This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

CC1125: CC1125 Failure /Damage

Part Number: CC1125

Hello,

I'm looking at the CC1125 datasheet SWRS120E from June 2011 revised october 2014. In page 7 chapter 4.1 Absolute Maximum Ratings it specifies the Voltage on analog pins to be max 2.0V. 

Does that also applies to the pin 17 where the PA  is feeded by VDD?

We are experienced a high failure rate (50%) of the CC1125 in our design and we think we are doing something that seems to be damaging the transceiver. We are excluding ESD, Solder reflow temperature and we have also checked for voltage pikes coming from the mother board. Our production engineer has pointed out this 2.0V max on analog pins and the voltage we are now aplying VDD to pin 17 wich is above 3V.

The CC1125 continues to communicate to the uprocessor but does not seem to lock to the intended frequency of 460Mhz, it seems to be transmiting at few MHz lower, unlocked and with plenty of spurious.  Could there be an explanation for this?

Thanks

  • The PA pin has the same rating as the VDD pins (up to 3.6 V).

    In the reference design we have a small resistor to VDD in the network between VDD and the PA pin to lower the bias point on the PA pin slightly. This is done to increase reliability on max output power. Do you follow the reference design on this?

    We have not seen any issues with the this pin.

    To rule out software issues, do you see the same if you use SmartRF Studio (you would need a TRXEB or a CCDebugger).

    I can do a quick review of the schematic and settings if you like. Send me a friend request if you want to send the data private.
  • Hi,

    PA pin (Pin-17) can be tied to VDD through a 18 Ohm Resistor as shown in our Reference Design (for 400MHz band design - SWRR101) .

    Did you follow our Reference Design (SWRR101) in doing your Schematic and PAB Layout?

    Did you check the Crystal Oscillator Frequency (40MHz) for Frequency Accuracy?

    Could you please post your Schematic in PDF for our review?

    Thanks,
    PM
  • Hello,

    I didn't use a resistor but only a 56nH inductor with two capacitors for decoupling and DC isolate the output. The fact is that we use the CC1125 as a receiver and we only use the transmitter once in its life for frequency calibration so we made it as simple as possible.

    We are now replacing this coil with a resistor to lower the voltage and see if it helps but in any case it never goes over 3.6V.

    The software seems to be ok and we compared damaged/undamaged CC1125 registers and I'm being told they are the same.

    I'll check with management about the schematic review but apart from the Tx line the rest is similar to the reference design.
  • Hello,

    As I said to your colleague above we didn't follow the reference design on that pin as we only use the transmitter once but we are now using a resistor to see if it helps.

    I checked the TCXO and it's oscillating nicely at 40MHz.

    Again I'll check if I can send our schematic but apart form that pin it is the same as for the reference design.
  • You don't need to use an resistor if you keep the output power below about 10 dBm.

    If you only use CC1125 in receive, did you see the issue when you do the frequency calibration?

    Could you post a screenshot of the spectrum you get when you see this error.
  • Hello,

    Yes we saw this problem while doing frequency calibration, in the picture above we should have a single carrier at 458.5125MHz  but as you can see thats what we got instead. It seems it is not locked and with lots of spurious.

  • Agree, it looks like the synth is out of lock.

    - If you can't send the schematic, check for assembly errors. The PCBs that work, does the spectrum look normal?
    - Could you give which settings you are using?
    - Could you confirm that you do a SCAL before going to TX (or having auto calibration on)
  • Thanks a lot TER. We actually had the CC1125 on manual calibration, we changed that to autocal and it has solved our problems. We will continue assessing our product and confirm it but it does looks like it is working now. Thanks a lot that was a great help.

    We had the register SETTLING_CFG at 0x03 and changed it to 0x0B.
  • Hello,

    We are implementing a new software that allows the autocal ON (SETTLING_CFG to 0x0B) which seems to be solving the problem.

    My question now is that a large number of units have been working until now and have been coming out of the factory without the autocal ON. Those units have been manually calibrated and have been fitted with the recomended TCXO. The receiver is constantly ON once it's powered up through its life so there are no more calibrations.

    What are the dangers we may encounter? What could be the biggest problem? Could we fall out of Lock due to TCXO aging, temperature & voltage drifts? 

  • If you do calibration once this calibration should be valid forever in theory if you don't change temperature or supply voltage. But:

    I would at least have some fail safe implemented. Either schedule a calibration at a given interval or if you haven't received a packet (within a time frame where it's expected to receive one), force a calibration.
  • Hello,
    Sorry when I say we manually calibrated the synthesiser it is a frequency calibration we do to compensate for TCXO error REG FREQOFF_CFG.
    We have not done the autocal but we may not have done the manual calibration of the Frequency Synthesizer either, we just programed the CC1125 from the SmartRF Studio 7. Is there a possibility in which there may be three states? Autocal, manual cal and not cal at all?
    In that case would the devices that passed the Tx/Rx test be prone to failure in the future?
  • As far as I understand you used SETTLING_CF = 0x03 originally. To confirm: I get the impression that in this case you haven't done a SCAL somewhere in the code? If that is the case I don't understand how your product actually can work since the synth has to be calibrated at least once.