On Jan 6, 10:49=A0pm, wzab <w...@wzab.nasz.dom> wrote:
> static inline uint8_t SPI_Transfer(uint8_t c)
> {
> =A0 const uint8_t status_mask=3D((1<<RXC0) | (1<<TXC0));
> =A0 /* Wait for empty transmit buffer */
> =A0 while ( !( UCSR0A & (1<<UDRE0)) );
> =A0 /* Put data into buffer, sends the data */
> =A0 UDR0 =3D c;
> =A0 /* Wait for data to be received and transmission completed
> =A0 =A0 =A0Please note, that it is not enough to wait for availability
> =A0 =A0 =A0of received data (RXC0), if you set SS high right after
> =A0 =A0 =A0SS is set, you'll experience SPI frame error in the RFM70!
> =A0 =A0 =A0I've lost quite a long time trying to debug this problem!!!
> =A0 */
> =A0 while ( (UCSR0A & status_mask) !=3D status_mask );
> =A0 /* Get and return received data from buffer */
> =A0 return UDR0;
>
> }
Ooops, it seems, that the TXC0 flag is not so easy to use. It is
cleared only after interrupt or after writing "1" to it (in polling
mode).
So the above function must be slightly modified!
static inline uint8_t USART_SPI_Transfer(uint8_t c)
{
const uint8_t status_mask=3D((1<<RXC0) | (1<<TXC0));
/* Wait for empty transmit buffer */
while ( !( UCSR0A & (1<<UDRE0)) ){};
/* Put data into buffer, sends the data */
cli();
UDR0 =3D c;
UCSR0A =3D (1<<TXC0); //Clear the completed transfer flag (possibly
set by previous transfers)!
sei();
/* Wait for data to be received and transmission completed
Please note, that it is not enough to wait for availability
of received data (RXC0), if you set SS high right after
SS is set, you'll experience SPI frame error in the RFM70!
I've lost quite a long time trying to debug this problem!!!
*/
while ( (UCSR0A & status_mask) !=3D status_mask ){};
/* Get and return received data from buffer */
return UDR0;
}
Reply by wzab●January 7, 20112011-01-07
On Jan 6, 11:31=A0pm, d...@kbrx.com wrote:
> If the compiler will generate a listing in assembley language, I'd check
> the "status_mask" first thing.
>
> Hul
avr-gcc treats it just as it was a constant defined by #define.
The generated code is quite efficient:
while ( (UCSR0A & status_mask) !=3D status_mask );
f6: 80 91 c0 00 lds r24, 0x00C0
fa: 80 7c andi r24, 0xC0 ; 192
fc: 80 3c cpi r24, 0xC0 ; 192
fe: d9 f7 brne .-10 ; 0xf6 <SPI_Transfer
+0xe>
/* Get and return received data from buffer */
--
Regards, Wojtek
Reply by wzab●January 7, 20112011-01-07
On Jan 6, 11:37=A0pm, hamilton <hamil...@nothere.com> wrote:
> > yet)")
>
> Thanks for the code, I'll take a look.
>
> Can you tell me what the 'X' on the front of so many lines is for ??
>
This is a "shar" archive (according to the alt.sources convention).
You can save the whole post (in an original form) and run "unshar" on
it.
Or you can remove all lines before the "#!/binsh" line, and run the
remainder
as the bash script. The above applies to Linux or other Un*x-like
systems.
I don't know how to unpack it in Windows (unless you make it Un*x like
by installing cygwin http://www.cygwin.com/ )
--
Regards,
Wojtek
Reply by hamilton●January 6, 20112011-01-06
On 1/6/2011 2:49 PM, wzab wrote:
> In different groups I have seen messages from people
> trying to get RFM70 to work with the AVR.
> I've faced the similar problem, when I needed
> to connect RFM70 to the USART working in the SPI mode.
> The problem which disturbed operation of RFM70
> was associated with incorrect SPI frames generated by
> ATmega88, when I set SS (CSN in RFM70) line high right
> after the RXC0 flag was set.
> To obtain the correct operation, I had to modify the
> standard SPI transfer routine in the following way:
>
> static inline uint8_t SPI_Transfer(uint8_t c)
> {
> const uint8_t status_mask=((1<<RXC0) | (1<<TXC0));
> /* Wait for empty transmit buffer */
> while ( !( UCSR0A& (1<<UDRE0)) );
> /* Put data into buffer, sends the data */
> UDR0 = c;
> /* Wait for data to be received and transmission completed
> Please note, that it is not enough to wait for availability
> of received data (RXC0), if you set SS high right after
> SS is set, you'll experience SPI frame error in the RFM70!
> I've lost quite a long time trying to debug this problem!!!
> */
> while ( (UCSR0A& status_mask) != status_mask );
> /* Get and return received data from buffer */
> return UDR0;
> }
>
> My whole working code is available in the alt.sources usenet group:
> http://groups.google.com/group/alt.sources/browse_thread/thread/19ce3c8b83458def
> (or search for "RFM70 - working example code for ATmega88 (no interrupts
> yet)")
Thanks for the code, I'll take a look.
Can you tell me what the 'X' on the front of so many lines is for ??
Thanks
hamilton
Reply by ●January 6, 20112011-01-06
If the compiler will generate a listing in assembley language, I'd check
the "status_mask" first thing.
Hul
wzab <wzab@wzab.nasz.dom> wrote:
> In different groups I have seen messages from people
> trying to get RFM70 to work with the AVR.
> I've faced the similar problem, when I needed
> to connect RFM70 to the USART working in the SPI mode.
> The problem which disturbed operation of RFM70
> was associated with incorrect SPI frames generated by
> ATmega88, when I set SS (CSN in RFM70) line high right
> after the RXC0 flag was set.
> To obtain the correct operation, I had to modify the
> standard SPI transfer routine in the following way:
> static inline uint8_t SPI_Transfer(uint8_t c)
> {
> const uint8_t status_mask=((1<<RXC0) | (1<<TXC0));
> /* Wait for empty transmit buffer */
> while ( !( UCSR0A & (1<<UDRE0)) );
> /* Put data into buffer, sends the data */
> UDR0 = c;
> /* Wait for data to be received and transmission completed
> Please note, that it is not enough to wait for availability
> of received data (RXC0), if you set SS high right after
> SS is set, you'll experience SPI frame error in the RFM70!
> I've lost quite a long time trying to debug this problem!!!
> */
> while ( (UCSR0A & status_mask) != status_mask );
> /* Get and return received data from buffer */
> return UDR0;
> }
In different groups I have seen messages from people
trying to get RFM70 to work with the AVR.
I've faced the similar problem, when I needed
to connect RFM70 to the USART working in the SPI mode.
The problem which disturbed operation of RFM70
was associated with incorrect SPI frames generated by
ATmega88, when I set SS (CSN in RFM70) line high right
after the RXC0 flag was set.
To obtain the correct operation, I had to modify the
standard SPI transfer routine in the following way:
static inline uint8_t SPI_Transfer(uint8_t c)
{
const uint8_t status_mask=((1<<RXC0) | (1<<TXC0));
/* Wait for empty transmit buffer */
while ( !( UCSR0A & (1<<UDRE0)) );
/* Put data into buffer, sends the data */
UDR0 = c;
/* Wait for data to be received and transmission completed
Please note, that it is not enough to wait for availability
of received data (RXC0), if you set SS high right after
SS is set, you'll experience SPI frame error in the RFM70!
I've lost quite a long time trying to debug this problem!!!
*/
while ( (UCSR0A & status_mask) != status_mask );
/* Get and return received data from buffer */
return UDR0;
}
My whole working code is available in the alt.sources usenet group:
http://groups.google.com/group/alt.sources/browse_thread/thread/19ce3c8b83458def
(or search for "RFM70 - working example code for ATmega88 (no interrupts
yet)")
--
HTH & Regards,
Wojciech Zabolotny