RS232 interrupt not working (Solved)

[harlequin2]

I am trying to use interrupt driven comms on a 12F1822 with no success so far. The thing sends and receives OK without using interrupts, but doesn't seem to recognise the rcv int. What am I missing? (V 4.12)

#include <12lf1822.h>
#device ADC=10
#fuses INTRC_IO, NOWDT, NOPROTECT, NOLVP
#use delay(internal=4000000)
#use rs232(baud=9600, xmit= PIN_A4, rcv=PIN_A5)

#define ShntDrv PIN_A2 //balancing shunt driver

//Globals
volatile unsigned int8 StoredVal;

void init(){
port_a_pullups(0b00100000); //internal pullup on rcv pin (RA5)
setup_adc_ports(sAN1, VSS_VDD); //
setup_adc(ADC_CLOCK_DIV_8);
set_adc_channel(1);
enable_interrupts(GLOBAL);
enable_interrupts(INT_RDA); //serial rcv
}

#int_rda
void serial_isr(){
unsigned int8 buffer, Msg=FALSE;
buffer = getc();
putc(buffer); //test - not working!!!!
if (buffer==0xFF) Msg=FALSE; //done rcving message
if (Msg) putc(buffer);
if (buffer==01){ //01 is command to add cell voltage to the string
Msg=TRUE; //receiving message
putc(buffer);
putc(StoredVal);
}
}

void main() {
unsigned int8 i;
unsigned int16 value ;
init();
do{
value=0;
for(i=1;i<=10;i++){
delay_ms(20);
value+=Read_ADC();
}
value/=10 ; //is now average of 10 readings
value-=316;
if(value>=0xFF) value = 0xFE; //is now 0 - 254
StoredVal=(int)value; //
// putc(getc()+1); // test comms - this works OK!!!!!!


}while(TRUE);
}

[temtronic]

you've got...

enable_interrupts(GLOBAL);
enable_interrupts(INT_RDA); //serial rcv


...maybe reverse the order to


enable_interrupts(INT_RDA); //serial rcv
enable_interrupts(GLOBAL);


???

[harlequin2]

[Douglas Kennedy]

We all know the RDA interrupt only works with the hardware UART.
It is never a good idea to put a putc in a RDA interrupt. Sure it will work but it takes time to send chars and you can guess it is about the same time as it takes to receive them, RS232 is asynchronous so an inbound char can and will arrive at any time (unless you have a handshake with the transmitter). The RDA needs to be called when the inbound char arrives but if the RDA is busy doing things like outputting a char or two it will not be able to respond. The hardware has a small 2 char receive buffer so you may get one putc char out and have the hardware capture the inbound char. It's still not a good idea. The rule with interrupts is that you should put as little code in them as you can.

[harlequin2]

[PCM programmer]

I'm not at a location where I can test this, but I think it might have a
chance to work: Give the compiler what it wants, to make it generate
code for a hardware UART. Then use APFCON to switch the pins.
That's what I've done below. Try it. See if it works.

[harlequin2]

[PCM programmer]

[harlequin2]

[harlequin2]

I have reduced the code down to the following and it still doesn't work. Because I initialise buf as true, it transmits a single byte (01) but does not respond to characters sent to it ie no interrupt. I know the thing can receive because if I put "putc(getc())" in main, it sends and receives fine.

[PCM programmer]

That's right, because you're not setting the APFCON register bits to
select the alternate UART pins. The compiler won't do that for you
automatically just because you specified the alternate pins in the
#use rs232() statement.

And also, putting 'UART1' in the #use rs232() while still keeping the
pin declarations, does not work. The pin declarations override
the 'UART1' directive. The compiler does not know that the alternate
pins can work with the hardware UART. It will generate code for a
software UART if you specify those pins.

You have to do it the way I showed in my post.

[harlequin2]

[PCM programmer]

Post your latest test program.

[harlequin2]

[PCM programmer]