Forums

Hi,

They way one usually handles this kind of issues is to make sure that the preamble is long enough for the RX to "find/hear". If you flush the TX FIFO (so that it's empty) and then enter TX mode, the CC1101 will continuously send preamble, then load the data into the FIFO when you want to send your payload. 

-Jonas

Thanks Jonas for your answer, 

so your suggestion is to send a loooong preamble lasting the time needed to have the RX surely "connected" (in my case about 700 ms) then add the payload? Sounds great!

Just one maybe stupid doubt: does the receiver in WOR mode stay in RX for the entire preamble (in case it "connects" as soon as I start sending it) or it goes to sleep at the end of the RX Timeout?

Thanks 

Miche

Hi again,

Yes, that is my suggestion.

The  RX_TIME bits set the timeout for sync word search in RX for both WOR mode and ordinary RX operation. When using WOR mode, RX_TIME can have a value between 0 and 6 when  WORCTRL.WOR_RES = 0, and between 0 and 3 when  WORCTRL.WOR_RES = 1. RX_TIME = 7 means that the radio will stay in RX mode until a packet is received, regardless  of the WORCTRL.WOR_RES setting. The timeout is relative to tEvent0. The longest RX timeout possible when using WOR  is 1.18 s. This is achieved by settings WORCTRL.WOR_RES[1:0] = 1 and MSCM2.RX_TIME[2:0] = 0.

You'll find all information you need in this Application Note.

Cheers,

Jonas

Hi Jonas thanks again,

one last question..

I know that AN, and  the DN111 but I thought that there wont be any RX Timeout in case of recognised packet (preamble)

I mean: setting WOR_RES=0 and RX_TIME between 0 and 6 the receiver would stay in RX for a certain time UNLESS during that time it gets a preamble, in this case it would stay in RX untill the end of the packet whenever it occurs...is that wrong? 

Or do I need to set RX_TIME=7 to obtain something like that? (timeout relative to tEvent0)?

Thanks

Miche

Hi,

If RX_TIME is between 0 and 6 you will timeout unless you find a SYNC WORD (not preamble, to be exact). If you find a SYNC WORD you will receive that packet.

Preamble is used for bit sync while sync word is used for byte sync of the data.

//Jonas

Hello, 

I'd like to clarify two things that I found out in the last two days for any interested reader...

1- it is possible to send packets back to back setting the TXOFF_MODE in Stay in TX (Start sending preamble);

send the first packet, then:

 - wait for the GDOx to de-assert (in case you use it in mode 0x06)

- load the packet in the TX FIFO and so on...

2- it is possible to avoid the RX timeout (in WOR mode or not) with just a preamble (before the sync word) setting the parameter PQT in PKTCTRL1 and RX_TIME_QUAL in MCSM2.  This way when the receiver gets 4*PQT  bits of preamble, remains in RX till the end of the packet.

This last consideration, together with Jona's suggestion (set the CC1101 in TX with no data in the FIFO: it will send an unlimited preamble) made my WOR work great with a minimum (0,781%) duty cycle!! 

Miche

Glad you solved your problem! :)

-Jonas

Miche, sorry to dig up this old thread but I am trying to get your second suggestion to work and wanted to ask if you could clarify a bit. What do you set RX_TIME_QUAL to? It works for me if I set it to 7, but that also disables the WOR duty cycle and it is on permanently. Any other setting of RX_TIME_QUAL does not work because RX times out before the sync word is sent.

Can you state exactly what setting you used please? Thanks.

Never mind, I fixed it. For anyone else who discovers this thread here is my explanation.

On the receiver set PQT to 1 and set RX_TIME to 1. Enable WOR.

On the transmitter set up a fixed 1 byte packet and send the extended preamble, followed by putting a single byte in the FIFO to generate the sync word.

Like Miche I can get down to a 0.781% duty cycle but no lower, using a 1 second polling cycle, or a 7.81ms on time. This is quite disappointing because with other radios like the RFM12B I can get down to 2ms. With the CC1101 some of that time is spent in idle while the RC oscillator calibrates, but even so it is a bit high. I am going to experiment with higher data rates to see if that helps (I am using 1.2k baud).

Hi Mojo, 

I'm glad you fixed it , I'll share my settings anyway. 

RX_TIME_QUAL = 1

RX_TIME = 4 (100)   with WOR = 0 makes a 0.781% duty cycle

PQT = 2 (010)    then it needs to recognize 8 bit of preamble 

Miche

Hello Miche.

Thanks for your reply. I increased the data rate to 4.8kbps and was able to reduce the duty cycle to 0.391% reliably. At 0.195% wake-up is only triggered about 75% of the time. This is all with PQT=1.