Interfacing PIC Microcontrollers 136 ; SUBROUTINES ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; INCLUDE "LCD2.INC" ; Include display routines ; ; Convert 16 bit binary result to 5 digits ; conv MOVF CCPR1L,W ; Get high byte MOVWF Lobyte ; and store MOVF CCPR1H,W ; Get low byte MOVWF Hibyte ; and store MOVLW 06 ; Correction value BCF STATUS,C ; prepare carry flag ADDWF Lobyte ; add correction BTFSC STATUS,C ; and carry INCF Hibyte ; in required CLRF Tents ; clear ten thousands register CLRF Thous ; clear thousands register CLRF Hunds ; clear hundreds register CLRF Tens ; clear tens register CLRF Ones ; clear ones register ; Subtract 10000d (2710h) and count sub10 MOVLW 010 ; get low byte to sub BSF STATUS,C ; get ready to subtract SUBWF Lobyte ; sub 10h from low byte BTFSC STATUS,C ; borrow required? GOTO sub27 ; no - sub high byte MOVF Hibyte,F ; yes - check high byte BTFSS STATUS,Z ; zero? GOTO take1 ; no - take borrow MOVLW 010 ; yes - load low byte to add BCF STATUS,C ; get ready to add ADDWF Lobyte ; restore low byte GOTO subE8 ; next digit take1 DECF Hibyte ; take borrow sub27 MOVLW 027 ; get high byte to sub BSF STATUS,C ; get ready to subtract SUBWF Hibyte ; sub from high byte BTFSS STATUS,C ; borrow taken? GOTO done1 ; yes - restore remainder INCF Tents ; no - count ten thousand GOTO sub10 ; sub 10000 again done1 MOVLW 010 ; restore BCF STATUS,C ; get ready to add ADDWF Lobyte ; restore low byte BTFSC STATUS,C ; Carry into high byte? INCF Hibyte ; yes - add carry to high byte MOVLW 027 ; restore ADDWF Hibyte ; high byte ; Subtract 1000d (03E8) and count subE8 MOVLW 0E8 ; get low byte to sub BSF STATUS,C ; get ready to subtract SUBWF Lobyte ; sub from low byte BTFSC STATUS,C ; borrow required? GOTO sub03 ; no - do high byte MOVF Hibyte,F ; yes - check high byte BTFSS STATUS,Z ; zero? GOTO take2 ; no - take borrow MOVLW 0E8 ; load low byte to add BCF STATUS,C ; get ready to add ADDWF Lobyte ; restore low byte GOTO sub64 ; next digit take2 DECF Hibyte ; take borrow sub03 MOVLW 03 ; get high byte BSF STATUS,C ; get ready to subtract SUBWF Hibyte ; sub from high byte BTFSS STATUS,C ; borrow taken? GOTO done2 ; yes - restore high byte INCF Thous ; no - count ten thousand GOTO subE8 ; sub 1000 again done2 MOVLW 0E8 ; restore BCF STATUS,C ; get ready to add ADDWF Lobyte ; restore low byte BTFSC STATUS,C ; Carry into high byte? INCF Hibyte ; yes - add carry to high byte MOVLW 03 ; restore ADDWF Hibyte ; high byte Program 6.3 Continued Else_IPM-BATES_CH006.qxd 6/29/2006 11:37 AM Page 136 Calculate, Compare & Capture 137 ; Subtract 100d (064h) and count sub64 MOVLW 064 ; get low byte BSF STATUS,C ; get ready to subtract SUBWF Lobyte ; sub from low byte BTFSC STATUS,C ; borrow required? GOTO inchun ; no - inc count MOVF Hibyte,F ; yes - check high byte BTFSS STATUS,Z ; zero? GOTO take3 ; no - take borrow MOVLW 064 ; load low byte to add BCF STATUS,C ; get ready to add ADDWF Lobyte ; restore low byte GOTO subA ; next digit take3 DECF Hibyte ; take borrow inchun INCF Hunds ; count hundred GOTO sub64 ; sub 100 again ; Subtract 10d (0Ah) and count, leaving remainder subA MOVLW 0A ; get low byte to sub BSF STATUS,C ; get ready to subtract SUBWF Lobyte ; sub from low byte BTFSS STATUS,C ; borrow required? GOTO rest4 ; yes - restore byte INCF Tens ; no - count one hundred GOTO subA ; and repeat rest4 ADDWF Lobyte ; restore low byte MOVF Lobyte,W ; copy remainder MOVWF Ones ; to ones register RETURN ; done ; ; Display period in microseconds ; disp BSF Select,RS ; Set display data mode MOVLW 'T' ; Time period CALL send ; Display it MOVLW ' ' ; Space CALL send ; Display it MOVLW '=' ; Equals CALL send ; Display it MOVLW ' ' ; Space CALL send ; Display it ; Supress leading zeros MOVF Tents,F ; Check digit BTFSS STATUS,Z ; zero? GOTO show1 ; no - show it MOVF Thous,F ; Check digit BTFSS STATUS,Z ; zero? GOTO show2 ; no - show it MOVF Hunds,F ; Check digit BTFSS STATUS,Z ; zero? GOTO show3 ; no - show it MOVF Tens,F ; Check digit BTFSS STATUS,Z ; zero? GOTO show4 ; no - show it MOVF Ones,F ; Check digit BTFSS STATUS,Z ; zero? GOTO show5 ; no - show it ; Display digits of period show1 MOVLW 030 ; Load ASCII offset ADDWF Tents,W ; Add digit value CALL send ; Display it show2 MOVLW 030 ; Load ASCII offset ADDWF Thous,W ; Add digit value CALL send ; Display it show3 MOVLW 030 ; Load ASCII offset ADDWF Hunds,W ; Add digit value CALL send ; Display it show4 MOVLW 030 ; Load ASCII offset ADDWF Tens,W ; Add digit value CALL send ; Display it show5 MOVLW 030 ; Load ASCII offset ADDWF Ones,W ; Add digit value CALL send ; Display it Program 6.3 Continued Else_IPM-BATES_CH006.qxd 6/29/2006 11:37 AM Page 137 SUMMARY 6 • The calculator demo performs single-digit arithmetic, using a keypad input and LCD display • The output pulse generator uses hardware timer compare mode and a virtual oscilloscope to display the output • Input period measurement uses the timer in capture mode and a virtual sig- nal generator to provide the input ASSESSMENT 6 Total (40) 1 Describe briefly how the keypad code is generated in the CALC application. (3) 2 State the advantages of using an include file for the LCD driver routines in the CALC application. (3) Interfacing PIC Microcontrollers 138 ; Show fixed characters MOVLW ' ' ; Space CALL send ; Display it MOVLW 'u' ; micro CALL send ; Display it MOVLW 's' ; secs CALL send ; Display it MOVLW ' ' ; Space CALL send ; Display it MOVLW ' ' ; Space CALL send ; Display it ; Home cursor BCF Select,RS ; Set display command mode MOVLW 0x80 ; Code to home cursor CALL send ; Do it RETURN ; done ; ; MAIN LOOP ; start CALL inid ; Initialise display BANKSEL PIE1 ; Select Bank 1 BSF PIE1,CCP1IE ; Enable capture interrupt BANKSEL PORTD ; Select Bank 0 BCF PIR1,CCP1IF ; Clear CCP1 interrupt flag loop CALL conv ; Convert 16 bits to 5 digits CALL disp ; Display period in microsecs GOTO loop END ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; Program 6.3 Continued Else_IPM-BATES_CH006.qxd 6/29/2006 11:37 AM Page 138 3 Outline the process required to display the negative result of a subtraction correctly on an LCD. (3) 4 State the meaning of the term ‘capture’ mode in an MCU timer. (3) 5 Explain why the main PULSE program consists of only one statement. (3) 6 Explain why the initial value in the timer preload registers in the PULSE program is 2710h. (3) 7 Explain why the ‘INCF CCPR1H’ instruction is needed in PULSE. (3) 8 Describe briefly the role of the carry flag in the division process. (3) 9 Explain the meaning of the term ‘compare’ mode in an MCU timer. (3) 10 Identify the Timer 1 interrupt flag by register and bit label. (3) 11 Describe the process for converting an 8-bit binary number to three ASCII digits for display. (5) 12 Describe the process for converting 16-bit binary into 5-digit BCD. (5) ASSIGNMENTS 6 6.1 BCD Addition Outline a process to read a sequence of decimal keys into the MCU which is terminated with the ‘ϩ’ key. A further sequence of digits is terminated with the ‘ϭ’ key. The numbers must then be added and the result displayed. Take ac- count of the fact that the numbers may not be the same length. Do not write a source code program. 6.2 CCP Control Refer to the PIC 16F877 data manual. Check the setup codes for capture and compare modes, and identify the function of each bit in the control registers initialised in the demo programs. Then check the setup of the interrupt in each program, and again identify the function of each relevant bit in the control reg- isters. Explain the significance of each setup option. Why is the use of the hardware timer helpful in these applications? Consider if there are any alter- native methods to achieve the same results. Calculate, Compare & Capture 139 Else_IPM-BATES_CH006.qxd 6/29/2006 11:37 AM Page 139 6.3 Pulse Detection A remote control receiver module generates a pulse whose length is controlled by the lever position on the transmitter console. The controller in the receiving system needs to determine the pulse length as short, mid-length or long. Outline a routine to check if an input pulse is longer than 1.4 ms, shorter that 1.2 ms, or in between these limits, switching on a ‘long’ or ‘short’ output bit accordingly, with neither for a mid-length pulse. Interfacing PIC Microcontrollers 140 Else_IPM-BATES_CH006.qxd 6/29/2006 11:37 AM Page 140 7 Analogue Interfacing Many control applications require the measurement of analogue variables, such as voltage, temperature, pressure, speed and so on. Selected PIC MCUs incorporate analogue inputs, which are connected to an analogue to digital converter (ADC); this outputs a 10-bit binary representation of an input volt- age. This result is then accurate to 1 part in 1024 (2 10 ), better than 0.1% at full scale, and precise enough for most purposes. In some cases, it is only neces- sary to use 8 bits of the conversion, which gives an accuracy of 1 part in 256 (<0.5%). In this chapter, programs to handle 8-bit and 10-bit data will be presented, and the additional software overhead required to achieve the higher accuracy can be seen. The ADC is controlled from special function registers ADCON0 and ADCON1, and can generate a peripheral interrupt if required. The output from the converter is stored in ADRESH (analogue to digital conversion result, high byte) and ADRESL (low byte). 8-bit Conversion The processing for an 8-bit result is simpler, so this will be described first. The ADC converts an analogue input voltage in the range 0–2.55 V to 10-bit binary, but only the upper 8 bits of the result are used, giving a resolution of 10 mV per bit (1/256 ϫ 2.56 V). Else_IPM-BATES_ch007.qxd 6/29/2006 11:38 AM Page 141 141 8-bit Conversion Circuit A test circuit to demonstrate 8-bit conversion and display is shown in Figure 7.1. The 16F877 MCU has eight analogue inputs available, at RA0, RA1, RA2, RA3, RA5, RE0, RE1 and RE2. These have alternate labels AN0–AN7 for this function. RA2 and RA3 may be used as reference voltage inputs, setting the minimum and maximum values for the measured voltage range. These inputs default to analogue operation, so the register ADCON1 has to be initialised explicitly to use these pins for digital input or output. INPUT & OUTPUT The test voltage input at RA0 (analogue input AN0) is derived from a pot across the 5 V supply. A reference voltage is provided at RA3 (AN3), which sets the maximum voltage to be converted, and thus the conversion factor required in the software. The minimum value defaults to 0 V. The 2.7 V zener diode provides a constant reference voltage; it is supplied via a current limit- ing resistor, so that the zener operates at the current specified for optimum voltage stability. This is then divided down across the reference voltage pot RV1 and a 10k fixed resistor. The range across the pot is about 2.7–2.4 V, and is adjusted for 2.56 V, which gives a convenient conversion factor. The LCD is connected to Port D to operate in 4-bit mode and display the voltage, as described in Chapter 4. Interfacing PIC Microcontrollers 142 Figure 7.1 8-bit analogue input test circuit Else_IPM-BATES_ch007.qxd 6/29/2006 11:38 AM Page 142 ADC OPERATION A block diagram of the ADC module is shown in Figure 7.2. The inputs are connected to a function selector block which sets up each pin for analogue or digital operation according to the 4-bit control code loaded into the A/D port configuration control bits, PCFG0–PCFG3 in ADCON1. The code used, 0011, sets Port E as digital I/O, and Port A as analogue inputs with AN3 as the positive reference input. The analogue inputs are then fed to a multiplexer which allows one of the eight inputs to be selected at any one time. This is controlled by the three ana- logue channel select bits, CHS0–CHS2 in ADCON0. In this case, channel 0 is selected (000), RA0 input. If more than one channel is to be sampled, these select bits need to be changed between ADC conversions. The conversion is triggered by setting the GO/DONE bit, which is later cleared automatically to indicate that the conversion is complete. ADC CLOCK The speed of the conversion is selected by bits ADSC1 and ADSC0. The ADC operates by successive approximation; this means that the input voltage is fed to a comparator, and if the voltage is higher than 50% of the range, the MSB of the result is set high. The voltage is then checked against the mid-point of the remaining range, and the next bit set high or low accordingly, and so on for 10 bits. This takes a significant amount of time: the minimum conversion time is 1.6 s per bit, making 16 µ s for a 10-bit conversion. The ADC clock speed must be selected such that this minimum time requirement is satisfied; the MCU clock is divided by 2, 8 or 32 as necessary. Our simulated test circuit is clocked at 4 MHz. This gives a clock period of 0.25 s. We need a conversion time of at least 1.6 s; if we select the divide by 8 option, the ADC clock period will then be 8 ϫ 0.25 = 2 s, which is just longer than the minimum required. The select bits are therefore set to 01 (Figure 7.2 (b)). SETTLING TIME The input of the ADC has a sample and hold circuit, to ensure that the voltage sampled is constant during the conversion process. This contains an RC low- pass filter with a time constant of about 20 s. Therefore, if the input voltage changes suddenly, the sample and hold circuit will take time to respond. This needs to be taken into account, depending on the type of signal being measured. If sampling speed is not critical, a settling time delay of at least 20 s should be included in the conversion sequence. In the test circuit, this is not a problem. RESULT REGISTERS When the conversion is complete, the result is placed in the result register pair, ADRESH and ADRESL, the GO/DONE bit cleared by the ADC controller, Analogue Interfacing 143 Else_IPM-BATES_ch007.qxd 6/29/2006 11:38 AM Page 143 Interfacing PIC Microcontrollers 144 (a) (b) Register Setting Flags Function ADRESH XXXX XXXX ADC result high byte ADRESL XXXX XXXX ADC result low byte ADCON0 0100 0X01 ADCS1,0 GO/DONE, ADON Conversion frequency select ADC start, ADC enable ADCON1 0 000 0011 ADFM, PCFG3-0 Result justify, ADC input mode control INTCON 1100 0000 GIE,PEIE Peripheral interrupt enable PIE1 0100 0000 ADIE ADC interrupt enable PIR1 0100 0000 ADIF ADC interrupt flag (c) ADFM = 1 Right justified ADFM = 0 Left justified R = Result bits - + Analogue to Digital Converter ADC MUX ADC Control Registers (ADCON0, ADCON1) Channel select bits Analogue Inputs RA0 RA1 RA2 RA3 RA5 RE0 RE1 RE2 External reference voltages Select external or internal reference voltage. Input Function Select Set mix of analogue or digital inputs GO/ DONE Divider Clock rate select System clock Vss Vdd Vadc Internal reference voltages RA3 RA2 ADRESH ADRESL ADIF 0000 00 RR RRRR RRRR RRRR RRRR RR00 0000 ADRESH ADRESL Figure 7.2 ADC operation: (a) ADC block diagram; (b) ADC control registers; (c) result regis- ters configuration Else_IPM-BATES_ch007.qxd 6/29/2006 11:38 AM Page 144 and the ADIF interrupt flag is set. Since the result is only 10 bits, the posi- tioning in the 16-bit result register pair can be selected, so that the high 8 bits are in ADRESH (left justified), or the low 8 bits are in ADRESL (right justi- fied) (Figure 7.2 (c)). Obviously, to retain 10-bit resolution, both parts must be processed, so right justification will probably be more convenient in this case. If only 8 bits resolution is required, the process can be simplified. If the result is right justified, the low 8 bits in ADRESL will record the low bits of the conversion, meaning that only voltages up to 25% of the full range will be processed, but at full resolution. If the result is left justified, the high byte will be processed, which will represent the full voltage range, but at reduced resolution. In our test circuit, the reference voltage is 2.56 V, and the justify bit ADFM = 0, selecting left justify. Only ADRESH then needs to be processed, giving re- sults for the full range at 8-bit resolution, which is about 1% at mid-range. The result will be shown on the LCD as 3 digits, 0.00–2.55. The test input pot gives 0–5 V, but only 0–2.50 will be displayed. Over range inputs will be displayed as 2.55 V. 8-bit Conversion Program The test program is outlined in Figure 7.3, and the source code listed in Program 7.1. The output port and ADC control registers are initialised in the first block, with the LCD include file providing the display initialisation, and driver routines. The main loop contains subroutine calls to read the ADC input, convert from binary to BCD and display it. The routine to read the ADC sets the GO/DONE bit and then polls it until it is cleared at the end of the conver- sion. The 8-bit result from ADRESH is converted to three BCD digits by the subtraction algorithm described previously. Full-scale input is 255, which is displayed as 2.55 V. 10-bit Conversion Figure 7.4 shows a circuit, which demonstrates 10-bit, full resolution, analogue to digital conversion. The reference voltage circuit now provides a reference of 4.096 V, giving a wider range of 0–4.095 V. With this reference voltage, and a maximum binary result of 1023 (2 10 −1), the conversion output will increase at 4 mV per bit. The result is displayed as a 4-digit fixed point decimal. The reference voltage circuit is a little different from the 8-bit circuit. The zener voltage is divided down using fixed-value resistors, and the final voltage tweaked by adjusting the current to the zener. This gives a finer adjustment than using the calibration pot in the voltage divider chain. Analogue Interfacing 145 Else_IPM-BATES_ch007.qxd 6/29/2006 11:38 AM Page 145 . Interfacing PIC Microcontrollers 136 ; SUBROUTINES ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; INCLUDE "LCD2.INC" ; Include display routines ; ; Convert 16 bit binary. bit accordingly, with neither for a mid-length pulse. Interfacing PIC Microcontrollers 140 Else_IPM-BATES_CH006.qxd 6/29/2006 11:37 AM Page 140 7 Analogue Interfacing Many control applications require. the GO/DONE bit cleared by the ADC controller, Analogue Interfacing 143 Else_IPM-BATES_ch007.qxd 6/29/2006 11:38 AM Page 143 Interfacing PIC Microcontrollers 144 (a) (b) Register Setting Flags