CC1101 SPI read always 0x0F

The status byte is always returned as the first byte on a read operation of the SPI port.  You must send additional dummy bytes to get the actual data from the register haver the read command and register address have been sent.  Note that the entire sequence is done with a single CS strobe otherwise the SPI state machine will be reset.

Jim Noxon

Hi Jim,

Thanks for the quick response.  I believe that I'm following the sequence you described, unless I'm missing something.  The read function I used is shown below.

char CC1101_SPIReadReg(char addr)
{
   char data;
   output_bit(UHF_ENABLE, 0);    //Enable CC1101
   while(input(SDI));                        //Wait for SDO to go low, indicating that the CC1101 is ready
   spi_write((addr&0x3F)|0x80);   //Send address - bit 6 low and bit 7 high to indicate a read single request
   if(spi_data_is_in())
      data = spi_read(0);                   //Send dummy byte
   output_bit(UHF_ENABLE, 1);    //Disable CC1101
   return data;
}

I have also verified that the read command and address is being sent with a scope.  If I monitor the SDO output from the CC1101 I can see the status byte after I issue the spi_write command, and after I send the dummy byte as well.

Thanks.

Regards,

Dino

So what I see in your code is you toggle the CS line with every byte written.  As explained in section 10 of the CC1101 spec (http://focus.ti.com/lit/ds/symlink/cc1101.pdf), you must leave the CS line low for the entire communication or the communications state machine will be cancelled and must be restarted.  Thus I would suggest the output_bit(UHF_ENABLE, x) statements should be outside of this function so the CS line will remain low for the bytes following the byte with the address and r/w bit etc.  This way, you can call this function multiple times for each byte you wish to transfer and only modify the CS line when appropriate for the type of communication you are trying to accomplish.

Jim Noxon

Thanks for looking at my code. Upon closer inspection I realised that I am not actually sending the dummy byte. As you mentioned, I am only sending one byte before toggling the CS line.  I have updated the code so that I am now sending the dummy byte after the read command before toggling the CS line.  This is in line with some of the TI software examples I have looked at.  I would expect to see the register contents after the dummy write, but I'm still only receiving the status byte (0x0F) in response.  Maybe my understanding is incorrect?

char CC1101_SPIReadReg(char addr)
{
   char status;
   char data;
   output_bit(UHF_ENABLE, 0);   //Enable CC1101
   while(input(SDI));                        //Wait for SDO to go low, indicating that the CC1101 is ready
   spi_write((addr&0x3F)|0x80);   //Send address - bit 6 must be low and bit 7 high to indicate a read single request
   if(spi_data_is_in())                     //Wait until data is ready
      status = spi_read();                 //Read status byte
   spi_write(0x00);                           //Send dummy byte to read register at "addr"
   if(spi_data_is_in())                     //Wait until data is ready
      data = spi_read();                    //Read register data
   output_bit(UHF_ENABLE, 1);   //Disable CC1101
   return data;
}

Regards,

Dino

Make sure you clear your status byte being tested in the spi_data_is_in() call otherwise your second read of the SPI read register will provide the same result as the first.  Also, you should probably be using a while loop to test the spi_data_is_in() call (unless you implement it inside this function) as you want to make sure you wait for the data byte to be completely recieved before reading it.

Other than that, all looks like it should work.

Jim Noxon

This is not a total surprise.  However, it wasn't my first choice due to the belief we were getting at least something back from the radio.  Perhaps it was just a timing thing which caused the data line to flip at the 4th bit.

Managing the ground connection under the radio has proven difficult for prototypes built when access to reflow ovens or proper BGA rework equipment doesn't exist.  Using hotplates, toaster ovens, and hot air guns all require a bit of artistic touch to get good solder connections on a regular basis.  Typically I build 4 to 5 prototypes at a time fully expecting 1 or 2 units to fail.  If I get some to work as expected, then I blame my lack of artistic capability in the assembly process.

Just a note, this can also cause wide variations in range testing so my suggestion is to have at least one small run built on proper assembly equipment for range testing as this will probably be most like what to expect for the final product.

Jim Noxon

sir , i have  implemented the logic of cc1101 register reading but after sending the dummy value im getting the value of the status byte only 

I was hoping someone could point me to where it says that the status byte is returned first in a read operation. I have seen in the datasheet that a status bit is returned concurrently with the read command, but not that it is the first byte returned.

I ask because I am currently experiencing unexplainable behavior from the device, and want to make sure that I have the appropriate literature to understand if I am doing something incorrectly.

Thanks,
Gary

Hi Gary

Page 31 of the data sheet:

"

10.1 Chip Status Byte

When the header byte, data byte, or command

strobe is sent on the SPI interface, the chip

status byte is sent by the CC1101 on the SO pin

"

This is also showed in Figure 15 og page 30.

Siri