• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 54 subscribers
  • Views 131 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

CC1120: EMC Susceptibility problem?

Peter CAMP-SMITH
Prodigy 10 points
Part Number: CC1120

Hi Texas Instruments
We are using a C1120 chip on a six-layer RF board that has the RF output feeding a 1 Watt PA Harmonic filter and Tx/Rx switch.
We configure the C1120 via SPI to Tx mode FSK and the chip generated FSK which was originally fed to the PA .
To identify/ isolate the problem the PA has been isolated with its own 3V3 power supply, The PA is currently ran in isolation with its own RF source.
. RF IN and Out soldered directly to the RF board IN/ Out
The C1120 output goes directly to a spectrum analyzer for display.
The PA input come directly from an external signal generator at 10 dBm . The PA output immediately after matching connects to an attenuator and then into a spectrum analyzer. The PA is about 3cm from the C1120. Ample plated through holes. Coplanar 50 ohms tracking. Separate ground planes
Problem sequence.
The PA is power up(RF OFF).
The C1120 micro is power up. The C1120 is set to Tx via SPI. The SPI EXCHANGE IS monitored on an oscilloscope.
The C1120 RF out is monitored on a spectrum analyzer.
As soon as the PA RF is turned on, the C1120 goes into Rx / idle mode ?. Out test frequency is 450MHz.
The C1120 RF output seems to turn off. The C1120 TRX_SW (Pin18) goes from Low to Hi indicating the C1120 has decide to go into Rx. The SPI commands are continued to be send, but the SPI MISO (Master In Slave Out) stops sending data and goes low- just stops.

My hypothesis is internally in the C1120 has a Sensitive RSSI Log amplifier that picks up the PA LEAKAGE RF and the C1120 decided it is going in to receive mode when RF leaks in to the C1120 internal log amp.
It is our probably unreasonably to expect sufficient isolation from a 1 watt PA with volts swing LEAKING in to the C1120 internal, since the chip is not really EMC/ EMC isolated. All the decoupling rules have been followed, buried ground plane with RF on one side and power and slow signal on the other / own side.
Each of the SPI line have a 10MHz filter.

We hope you can offer some discussion / conjecture on our theory that RF is LEAKING into C1120 internal log amplification and this result in a CHANGE of mode to receiver mode.
Let me know for more information is required? Sketch attached. 
Look forward to any suggestions.
Cheers. Peter.

  • 0
    TI__Mastermind 29570 points

    HI Peter,

    Apologies for the delay in replying.

    CC1120 works fine with a 30 dBm Skyworks FEM. We have no reference designs with Qorvo.

    We have had several customers operating at 30 dBm at 915 MHz and 169 MHz with the Skyworks FEMs.

    Regards,

       Richard

  • 0
    Prodigy 10 points


    Hi Richard, thanks for your response.
    We have moved the RF configuration, along to try and learn more about the mechanism.
    Out operating frequency is 450 MHz.

    Referring to the original sketch, the C1120 RF output now connects to the PA, via the recommended matching. The RF output of the C1120 is 13 dBm( I recall..). The PA delivers 29 dBm after a Harmonic filter and a Tx/Rx RF switch in Tx position to a SMA then to a spectrum analyzer.

    Interestingly the FSK is happy working un till we cool the C1120, we don't have to cool it much, and the FSK stops, with the RF stopping on the lower RF frequency. As the C1120 warm up the FSK resumes !
    Initially we were using freezer spray. Then we moved to mylar tape, and a -10 degree Cooled Drill "dry" punch (blunt end) directly on the C1120 chip. The FSK stops.
    On interesting observation is if the PA input is opened circuited - so the PA does not output any RF - the C11120 out puts it ~~ 13 dBm, the FSK works as expected/ reliably, no matter how much we cool it !

    Another observation is if we put a Small chip size Sub Zero ICE BLOCK over the chip through the mylar barrier- it keeps working!
    So we think it is an EMC coupling into the C1120 and a combination of been cold+ metal.
    We do not know what is occurring exactly at this time.
    I even brought a RF HOT wire directly over the C1120, hoping to get the C1120 to mis behave with a Hi RF field through the chip- but no effect!

    Another theory is one of the decoupling caps may have a microscopic crack- but it all looks fine through the microscope. We will be continuing experiment / testing to attempt to understand the mechanism!

    Appreciate any suggestions- but it looks like you may not have experienced this problem.

    Our software engineer is also making changes to the internal configuration. Our software engineer pointed out an internal PPL lock indicator is possibly available, and he may be able to route this signal to an external General Output pin for us to measure observe. He is currently investigating this approach as a diagnostic tool. His suspicion/ hypothesis is the PLL unlocks and the FSK stops..... 

    Out investigations persist...

    Cheers,

    Peter.

  • 0 in reply to Peter CAMP-SMITH
    TI__Mastermind 29570 points

    Hi Peter,

    There could be a large temperature difference when there is a FEM operating at 30 dBm in close proximity and then applying freeze spray.

    If there is a large delta in temperature make sure you issue a new SCAL strobe for frequency synthesizer calibration.

    Regards,

      Richard