CC1101: Fail to send an PLL lock

We always recommend to do a calibration at a certain interval, and not just do one at startup. Doing a calibration every 10 s sounds good.

I sounds like you have some other problems not related to the synth calibration. If you look at Figure 25 in the data sheet, you will see that after calibration state you ill always go to TX or RX state, and you will not go straight back to IDLE even if the calibration would fail.

Is this a problem you observe on old product or is this something you experience on a new product (new HW and/or new SW)? How often does it happen and on how many devices?

How do you know that the transmission does not start? Is it because you do not receive anything, because MARCSTATE indicates that you never enter TX state or something else?

Siri

Please reopen this thread, the reason for lack of feedback is that the feed was sent to the supprt page. Here is the reply I sent a week ago:

he chip with the problem is marked "CC1101 TI 56I ALVL G4". We made calibrations every hour and could see that there was no radio traffic for an hour (next calibration). It happens on a few units, <1% I would guess, but it's hard to say since the units might work during our tests and ends up at customers. When we get them back, they might work as they should. I have checked the FW, it believes that the message is sent as intended since GDO Tx interrupt comes at it should (GDO set to 6, interrupt asserted when SYNC word is sent/received). I have checked both MARCSTATE and the chip status word, these also reacts as if the message is sent. I have also made a test FW that is making 100 calibrations with 1s sleep time inbetween, this chip fails ~5 times and the failures seems to come in bursts; one test was OK for 75s, then there were 3 errors during the last 25s (erros is detected by looking at FSCAL1 that should be !=0x3f if the PLL is locked). Other chips works as expected, My first wish is to correct the problem (workaround) and if that is not possible, find a way to check the chip during production to sort out chips that have this problem. My test above takes 100s, if I go faster, the failure rate goes down, I really dont know if 100s is safe enough.

Siri is on vacation, back 30th. Could this wait until she is back? 

I think have to, I will also soon be on vacation and we should wait for the correct answer

Hi Mats

Sorry for the late reply. If I understand you correctly, the radio does transmit as it should, but it seems like it is transmitting with the PLL not locked so that it transmits on the wrong frequency, is that correct? Also, as I understand it, the units that failes, will also fail when you are running your test FW with 100 calibrations, correct?

Can you please share with me your full set of registers, and also some pseudo code on what you do in your test FW (or send me code snippest of the code).

BR

Siri