I2C Slave Tx doesn't work, both data and clock pulled low

[nehallove]

I have master PIC18f25k20 and slave 16f1937.

compiler version 4.0139

slave code:
//This is main.h

//#include <16F1825.h>
#include <16LF1937.h>
#device *=16 ICD=TRUE
#device adc=16

//#FUSES LP
#FUSES HS
#FUSES NODEBUG
#FUSES NOWDT //No Watch Dog Timer
#FUSES WDT_SW //No Watch Dog Timer, enabled in Software
#FUSES NOBROWNOUT //No brownout reset
#FUSES NOLVP //No low voltage prgming, B3(PIC16) or B5(PIC18) used for I/O

//#use delay(clock=20000000)
#use delay(internal=16mhz)


//this is main.c

//! Library Files
#include <main.h>

#use I2C(SLAVE,sda=PIN_C4,scl=PIN_C3,address=0xB0,SLOW,FORCE_HW)


unsigned int8 address=0, buffer[16];

#INT_SSP
void ssp_interrupt ()

{
unsigned int8 incoming, state;

state = i2c_isr_state();
if((state == 0)||(state == 0x80))
i2c_read();
if(state >= 0x80)
i2c_write(0xAA);
else if(state>0)
buffer[address++] = i2c_read();

//if((address == 17)||(buffer[address-1] =='\r'))
//address = 0;

/*
if(state <= 0x80) //Master is sending data
{
if(state == 0x80)
incoming =i2c_read(2); //Passing 2 as parameter, causes the function to read the SSPBUF without releasing the clock
else
incoming = i2c_read();

//if(state == 1) //First received byte is address
//address = incoming;
if(state >=2 && state !=0x80) //Received byte is data
buffer[address++] = incoming;
}

if(state >= 0x80) //Master is requesting data
{
i2c_write(buffer[++address]);
}
*/
}

void main ()
{

enable_interrupts(GLOBAL);
enable_interrupts(INT_SSP);

while (TRUE)
{
}
}


Master Code:

try to read from I2C slave

while(1){
i2c_start();
i2c_write(0xB1);
x = i2c_read();
x = i2c_read();
i2c_stop();
}

Result:
if i use only one read operation, it doesn't read right. If i put two read operation then it reads 0xAA. In both condition after one loop cycle, slave pulls both data and clock line low. And both side freezes on #I2C() line.

Does anybody have a clue what is going on? Master write operation works.

nehal
_________________
nehal

[Ttelmah]

1) Learn to use the code buttons.
Your code is 99% unreadable.

2) Trim code. We aren't here to read through swathes of stuff that is remmed out. Test programs should be as short as possible.

3) When doing multiple read transactions, the master needs to not ACK the last byte.

4) Time. Do you really think you can send stop, then immediately start as fast as possible, without pausing at all?.

5) As currently written, address will overflow buffer if you write, and will destroy the contents of the RAM.

6) Have you bothered to check for errata?. The 25K20, on older revisions has a completely non working master mode, and on all revisions has problems handling the stop event, which can be triggered by poor pull-ups, or not allowing any delays..... It is actually very difficult to get reliable I2C with this chip.

[nehallove]

[PCM programmer]

[nehallove]

Thanks PCM Programmer.

By putting i2c_read(0) on the last byte reading stopped the pulling of data and clock bus. But

i2c_start()
x[0] = i2c_read();
x[1] = i2c_read();
x[2] = i2c_read(0);
i2c_stop()

but with above setting first byte x[0] is always 0x62 no matter what i send.
if i use below code the receive byte is also only 0x62

i2c_start()
x[2] = i2c_read(0);
i2c_stop()

Any reason, it seems weird to me.
_________________
nehal

[PCM programmer]

You don't send the slave's read address in either of those i2c transactions.
How can the slave respond ? It has not been addressed.

[nehallove]

oh sorry i accidentally cut the line while copying on my reply i do use the address.

i2c_start()
i2c_write(0xB1);
x[0] = i2c_read();
x[1] = i2c_read();
x[2] = i2c_read(0);
i2c_stop()

and

i2c_start()
i2c_write(0xB1);
x[2] = i2c_read(0);
i2c_stop()
_________________
nehal

[PCM programmer]

Instead of using your programs to test the i2c slave, do this:

Use the CCS example program Ex_Slave.c for the slave PIC:
c:\program files\picc\examples\ex_slave.c

Use the code in this post for the Master PIC:
http://www.ccsinfo.com/forum/viewtopic.php?t=32368&start=3
Note that the master code uses software i2c. It doesn't have FORCE_HW
in the #use i2c() statement. Software i2c master is more reliable than
hardware on many PICs.

What is the Vdd voltage on each PIC ? Is it 5v or 3.3v ? Or are the
PICs running at different Vdd voltages ?

What is the value of the pull-up resistors used on SDA and SCL ?

What is the length of the i2c bus ? (in inches or cm)

Do you have a ground connection between the two PICs ?
(Are they on the same board ?)

[nehallove]

Hi I used the code in ex_slave.c and use that code modify the routine a little bit since it reads the memory. But still the same thing. First byte is wrong on reading. Even on oscilloscope it show 0x62.

PICs running at different Vdd voltages ?
Both pics are at same voltage.

What is the value of the pull-up resistors used on SDA and SCL ?
Pull up is 1.4 k

What is the length of the i2c bus ? (in inches or cm)
An inch

Do you have a ground connection between the two PICs ?
(Are they on the same board ?)

They are different boards but they have common ground and supply voltage.
_________________
nehal

[nehallove]

I also removed the force hardware on master side.
_________________
nehal

[Ttelmah]

I notice you have several conflicts in the fuses on the slave device:
The watchdog is specified to be both software enabled, and disabled. Only one can be used at a time. You also have the oscillator set to use a crystal, but the delay setup to use the internal RC. Things like this can lead to unexpected problems. Also, a slave device does not have a speed. Only master devices should specify this.

Best Wishes

[nehallove]

I changed the fuses as below but still the same.