USART problems

[Bruce]

What's the purpose of two HSERIN's?

[Demon]

I have a variable length file. The first SERIN2 processes a RECORDTYPE, the second processes VARIABLEDATA.

That way I can have record A: 33 characters, and record B: 24 characters. My file consists of 6 different record layouts at the moment, and I'd like to process all the records in one batch. I'd like to avoid having to use 6 different PICs to load them onto EEPROM.

I can't find any reference to where the flow pin is connected on the DB-9 connector. I checked here:

http://www.klm-tech.com/technicothic...2.html#pinouts

And my best guess would be pin 8, but it's not working, neither is 7. Actually, I tried all the empty pins without success. Maybe I need a resistor like on the RX and TX pins?

Robert

[Demon]

I found this Assembler program that uses CTS and RTS:

http://www.ecomorder.com/techref/mic...tinttst-rh.htm

I kinda follow the code, sorta. At least I understand his note about how RTS and CTS have to be opposite since MAX232 will invert their value; so you set 0 to get 1 and vice versa.

I figure the same routine could be written in BASIC Pro, I'm just not sure on some of the variables and sizes (like for the input/output buffers, stuff like that).

So now I gotta tear at my MAX232 circuit, I had only used one line, leaving the other unused. I put it on a small board for the sake of convenience when moving it about my projects. Well, having no working room on it isn't proving convenient any more.

At least it's a start as to how to use USART with RTS and CTS. I can't find any references for the 'flow pin' used in SERIN2 so my options are getting slimmer.

Robert

[Bruce]

With hardware handshaking, the PC transmits data when CTS is asserted, but you'll also need to setup your PC & terminal software to use hardware handshaking.

[Bruce]

How about something like this?

' ** // PIC16F877, 20MHz crystal, MCS+ loader
' ** // Interrupt driven serial receive

' ** // Buffers up to 80 characters using the hardware
' ** // USART. Buffering ends when the EOM "~" character
' ** // is received.

' ** // Includes auto over-run reset + RCIF bit polling
' ** // for new serial characters

define  LOADER_USED 1
define  OSC 20
define  HSER_BAUD 19200 ' Set data rate
DEFINE  HSER_CLROERR 1  ' Automatically clear over-run errors
DEFINE  HSER_RCSTA 90h  ' Enable USART receive
DEFINE  HSER_TXSTA 24h	' TXSTA=%00100100. TX enable, BRGH=1 for high-speed
                        ' TXSTA=%00100000. TX enable, BRGH=0 for low-speed

GP      VAR BYTE        ' GP variable
BytesIn var byte[80]    ' Up to 80 bytes
ByteCnt var byte        ' Indicates number of bytes in buffer
X       VAR BYTE        ' GP var
EOM     CON "~"         ' "End Of Message" marker
OERR    VAR RCSTA.1     ' Alias USART over-run bit
CREN    VAR RCSTA.4     ' Alias USART continuous receive enable bit
RCIF    VAR PIR1.5      ' Alias USART received character interrupr flag bit

' ** // Zero vars & A/D off
ByteCnt = 0             ' Zero counter
ADCON1 = 7              ' A/D off, all digital

' ** // Setup serial interrupts
INTCON = %11000000		' Enable interrupts
PIE1.5 = 1				' Enable interrupt on USART

    On Interrupt Goto Main
    GOTO Doodle
    
    Disable    
Main:
    WHILE RCIF              ' If RCIF=1 there's a new character in RCREG
     BytesIn[ByteCnt]=RCREG ' Yes. Then store it in array
     If BytesIn[ByteCnt]=EOM then OutMsg ' Display on receipt of terminating char 
    ByteCnt=ByteCnt+1      ' Increment array index pointer
    WEND                    ' Exit only when RCREG is clear
    RESUME                  ' Return to normal operation prior to interrupt
    ENABLE

OutMsg:   ' Note: Interrupts are still disabled here
    HSEROUT [13,10,dec ByteCnt," Bytes Rcvd",13,10]
    HSEROUT [str BytesIn\ByteCnt,13,10]
    FOR GP=0 to ByteCnt   ' Clear array bytes 0 to ByteCnt
     BytesIn[GP]=0        '
    NEXT                 '
    ByteCnt=0             ' Reset index pointer back to first element
    WHILE RCIF            ' Trash any left over characters in RCREG buffer
     GP=RCREG             ' after EOM has been received
    WEND
    GOTO Main             ' RCIF is clear, so return
    
Doodle:  ' Waste time here waiting for serial interrupt
    HIGH PORTB.0
    FOR X = 0 TO 250
        PAUSEUS 20        ' Do not use long pauses. It slows down interrupt
    NEXT X                ' processing, and requires a slower data rate to work
    LOW PORTB.0           ' with on interrupt
    FOR X = 0 TO 250
        PAUSEUS 20
    NEXT X
    GOTO Doodle

    end

If you have a terminator on the end of your data packets, this makes the whole process pretty simple.

If you just need to PIC to tell the PC to stop sending data, connect the PIC flow control pin to the PC CTS pin.

[Demon]

That's perfect Bruce! Thanks!

I read a tutorial on controlling the DB9 pins directly, modem style, and I was going nowhere fast. Your technique is much better, I had read about it earlier but I didn't know how to tie it all together. And best of all, it uses USART instead of SERIN2.

Yup, I had an asterisk as end of record marker, but I like tilde even better. There's no chance at all that it will appear in the text so it makes a better choice.

I don't really care about telling the PC to stop sending, but I'll keep your solution in mind. My main goal was just to download a file onto EEPROM, once complete, I didn't really care about cleaning up. This will be used for one-shots, once the data is on the EEPROM, I don't alter the data.

Welp, I'm off to try this. If I got it right; MAIN is the USART I/O routine, DOODLE is waiting for input and OUTMSG is an end of record summary.

Robert


Stupidlesser as each day goes on...

[Demon]

Bruce,

Can you PM or email me your last name please?

roberthedan AT hotmail DOT com

I'd like to give credit where credit is due in the comments.

And you've also made yourself a new customer at Rentron.

Robert

[Demon]

On Interrupt Goto Main
GOTO Doodle

Disable
Main:
WHILE RCIF ' If RCIF=1 there's a new character in RCREG
BytesIn[ByteCnt]=RCREG ' Yes. Then store it in array
If BytesIn[ByteCnt]=EOM then OutMsg ' Display on receipt of terminating char
ByteCnt=ByteCnt+1 ' Increment array index pointer
WEND ' Exit only when RCREG is clear
RESUME ' Return to normal operation prior to interrupt
ENABLE

---------------------------------------------------

Isn't DISABLE an unexecutable statement? Shouldn't it be the 1st statement of Main?

Then at the bottom of OutMsg, we could GOTO the WHILE statement as it is doing now; label Main1 on Disable, and Main2 on While?

Or am I missing something?

Robert