software usart

[Ricardoco]

Hi all, if i was using a software usart to read data from an input pin, do i have to use a particular pin on a given chip say a 16f877 or similar, and if not is there a maximum amount of pins that could be used, not all at the same time of course, and how would i change which pin is being read.

 

[Nigel Goodwin]

Hi all, if i was using a software usart to read data from an input pin, do i have to use a particular pin on a given chip say a 16f877 or similar, and if not is there a maximum amount of pins that could be used, not all at the same time of course, and how would i change which pin is being read.

You can use as many input pins as you like, check my tutorials for suitable routines. The easiest way would be to have a seperate subroutine for each input pin, as the routine is only small this wouldn't be a problem - the routines could all share the same delay routines.

 

[Ricardoco]

thankyou now im getting somewhere although i must admit im finding this assembler a task to say the least, you couldnt by any chance lend me some IQ for a while could you nigel Oh yes thats exactly what you do do on this forum..... MORE MORE lol

 

[Ricardoco]

You can use as many input pins as you like, check my tutorials for suitable routines. The easiest way would be to have a seperate subroutine for each input pin,

Ok call me stupid but could you go through one of your routines with me and show me where you would set the input pin for the data, please.

 

[Nigel Goodwin]

You can use as many input pins as you like, check my tutorials for suitable routines. The easiest way would be to have a seperate subroutine for each input pin,

Ok call me stupid but could you go through one of your routines with me and show me where you would set the input pin for the data, please.

This is the actual receive subroutine, the port and pin are defined in SER_PORT and SER_IN.

Rcv_RS232   BTFSC   SER_PORT, SER_IN      ;wait for start bit
            GOTO    Rcv_RS232
            CALL    Start_Delay	          ;do half bit time delay
            BTFSC   SER_PORT, SER_IN      ;check still in start bit
            GOTO    Rcv_RS232
            MOVLW   0x08                  ;set up to read 8 bits
            MOVWF   Bit_Cntr
            CLRF    Rcv_Byte
Next_RcvBit CALL    Bit_Delay
            BTFSS   SER_PORT, SER_IN
            BCF     STATUS    , C
            BTFSC   SER_PORT, SER_IN
            BSF     STATUS    , C
            RRF     Rcv_Byte  , f
            DECFSZ  Bit_Cntr  , f         ;test if all done
            GOTO    Next_RcvBit
            CALL    Bit_Delay
            MOVF    Rcv_Byte, W
            RETURN

The equates in that particular example (tut7_1.asm) are:

SER_PORT	Equ	PORTB
SER_TRIS	Equ	TRISB
SER_IN		Equ	0x07
SER_OUT		Equ	0x06

For your purpose of multiple input pins, I would suggest NOT using equates, and manually setting the port and pin in each seperate subroutine to the pin you require. I would suggest naming the subroutines 'sensibly', something like 'Rcv_RS232_B1', 'Rcv_RS232_A2' etc.

 

[motion]

Hi all, if i was using a software usart to read data from an input pin, do i have to use a particular pin on a given chip say a 16f877 or similar,

You don't have to use a particular pin. In assembly code, you can write code as follows:

#define USE_THIS_PIN  PORTB,5      

           CLRC
           BTFSC   USE_THIS_PIN
           SETC
           RRF     SERIAL_REGISTER,F

and if not is there a maximum amount of pins that could be used, not all at the same time of course, and how would i change which pin is being read.

Since you don't seem to have a clue on where to begin, let me make a few assumptions on what you intend to do. First, only one pin at a time is being sampled. Second, you want to save on code and write a common serial receive routine for all port pins.

Therefore, the example above no longer applies because it is hard coded to a particular port pin. No problem. Here is a code sample that should work on all ports.

;
;  Set to 1 the bit of the particular port you want to sample
;  Set all other bits to 0.
;
PORTA_MASK EQU 20h   ; 20h-24h are arbitrary memory locations
PORTB_MASK EQU 21h
PORTC_MASK EQU 22h
TEMP       EQU 23h
SERIAL_IN  EQU 24h

          MOV   PORTA,W                ; read PORTA
          ANDWF PORTA_MASK,W           ; mask out unused pins
          MOVWF TEMP                   ; store into temporary register.          
          MOV   PORTB,W                ; read PORTB
          ANDWF PORTB_MASK,W           ; mask out unused pins
          IORWF TEMP,F                 ; store into temporary register.          
          MOV   PORTC,W                ; read PORTC
          ANDWF PORTC_MASK,W           ; mask out unused pins
          IORWF TEMP,F                 ; store into temporary register.          
          CLRC
          SKPZ
          SETC
          RRF   SERIAL_IN,F

Now, if you want to work with several port pins concurrently, there is a very old application note at the Microchip's website that shows how to do concurrent tasks. It even has a code sample for two concurrent software serial routines.

How old? It was originally written for the PIC16C5X family.

**broken link removed**

 

[Ricardoco]

Firstly thankyou to motion.

Since you don't seem to have a clue on where to begin

You are so right.

First, only one pin at a time is being sampled. Second, you want to save on code and write a common serial receive routine for all port pins.

yes only one pin at a time needs to be read, and the smaller i keep the routine the easier it will be for me to understand what is going on.
It is hard to wade through everything and just pick out what i need to do to get where i want to be with a particular project,when what seems in my head to be a simple concept seems so hard to do without the experience you guys have no matter how many times i read the data sheets. If i knew someone in person i think a real time chat would make it much easier but i dont so i come on here and if i look stupid then so be it, as long as you guys dont get bored educating me.

Now, if you want to work with several port pins concurrently, there is a very old application note at the Microchip's website that shows how to do concurrent tasks. It even has a code sample for two concurrent software serial routines.

well the way i understand it (correct me if i am wrong) if i am recieving two or more lots of data at 4800 baud each and the outgoung data
has to be at 4800 baud then i can or will have to sacrifice at least 50% of the combined incoming data just to get it out again.at the correct baud.

 

[Nigel Goodwin]

well the way i understand it (correct me if i am wrong) if i am recieving two or more lots of data at 4800 baud each and the outgoung data
has to be at 4800 baud then i can or will have to sacrifice at least 50% of the combined incoming data just to get it out again.at the correct baud.

The 16F877 has a hardware USART, you can use that as the single serial output - that takes care of all your transmission worries, without taking any time up (other than a few microseconds, with a 4MHz clock).

 

[Ricardoco]

Hi
A standard NMEA Talker sends out 1-3 sentences per second
(depending on the type of talker and length of the sentence, max 80 characters)
The standard baudrate is 4800


If i was to use just two inputs would the pic be able to collect ALL the data from two inputs, sort it and send it without missing any data?

 

[motion]

Hi
A standard NMEA Talker sends out 1-3 sentences per second
(depending on the type of talker and length of the sentence, max 80 characters)
The standard baudrate is 4800


If i was to use just two inputs would the pic be able to collect ALL the data from two inputs, sort it and send it without missing any data?

I think you should buffer one full sentence before retransmission to assure you don't mixup sentences from two channels. It would simplest to use two buffers, one per input.

 

[Ricardoco]

thanks motion, but the more i look into this the more negative i feel about using one pic to do it, im told from more than one source that reading information from more than one nmea talker (4800baud serial input with upto 80 characters) is just not possible without loosing data. i am begining to believe them. Of course if i was to use an atmel im told it is feasable but not with a pic, all the information i can get would suggest that the only way to do it with the pic chip would be to use a 12f for every nmea talker each buffering while information is read from other sources. and a master pic to collect all the information and send it out

 

[motion]

To illustrate what happens when two sentences arrive and you don't buffer:

Sentence #1: This is sentence one.
Sentence #1: Next is sentence two.

Retransmit without buffering: This Next is is sentence sentence one two.

Secondly, without buffering, it is impossible to retransmit at 4800baud when both incoming channels are also 4800baud.

Of course if i was to use an atmel im told it is feasable but not with a pic

I can write an assembly language program for a PIC16F876 running @20Mhz that can receive from a least a couple of bit-banged serial ports @4800baud. I'd sample the port pins at 4X the baud rate (19.2K per sec) under interrupt from a "Special Event Trigger". Using a 40Mhz PIC18F452 would add speed and more memory for additional channels. This would be much easier than using multiple slave PICs and trying the make them talk to a master PIC.

Having said all that, unless you have substantial experience working with PICs and interrupts, it would sound impossible to do. Using an Atmel ATMega with two hardware uarts and writing in C language would greatly simply things. So, I can understand other people coming up with the above conclusion.

BTW, here is a link to an integrated uart and rs-232 tranceiver from maxim, the maker of the popular max232. I think you can connect at least a couple of these to the spi port. This should solve the extra uart problem.

**broken link removed**

 

[Oznog]

Some PICs have 2 USARTS, starting with the PIC18F6520 for example. Just look under "Peripherals" for "2 AUSART" or "2 EUSART".
https://www.microchip.com/ParamChartSearch/chart.aspx?branchID=1004&mid=10&lang=en&pageId=74

Bit banging is quite tough on resources. It's a bit tough to perform other functions while reading the serial port, especially if you're doing oversampling on the signal.

 

[Ricardoco]

Ok im not good a wording what im thinking which may be why my coding is so bad but i will give it a go:

would a pic be fast enough to do it like this:

memory locations from A01-A99 & B01- B99


read byte from pin 1 into memory location A01
is byte "crlf"
if yes move from memory location A** to tx & clear memory reset memory location to A01
if no increment memory location A01

read byte from pin 2 into memory location B01
is byte "crlf"
if yes move from memory location B** to tx & clear memory reset memory location to A01
if no increment memory location B01

start again