RS232 receive

[[email protected]]

I used the following simple program for an experiment

#include <16F628A.h>

#FUSES NOWDT //No Watch Dog Timer
#FUSES NOLVP
#FUSES HS
#use delay(clock=8000000)
#use rs232(baud=9600,parity=N,xmit=PIN_B5,rcv=PIN_B2,bits=8,enable=PIN_B3)

void main()
{



printf("Enter a char");
while (true) {
putc(getc());
}
}

PC received the string "Enter a char" but the progrma stuck on getc(). No data received. Wiring and RS232 interface were carfully checked and 100% no problem. I used 18F1320 for test again, the result was exactly the same. I am using PCWH 3.249.

Is there a possibility this is caused by the compiler bug?

[kender]

By design, getc() waits until a character is received. The program will hang in the getc() until it receives a character. This is a normal behavior and not a bug. If you don't want the program to wait for the character, use kbhit() to check if there is a character available.

[ckielstra]

[[email protected]]

Sorry. The description is not clear. The program stuck on getc() no matter how many times i hit the keyboard. kbhit() funtion always return zero.

[[email protected]]

I am using LT1280 for RS232 interface. The On/off pin is connected to RB3, therefore I need to add enable=PIN_B3 to #rs232 clause

For testing with 18F1320, the tx and rx pins are updated in #rs232 clause.

The problem is why kbhit() always return zero and getc() is keep on waiting for input (because kbhit() reurns zero) no matter how many times i hit the keyboard?

[Ttelmah]

The reason is simple. You are turning off the LT1280, by using the enable pin.
If you want to receive, the chip _must_ be turned on. The 'enable' output on the serial connection, is for things like RS485 drivers, where you can turn off the _driver_ component only, and leave the line idle, when not sending, with the reciever still wroking. The PIC is doing exactly what you are telling it, and switching the driver off, when you are not sending. On the LT1280, the on/off pin, controls both the driver and the receiver, so you never see the incoming characters.

Best Wishes

[Guest]

Does this means that teh following program will work?

#include <16F628A.h>

#FUSES NOWDT //No Watch Dog Timer
#FUSES NOLVP
#FUSES HS
#use delay(clock=8000000)
#use rs232(baud=9600,parity=N,xmit=PIN_B5,rcv=PIN_B2,bits=8)

void main()
{

set_adc_ports(no_analogus)l
output_high(PIN_B3);
printf("Enter a char");
while (true) {
putc(getc());
}
}

[nasbyc]

[temtronic]

Your hardware is too complicated.
Just use the folowing.

PIC<- max232-> PC.

example wiring is in the CCS Help files.

There is no need for the RS485 chips in this setup.

Also, you can get rid of the 'enable' option in the use RS232(...) code.

[nasbyc]

[temtronic]

That 'system' is much better and far easier to work with .
Work in stages..
You can get the 'master' PIC 'up and running' , then add the slave PICs as required.Just be sure to add the RS485 bus resistors.

[Ttelmah]

The problem here, is that the thread is about RS232, not RS485. The problem of hijacking a thread....

Now, step back. Original thread is about RS232, using a LT1280 driver. The LT1280, has an 'enable' pin, which turns the whole chip on/off, for power saving. If this is connected to an 'enable' output from the PIC, then when the chip is turned off, you can't receive.

RS485 devices are different. They normally have separate 'receive enable', and 'transmit enable' pins, and these are usually opposite polarity. If you look at the standard MAX485, it has a /RE (signal low, enables receive), and DE (signal high enables transmit) pins. With these, you need the following wiring:

1) Connect /RE and DE together.
2) Connect these to the enable output of the PIC.
3) Have a _pull up_ resistor on the RO pin of the device, connecting to he RX pin of the PIC.
4) Have a terminating resistor between the A & B lines at each end of the bus.
5) Have two bias resistors, ensuring that the 'B' line is 200mV above the 'A' line when the bus is not driven. Typically the three resistors can be (for instance) 1.2KR from 5v to line B, then 120R between A and B, then 1.2KR to ground from A. This gives 100R termination (ideal for twisted pair wiring), and the bias as well.

Now on this biasing, with _some_ driver IC's it is not needed (devices exist where an undriven bus is warranted to give a 'high' output), but in general it is safer to have it. Maxim do also do active termination devices, that also provide the bias as well.

What happens then is that the DE pin is driven high _only_ when the PIC is sending data. When this happens the /RE pin turns off the receiver, and the pull up resistor ensures the PIC see nothing when the transmission is being done. As soon as the transmission finished, DE goes low (turning off the transmitter), and /RE also goes low _turning on the receiver_. The bias resistors ensure that nothing is received when the bus is idle, but since the receiver is now 'on', if somebody else sends something, this is received.

Remember though you should only have one set of terminators at each end of the bus.

If you are running with a low data rate, termination may not be needed, but biasing still is.

Best Wishes

[nasbyc]

[Ttelmah]

You need to be operating the /RE and DE pins at the PC end.
Easiest solution, get yourself a RS485 adapter for the PC, rather than trying to go RS232-TTL-RS485. Devices that support half duplex, will already control the enable pin for you.
If you really 'must' use RS232 on the PC, then you need some extra circuitry. After you have the Rs232 converted to TTL async, you need to add a monostable multivibrator, or a simple 8pin PIC, setup so that when the falling edge on the TTL serial from the PC is seen, it operates the /RE and DE pins for 9.5 bit times. Remember you also need the pullup resistor on RO at this end.

Best Wishes

[nasbyc]

thanks