Getting SERIN2 to work?

[BobPigford]

Joe S. and mackrackit,
I have some success by going farther back to basics (sending ASCII numbers from a terminal). I'll try to post the code and results later today.
Thank you for your help and patience,
Bob

[BobPigford]

Joe S. and mackrackit,

Background: I had written a program with PBCC (PowerBASIC Console Compiler for WinXP) that:
1) Waits for the PIC to send a "0" to the PC, then
2) the PC sends 12 consecutive 16-bit words to the PIC expecting the PIC to receive them and do other things with them.
I had never been able to get the PIC to receive them. It would always just "hang up" at the SERIN2 statement.

I went back to basics and worked to send a set of characters (numeric) to the PIC via a PC terminal program and had the PIC send them back. I (finally) got it to work just fine:

My TX-RX code for this simple test is:
'--------------------------------------------------------
INCLUDE "MODEDEFS.BAS"

ADCON1 = 7 'set analog pins to digital
TRISA = 1 'set port A to inputs
TRISB = 0 'set port B to outputs
TRISC = 0 'make it an OUTPUT for serial comm on it's pins, even though
'RX is sort of an INPUT.

DEFINE OSC 20

MD VAR WORD
LedPin VAR PORTC.3
RXData VAR WORD
RXPin VAR PORTC.5
TXPin VAR PORTC.4

MD = 16780 '16780 = 2400 driven, 8 bit, 1 stop, no parity, inverted

main:

SERIN2 RXpin, MD, [DEC RXdata] 'waits for a character then moves on

SEROUT2 TXpin, MD, ["I am sending it back -> ", DEC RXdata, 10, 13]

HIGH LedPin 'hook LED to Pin 14 = RC3
PAUSE 20 'a flash of the LED means we got
LOW Ledpin 'this far OK
PAUSE 20

GOTO main
'--------------------------------------------------------

I got this to work with Hyperterm and with a simple terminal program I wrote in PBCC (PowerBASIC Console Compiler for WinXP).
I used a Serial Port Monitor to look at all the bytes going and coming.

What I learned:
1) I need a 1k resistor in the TX line and another 1k resistor in the RX line (I am not using an RS232 converter). Gee, this is exactly what the PBP manual said to do! Other references to a 22k or 10k resistor in the RX line were not good for me. What a wild goose chase that was. It really masked the problem for a long time.
2) The PC transmit end must be set for 8-bit whenever the DEC qualifier is used for variables within the SERIN2 statement. A transmitted 7-bit byte will not do (even though a 7-bit byte can still represents an ascii number).
3) When I use the qualifer DEC, SERIN2 will accept ONLY NUMERIC ascii characters, and it will keep accepting them until any NON-numeric character is received. THIS IS AS IT HAS TO BE. I turns out that my PBCC program had been sending a space character prior to the string of numeric characters. Then DEC caused that variable to be bypassed and nothing was stuffed into that variable at all.
4) The character that tells SERIN2 (with DEC) that the string of numeric characters has ended can be any non-numeric character. It does not have to be a CR.
5) When testing the TX-RX code, remember to send NUMERIC keystrokes in addition to alpha keystrokes. Afterall, wasn't it NUMERIC values I was trying to stuff in there? How could I have missed that one?
6) Breaking my 16-bit value into two bytes (high and low) then sending them to the PIC while SERIN2 uses the DEC qualifier will not work either! That technique must need some other qualifier (or no qualifier ... will have to try that again later).
7) If I set the baud rate mode via a variable, and that mode number is larger than a single byte's worth (like 16780) , then that variable better be defined as a WORD, not as a BYTE. Yet another noobie mistake!
8) SERIN2 is UNFORGIVING and will tell you nothing if:
* DEC is used and if
* Non-numeric STRING characters are sent, or
* If a non-numeric string character is sent
before the numeric string characters are sent.
9) SERIN2 with the DEC qualifier will accept lots of numeric characters, but if they represent a number higher than 65535 (max 16-bit word), the numbers that are stuffed into the variable will not be accurate.
10) While one would think that a Receive pin should be set as an INPUT, no, set both TX and RX pins as outputs. Learned that one by luck.

Now I can even get my 12 word transmission to work with the PIC. Hooray! Now I will move on to making the PIC program do more than receive the needed values. Progress at last!

It has been a painful learning experience. But I am stronger for it, I suppose.
Once I understood it, the PBP manual was right afterall. Ha! How dumb could I be?

Many thanks for your suggestions. One of the reasons I did this long post was to (hopefully) help other beginners navigate around some of these TX/RX and SERIN2 DEC traps.

Best regards to all,
Bob

[Archangel]

My hat's off to you Bob, I don't know where you found that, it's not in the manual that I can find . . . But it works! I've spent 1/2 the night trying to make it work some other way and NADA. You have earned your keep, fixing your own problem . . . Thanks for the lesson !

[BobPigford]

I looked at the resistors that I used successfully. My previous post was in error with regard to the resistor value used in the TX line from the PIC. I used a 10k, not a 1k. I don't know if a 1k will work in the TX line, but I know the 10k works.

I still use the 1k in the RX line to the PIC, though.

I am next going to install an RS232 converter to avoid all this resistor roulette!

Best regards to all,
Bob

[Bronurstomp]

Bob,

I've used the ubiquitous MAX232 to a PC RS232 port with success in an automated tester using Hserin and Hserout. I'm jumping on on this cause I use the PIC16F876A also.
The DB9 on this end is Female so you can use a standard male to female cable to the PC.
Here's some syntax and I'll drop in the related schematic. I have it set at 4800 baud, but I'm sure it'll run much faster. I deleted a bunch of stuff from the diagram, but I left the ISP hookup, as well as the reset button.

This bit of code will wait for *T from the PC, run the TECH subroutine, then after sending the A/D results, waits for 3 numbers, converts them into decimal, uses that for the PWM value, then waits for FN to go on it's way.
The formating stuff in the HSEROUT code is to pad 0s cause the PC (running Labview) was looking for precisely 3 bytes at that point.

DEFINE OSC 8'  		        USE 8MHZ CRYSTAL OSC.			
DEFINE ADC_BITS 10'	        +VREF= 2VDC 1 = .00195V OR ~ 2MV
DEFINE ADC_SAMPLEUS 50'   SAMPLE TIME .1MS
DEFINE CCP1_REG PORTC' 		Hpwm channel 1 pin port		
DEFINE CCP1_BIT 2' 		    Hpwm channel 1 pin bit		 
DEFINE CCP2_REG PORTC'		Hpwm channel 2 pin port		 
DEFINE CCP2_BIT 1'		    Hpwm channel 2 pin bit		 
DEFINE HSER_RCSTA 90h'	Set RECIEVE Enabled     
DEFINE HSER_TXSTA 24h'	Set XMIT Enabled HS     
DEFINE HSER_BAUD 4800'		IL1700 BAUD RATE		 
DEFINE HSER_CLROERR 1'		CLEAR OVER RUN ERRORS		 
DEFINE CHAR_PACING 1000'	CHAR PACING(uSEC) TO 1MS		 
TRISA=%00001111'	SET UP PORTA I/O (1=IN,0=OUT)		 
TRISB=%11001000'	SET UP PORTB I/O 			 
TRISC=%10000000'	SET UP PORTC I/O			 
ADCON1=%10100101'	USE PORTA.3 PIN5 AS VREF+ @ 2.0VDC 	 
ADCON0=%01000000'	A/D OFF -- RIGHT JUSTIFY. 6MSB=0 ADRESH	
 HSERIN[WAIT("*"),BUNDONE]
        IF BUNDONE = "T" THEN GOTO TECH ' T FOR TECH

TECH:'                  ****************  TECH ************
'
BLEDOUT = 0'		          TURN BLUE LED ON
HPWM 1, BLUELVL, 5000'		    CCP1=BLUE SET CURRENT		
ADCON0 = %01001101'	              BLUE ADC		
ADCIN 1, BLUSNS'                             READ IT				
BLEDOUT = 1'		          TURN BLUE OFF			
RLEDOUT = 0'			   TURN RED ON
HPWM 2, REDLVL, 5000'		   CCP2 =RED SET CURRENT	
ADCON0=%01000001'		 RED A/D		
ADCIN 0, REDSNS:REDSNS=REDSNS MIN 999	
RLEDOUT = 1'			    TURN RED OFF
'HPWM 1, 0, 5000'          STOP PWM 
'HPWM 2, 0, 5000 
IF BLUELVL <10 THEN
HSEROUT["00",DEC BLUELVL]:GOTO RED1
ENDIF
IF BLUELVL <100 THEN
HSEROUT["0",DEC BLUELVL]:GOTO RED1
ELSE
HSEROUT [DEC BLUELVL]
ENDIF
GOTO RED1
RED1:
 IF REDLVL <10 THEN '           FORMATING
HSEROUT["00",DEC REDLVL]:GOTO BLU1
ENDIF
 IF REDLVL <100 THEN
HSEROUT["0",DEC REDLVL]:GOTO BLU1
ELSE
HSEROUT [DEC REDLVL]
 ENDIF
GOTO BLU1
BLU1:
    IF BLUSNS <10 THEN
    HSEROUT["00",DEC BLUSNS]:GOTO RED2
    ENDIF
    IF BLUSNS <100 THEN
    HSEROUT["0",DEC BLUSNS]:GOTO RED2
    ELSE
    HSEROUT [DEC BLUSNS]
    ENDIF
GOTO RED2
RED2:
     IF REDSNS <10 THEN
    HSEROUT["00",DEC REDSNS]:GOTO WAITF
    ENDIF
    IF REDSNS <100 THEN
    HSEROUT["0",DEC REDSNS]:GOTO WAITF
    ELSE
    HSEROUT [DEC REDSNS]
     ENDIF
GOTO WAITF
WAITF:
HSERIN [WAIT ("F"),EFF]
IF EFF = "N" THEN 
GOTO BEGIN
    ELSE
     GOTO BLULEDSET
      ENDIF
    GOTO TECH
BLULEDSET:'----------------------SET BLUE--------------------------
HSERIN [WAIT("@"),STR INCOM\3]
INCOM[0] = (INCOM[0]-"0") '
INCOM[1] = (INCOM[1]-"0") ' Convert to decimal
INCOM[2] = (INCOM[2]-"0") 
X = 0 '                     RESET THE VARIABLE             
X = X + INCOM[0] * 100 
X = X + INCOM[1] * 10 
X = X + INCOM[2]
BLUELVL = (X MIN 255)
X = 0
    HPWM 1, BLUELVL, 5000'		   CCP1  AND SET CURRENT

Then more of the same for the Red channel.
I hope this helps, and I pasted everything correctly!

Bron