The alternative was to use the external memory supported by the 68HC11 microcontroller for 64K bytes to store recorded messages in digital format. However, 8K is reserved by EPROM buffalo monitor, about 8K needed for program and data variables, some space needed for I/O registers and base RAM which gives maximum of 48K memory for data. The voice samples have to be measured in 8KHz for fairly recognizable clear voice that gives about 5-6 sec of recording capacity which is too low for my original intent of recording multiple messages where each message can take about 3-5 sec on minimum. The other alternative I proposed was to use out of chip external 128K ram to store recorded messages, but it needed to control external RAM with many address pins which seemed fairly time consuming than doable in the short implementation duration of the project in the semester. So, I finally decided to use Voice Chip for storing recorded message upto 60 sec and since it has built in A/D and converting digital to analog signal to speaker, I decided to focus more on various timing aspects of timing unit of microcontroller relevant to my project and handling multiple message operations with cascaded interrupts and logic and various control features.

The internal I/O devices that I used are Timer Input capture 1 and Timer Counter features with I/O ports and interrupt logic:

It provides the interrupt for key pressed in the keypad. The Output Enable pin of Key Encoder produces a high pulse whenever a key is pressed in keypad and it is connected to the Port A pin 2 (counting from 0) which is for Timer Input Capture 1 input. The interrupt is set for rising edge active for this device. This device generates interrupt when key is pressed and records cycle time in 2 bytes TIC1 I/O register of when interrupt is occurred. This is used in determining start and stop cycles in conjunction with number of overflows of Timer Counter for measuring the time interval of each message while recording.

This device generates the interrupt whenever the counter overflows. This interrupt is enabled when the function for Record is called and the interrupt service routine counts number of overflows occurred and determines time remaining in total and time interval of each message. The interrupt is disabled when Stop is pressed so that counting of overflows do not occur out of Record mode.

The external I/O devices that I used are:

This device is used to interface with matrix keypad. It scans rows and columns of keypad and produces Data Available pin high when key is pressed with output 4-bit code of key pressed. The Output Enable\ pin of this device is grounded to make the 4-data output pins always active The Data Available pin is connected to PortA pin2 for invoking interrupt. The 4 output pins are connected to Port E pins 0 to 3 for input.

This device is used for showing menu and other messages relevant to operation mode to the user. It is also used to ask response from the user. The eight Data pins of this device is connected to Port B eight pins. The three control pins RS,R/W,E are connected to Port A pin 3,4 and 5. The microcontroller sets the control pin R/W to 0, RS to 1 for data or 0 for instruction output and drives E low and high to enable writing of data or control instruction to LCD.

This device is used to record and store messages in multilevel storage cells. It is also used to playback recorded messages through speaker. The device is set in operational mode push button by setting M6(A6)pin to logic 1. The device is connected to microphone through series of analog circuits with resistors and capacitors. The speaker is connected to its output terminal SP+ and SP-. The three control pins CE, PD, P/R are connected to Port D pin 2,3&4 for controlling operations of Voice chip. PD low would make activate power on Voice Chip for the operations, P/R is 1 for play and 0 for record and CE will initiate the corresponding action.

This device is used to convert analog audio voice to electric signal and is connected to MIC and MICREF pins of Voice Chip with resistors and capacitors for biasing and filtering out DC .

This device is used to convert electric signal to analog audio voice and is connected to SP+ and SP- pins of Voice Chip.

All of the above devices were tested and implemented successfully in the project.

The large modifications to the 68HC11 board was not needed but extra 8K RAM was used in U2 slot for extra storage needed for extra .text (code) required by including stdio.h header files while testing with the program code.

Initially, I could not make up the LCD to show up anything even when supplying power. Later, I found out that the contrast pin (Vee) has to be low voltage. Another problem encountered was not being able to make voice chip circuit to work to record messages in operational mode setting with active message cueing mode (pin A0 of Voice Chip high) on testing. Finally, I had to make pin A0 low and used push button mode(A6 high) to make it to work to record one message and then to record multiple messages with playing it back.

1 – Record

2 – Stop

3 – Play

4 – Next

5 – Erase

6 – Time Remaining

When 2-STOP is pressed, it stops recording and go to default state to show ‘default display’

When 3-PLAY is pressed, the LCD will display

When 4-NEXT is pressed, next message will be played with corresponding PLAY message.

When all messages are played ‘default display’ will be shown.

When there are no messages, if the 3-PLAY is pressed, the LCD will display

for 3 sec and goes back to the display menu.

When 5-ERASE is pressed, the program will ask the user to erase or not.

When 7 is pressed, the messages are erased and the display will show

for 3 sec and ‘default display’ will be shown.

When 6-TIME REM. is pressed, the display will show

The main program starts and initializes direction for ports and does setting for interrupts enabling for Timer Input Capture1. It initializes LCD by initialization routine and makes it ready for data output later. It initializes the control lines of Voice Chip then it shows the main menu of different options and wait for input interrupt. When 1-Record is pressed, the interrupt service routine is called which then identifies the key pressed and calls record function which activates the Counter Overflow interrupt and clears I flag. This allows other external interrupts and sets the Voice Chip to start recording and meanwhile the microcontroller shows updated time remaining in seconds to the LCD. If the 2-Stop is pressed, another interrupt occurs and gives indication for previous Record function to stop and stops timing measurement of the present message and stores it in the correct position of the array for that message and go back to display menu. When 3-Play is pressed the Play function is invoked which checks if any message is present and if present plays first message with displaying its number and length and stops and waits for 4-Next key to be pressed. When 4-Next is pressed it goes on to play next message and the similar process occurs until the last message and then it goes back to display mode of main menu and wait state. If 5-Erase is pressed, the program asks for conformation message to erase and after getting confirmation deletes all messages otherwise just goes to default mode of menu display and wait. If 6-Time Remaining is pressed, the time remaining is converted to character and displayed to LCD and waits for key 9 to continue to go to default mode of displaying menu.

The choice between C and assembly language was made based on strength and capabilities of C and assembly language. C is more user friendly and makes it easy to handle the testing of multiple conditions and implementing multiple selective structure and looping executions. Moreover, calling function with many arguments is simple in C. So, definitely my main() function is in C, so are other functions like key_isr(), Record(), Stop(), Play(), Erase(), Time_Rem(), some LCD functions and others. Assembly language is powerful in handling program of timing operation. So, delay functions definitely are in assembly language. Also set_jmp_table() function is in assembly by its nature. Moreover, assembly language function handles type in loose manner, so it is easy to go from byte variable to int variable of 2 bytes and long types and type changing is easy in assembly and there are different strong byte manipulation instructions in assembly language. So I chose assembly language for operations of converting binary numerical value to two byte character representations for output to LCD properly. Also, in assembly, the codes are tight, takes little space, executes faster than C, so some of my LCD functions, I chose in assembly language like LCD_initialize().

I chose interrupt because I wanted to know more about how interrupts work and handling interrupts is a challenging task which I like to do. The choice of interrupt implementation can give me a feel of how regular operation is interrupted and interrupt operation is handled and previous interrupted operation is resumed. Key pressed on keypad causes interrupt on Timer Input Capture 1 and in fact, I made cascaded interrupt provisions inside Record, Play and Erase functions. Also, timer counter worked under timer overflow interrupt. Moreover, the interrupt frees the processor to do another job without having to be busy waiting for flag to be set.

In my project, while pressing record, the interrupt service routine identifies key pressed and calls record function which has to enable certain interrupts, clear flag, write to LCD before starting recording, so recording does not actually start at the instant the key is pressed , however, the delay are too minor of few tens of cycles which is negligible compared to the length of recording purpose. Also, the TIC1 register actually stores the starting time when the key is actually pressed and also the ending time when the key is actually pressed.

Initially there were many syntax errors in program functions like of same identifier but difference in capital and small letter and the debugger was not so good to identify that. Afterwards, I tried to run initialization routine on LCD and some display functions, but nothing showed up. I rechecked my initialization routine many times and compared that with specification sheet. Then I ordered up the sequence of procedures with right delays in initialization routine to finally get something in the display unit.

The other problem was to make the code to work. I had to debug many times for LCD routines to work and keypad input testing to work.. The most difficult was the main software procedures on controlling recording and playing to work. Even minor error would take a lot of time to be fetched out. I had to use many printf statements in program to find when the interrupt occurred, whether certain subroutine was called or not and some execution statements were performed or not.

The tradeoff basically occurred for keypad interface. If I were to directly connect keypad lines to input/output pins of 68HC11, I would have to make more software to scan columns while one of the rows are driven low and repeat that to each row to find the key pressed and its identification. However, if I used the 16-Key Encoder(MM74C922), the chip does the scanning of the rows and columns by itself and takes care of debouncing by certain external capacitors to its two output pins(OSC and KDM). The Encoder generates its Data Available output pin to high pulse while key is pressed, then four output data pins can be checked to identify the key pressed. So, it will reduce the software load and also take load out of controller being busy to scan most of the time. So, I chose the 16-key hardware which decreases the software load but also meanwhile increases more hardware connections and components(capacitors).

After then I went on to test input interrupt function key_isr() with keypad connected to the board through 16-Key Encoder. I had already done hardware testing of the 16-Key Encoder with keypad to light up combinations of LED’s. So, I know beforehand before software tests that hardware functions right and went on to have input key interrupt function that transfer signal values of Key Encoder’s output to output port B to light up LED’s. I found out bugs with the testing and was able to make it work.

Then I tested functions relating to record and playback on voice chip having already done hardware testing on voice chip circuit to record and play messages. Now, since I know my input interrupt functions and LCD functions work right, I tested my Record(), Stop(), Play() function for one message, I debugged many errors with the help of printf() statements at certain position in code. Then I went on to add my timer interrupt and time measurement functions on previous working codes and found out errors and corrected them. Then I added converting binary number to character representations for output to LCD on top of that. In that way, I finished testing my whole program.

The project took the alternate implementation of Initial Software/Hardware Design using Voice Chip with more emphasis on timing issues while recording and playing back. The Skip function was not used in the final implementation because it was not supported by the Voice Chip which I tested many times to support for skipping function. Instead as an alternative there is Next() function and 4-Next key that makes the Audio Repeater to move on to the next message when the present message is being finished playing. The interrupt for Data Available output of key-encoder is changed from STRA to TIC1 pin to use the Timer Input Capture function. Output Enable pin is always grounded as the microcontroller will check for four binary inputs of data only when the Data Available interrupt occurs. The control line EOM(End of Marker) from Voice Chip to microcontroller was not used because while testing it caused unnecessary interrupts even when the playing back of one message was not finished. The control line OVF from Voice Chip to microcontroller was not used because microcontroller itself keeps track of time and stops recording when time remaining goes to 0 sec. Also, XCLK line to Voice Chip was not used as I had originally thought of using Output Compare function. But as my peer reviewer Dave Dowgiallo has pointed out that the line needs the frequency input of 1024 KHz and not 8KHz as shown in specification of Voice Chip and generating 1024 Khz could not be done by output compare feature as even 1 instruction takes more than 2 cycle, so I decided to use the internal clock of Voice Chip. These modifications eliminated the need of software functions like __interrupt EOM(), __interrupt Overflow(), __interrupt void Sq_wave-OC2_isr(void), init_OC2(void) as proposed in Final Hardware/Software design. The buzzer was also not used because the messages were played one at a time and user presses key for next message instead of playing all messages continuously with beep in between.

Moreover, new enhancements were added like measuring the time length of each message while recording through use of counter overflow and timer input capture and storing them in array in corresponding message’s position in character representation which requires addition of many complex functions and some variables.

The final result that I got was the complete workable circuit with microcontroller which can store multiple messages upto 60 seconds and plays them back when desired. The user can see the updated version of time remaining in intervals of 2 sec to optimize the duration of message provided for new message on sequence of old messages. Moreover, the user can erase all messages and record new ones from the beginning and the user can press 6-Time Remaining to look at time remaining to record. Moreover, the properties of message which are message number and message length are displayed to the user when playing them back.

The possible extension would be to interface this circuitry with telephone line to build answering machine and also the capability to erase particular message in the sequence of many messages.

The lessons that I learned are to plan properly beforehand for the design as time becomes very scarce at the last minute while actually implementing the designed software and hardware to test each components for errors which can take a lot of time. Also, the other thing I learned was the significance of testing part by part and step by step which helps a lot in the design of large project.

A6. User Interface printouts

When 2-STOP is pressed, it stops recording and go to default state to show ‘default display’

When 3-PLAY is pressed, the LCD will display

When 4-NEXT is pressed, next message will be played with corresponding PLAY message.

When all messages are played ‘default display’ will be shown.

When there are no messages, if the 3-PLAY is pressed, the LCD will display

When 5-ERASE is pressed, the program will ask the user to erase or not.

When 7 is pressed, the messages are erased and the display wil show

for 3 sec and ‘default display’ will be shown.

When 6-TIME REM. is pressed, the display will show