©©©©..©©©©©©........©
7 ©..........©©©©..©©©©..........©
0 ©..............................©
1 ©..©©©.©..©..©©..©...©.©..©©©©.©
2 ©.©....©..©.©..©.©©..©.©.©.....©
3 ©.©....©©©©.©©©©.©.©.©....©©©..©
4 ©.©....©..©.©..©.©..©©.......©.©
5 ©..©©©.©..©.©..©.©...©...©©©©..©
6 ©..............................©
7 ©©©©©©©©©©©©©©©©©©©©©©©©©©©©©©©©
The Logo.

To be able to fill these data to the buffer with ease, then we can create an image data table. This technique is similar to how to show the variation of Running LED on exercise 5 of the Easy and Fun Learning Microcontroller book or on the LED Simulation article. Each image table has the same size with the size of the buffer. And for the case of the logo above, we need four 32 bytes size image tables.

Like on the Running LED exercise, the work of preparing the table is a tedious job and requires precision. But you should not be discouraged if you want to make the table of the large and complicated image because you can use the aid program ImgTable.exe. The program will help you to convert bmp format images into an image table that we can use it in the program using either the M51 format or using the SDCC format. So to get an image table, you simply create an image using MSPaint, and then do the conversion using ImgTable.exe.

ImgTable.exe can do the conversion from bmp file with a monochrome format up to 24 bits format. But because the table will be used to lit the LED which of course is equivalent to a monochrome format, the 24 bits format would be first converted by ImgTable.exe into the monochrome format. Therefore, it is better if you save the images you create with monochrome format to save on file size.

Rules applied by ImgTable is that the pixels that are black will be converted into 1, while the white pixels into 0. For colors other than black and white, then the three basic color elements will be summed first, then divided by three. After that, if the value is closer to the black will be interpreted as black, and if it is closer to white will be interpreted as white.

Table creation sequence is from left to right and from top to bottom. From top to bottom means every eight pixels will be taken as one byte. For example, an image with a size of 32 x 32, it will be four tables with the size of 32 bytes. The first table stores the data of the coordinates (0,0) to (31,7) by the rule that the upper left pixel is (0,0) and the lower right pixel is (31.31). The second table is (0,8) to (31,15). So forth.

In this example, all tables store the entire image data appropriately because the high of the image size is a multiples of eight. If the size of the image is not a multiples of eight, then the final table will have the remaining bits. For example if the image size is 32 (width) x 33 (height), then the number of tables are five tables with size of 32 bytes. And the last table only use bit 0. While bit 1 to bit 7 will be filled with 0.

Return to the LED matrix, we must display the four parts of the logo one by one. But if we display them alternately, then the shape of the logo will not be seen clearly. We will only see the incomplete images that appear alternately. Therefore we have to show by shifting it up or down. However, we have created tables that contains the data of the image blocks. So we can not retrieve data from the image coordinates (0,1) to (31,8) directly. For that, we need one more buffer with the same size as the first buffer. If we call the first buffer as the display buffer, then we call the second buffer as hidden buffer because it was not for displaying data to the LED matrix. This buffer is only used as an aid in manipulating the display data.

In case we want to shift the image up, then we can describe the arrangement as if the buffer is as follows:

01245...
abcde...
Numbers = the display buffer, alfabets = hidden buffer.

The first time, we fill the display buffer with data 0, so that all the LEDs will be turned off. Then we put the first image block table on hidden buffer. After that do a bit shifting between a and 0, b and 1, c and 2, and so on. Do these shifts of up to eight times, which means that all parts of the image block has been on display buffer. And before doing the ninth shift, we must place the second image block into the hidden buffer. And the next step is an iteration of the previous step.

For more details, you can see the program listing for the demo above that is written in the format of M51 and c. In the listing, there is DelayTime constants that you can replace to change the speed of picture motions.

To try on the simulator, then you must set the port link so that both Port0 and Port1 are linked to MXLED.exe by the message link with the WM_USER message, and lParam of Port0 is 2 and lParam of Port1 is 1. You also have to set the configuration of MXLED size as 8 x 32 landscape.

Happy trying and please do not hesitate to ask if there are less clear. Hopefully useful.

Actually, he asked me to place the digits control using P0, and P2 for segment data. However, because previously I had made the program with P0 as the control of digits and P1 for segment data, so that I don’t need to reconfigure when I tried on the simulator, I keep using P0 as the control of digits and P1 for segment data. Therefore if you have already used P0 as the control of digits and P1 for segment data, then you need to change the PortDigit and PortData constants at the top of the program.

To realize the friend request, I use the way I usually do if I want to display data on the seven segment, i.e. put seven segment manager procedure on the timer interrupt. This ensures our programs to be able to display data on the seven segment with good brightness. There are several examples of programs that put the seven segment scanning process in the main program. But we know that the seven segment which installed in a multiplex must continuously scanned to be able to display the data. If the program are less keen in doing the scanning, then the display will be disturbed. Probably will not light up, and this is the worst, or at least the light intensity of the seven segment will be dimmed. And it is very unpleasant. By using a timer interrupt to do the scanning, we can relieve our attention from the scanning process, because it has been fully handled the interrupt handler. So this way is the best way (at least I think so, you may leave a comment if you do not agree).

Then, because the program is indeed a very simple program, then we can use all facilities owned by the microcontroller without much consideration. For easier handling in counter addition or subtraction, we can use the external interrupts, i.e. ext0 and ext1. So we just attach the button at P3.2 (ext0) and P3.3 (ext1). Of course we must use the button that really bounce-free (bounceless), so that if we press the button once, then we also get only one count. To raise the count, we use ext0. As for lowering the count, we use ext1. In the main program, we do not even do anything.

Using interrupts to handle button is very simple in terms of response. But the drawback is that the button used should be a bounceless type. In addition, only P3.2 and P3.3 can be used. Therefore, using debouncing technique as I have discussed in the Easy and Fun Learning Microcontroller book will probably be a better solution in handling button. For that, I also will feature a program that handles the button not on the external interrupt, but placed on the timer interrupt. Thus, in addition to the timer interrupt is used to perform seven segment scanning is also used for button scanning.

In scanning through the timer interrupt, we can use any button, does not have to use the bounceless. We just need to adjust how quickly button bounce will be addressed. To try it, you can change the value MaxWait to 255 to slow down the response of the button bounce. If you use a simulator, and a computer that used is quite a fast computer, it may be the key response is still too fast. You simply change the TimeOutDown and TimeOutUp type to be unsigned int, and change the MaxWait value to higher rates. After that, try to press the button on Tombol.exe with high frequency. You will find that if the button is pressed too quickly, then presses the button too fast will be considered as a bounce. Thus, we can use the buttons that are not really bounce free. All you need to do is adjusting how sensitive your buttons, i.e. by setting the MaxWait value. And generally, you will not use numbers higher than 255. Therefore, the data type I use is unsigned char.

Remember, if you use Tombol.exe as simulated buttons, then you must set the used port (in this case is P3) on the simulator to be linked in Link Message. This is necessary because Tombol.exe send the keystrokes signals to the simulator through Window Message. As for the program that handles button via the external interrupts, then you can not use Tombol.exe, because Tombol.exe could not provide an interrupt signal to the simulator. You must use the existing buttons on the simulator. To display this button port window is from the View-Port menu. Then from the window, double-click on the desired port, which is the name of the port, not on its value, because if you double-click on the value, then it will appear the dialog to change the value of the port or the relevant registers.

Another friend asked me, “How do I create a program that constantly monitor keystrokes, but other operations continue to run without being interrupted by the oversight of the key”. Well, two examples of the above program, whether using an external interrupt and those that use the button scanning on the timer interrupt, has conducted monitoring of key presses continuously, while at the main program does not even do anything. Thus, the above example is also an answer to this question. You simply put the other operations in the main program, while the supervision of the button has been continuously monitored.

I wrote the program in c and M51 language only. So if you just have a Micro v5.4, which is the IDE that I include in the Easy and Fun Learning Microcontroller book, then you should first download the latest version of the IDE that supports the C language.

Happy trying, hopefully useful.

Source