SoftwareSerial with 1284P vs 644P

Other Hardware 3rd Party Boards
1

To cut to the chase, I have a simple sketch

#include <SoftwareSerial.h>

SoftwareSerial mySerial(21, 20);

void setup()  
{
  Serial.begin(57600);
  Serial.println("Goodnight moon!");

  // set the data rate for the SoftwareSerial port
  mySerial.begin(9600);
  mySerial.println("Hello, world?");
}

void loop() // run over and over
{
  mySerial.println("ping from the board");
  delay(2000);
  
  if (mySerial.available())
    Serial.write(mySerial.read());
  if (Serial.available())
    mySerial.write(Serial.read());
}

that works fine on 644P (both way communication) and it does not work on 1284P (no communication at all).

The "...644P__" and "...1284P__" macros are paired correctly in the sanguino files.

Any ideas?
Thanks.

2

Humous seeing this, I was going to test out this issue, in a way.

My approach is to test out the fuses settings. Default high fuse is 0x9E, but rumor is that 0xDE is better. Tomorrow, I will set the fuses at:
avrdude -v -p m1284p -c usbasp -U hfuse:w:0xDE:m -U lfuse:w:0xFF:m -U efuse:w:0xFD:m

3

These are the values I use for the fuses:

atmega1284.bootloader.low_fuses=0xFF
atmega1284.bootloader.high_fuses=0x98
atmega1284.bootloader.extended_fuses=0xFD

and I never had a problem until now.

4

Are the boards' pinouts different? i.e. the digital pin-to-MCU pin# association?

5

Are the boards' pinouts different? i.e. the digital pin-to-MCU pin# association?

No, I am using the same Sanguino files for both.

6

57k baud has a pretty high error rate, if you are using the internal oscillator, it might be that the 1284p deviates enough to prevent serial from working...

7

My approach is to test out the fuses settings. Default high fuse is 0x9E, but rumor is that 0xDE is better. Tomorrow, I will set the fuses at:
avrdude -v -p m1284p -c usbasp -U hfuse:w:0xDE:m -U lfuse:w:0xFF:m -U efuse:w:0xFD:m

That did not help. Actually it makes matters worse (I tried both 9E and DE), since even the upload fails.

8

57k baud has a pretty high error rate, if you are using the internal oscillator, it might be that the 1284p deviates enough to prevent serial from working...

I am using external crystal (16MHz). The board has been working reliable for quite a while, so hardware-wise is solid.
I am using 57k for the serial monitor, 9,600 for SoftwareSerial.