• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Subscribers 54 subscribers
  • Views 1507 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

CC1200 pcb layout critical points

Michele Santucci
Intellectual 675 points
Other Parts Discussed in Thread: CC1200

Hallo everyone,

I'm using CC1200 for testing purposes from several months and now is time to design a pcb to fit my needs. The starting design I used is the CC120xEM 420-470MHz Rev 1.1 board.

In my own design I just changed the SMA connector with an U.FL (hirose U.FL-R-SMT-1). I kept the same BOM as the original design. Unfortunately the resulting PCB manifactured and assembled with components is not working: I can 'talk' with the CC1200 (but this was the easy task) but as long as I can see no data is received or transmitted.

Using SmartRF Studio and stacking my own board onto a TrxEB board I made the following tests (using an original TI CC120xEM 420-470MHz board as a counterpart):

1) using PacketTX mode (default configuration) I'm not able to transmit a single packet on the other side;

2) using PacketRX mode (default configuration) I'm not able to receive  a single packet from the other side; 

3) using ContinousRX mode I can clearly see the RSSI signal level grow up of 60/70 dB;

4) while in TX (using a Spectrum Analayzer) I can clearly see the proper signal (centered on 433.9 MHz) level-up to the same values I typically see when using the original CC120xEM boards.

all tests have been made using "Generic 434Mhz -> Symbol rate 1.2Kbps , 2FSK, 12.5 kHz, Channel Spacing (434Mhz)" pre built configuration with units placed about 1-2mt far.

I cannot undestand why data are not 'coming' in (or out).

Excluding design faliures my doubts are about the pcb manifacturing and/or assembly but of course I'm open to any other advice/suggestion.

Supposing an assembly issue my final questions are:

What are critical components into RF signal integrity ?

How can I check if the RF signal is 'correctly' received/transmitted ?

Best Regards,

 Michele Santucci

  • 0
    TI__Guru**** 317180 points
    - Have you routed out the GPIOs on your board?
    - Have you tried to increase the Rx bandwidth? With 12.5 kHz RX BW a small offset between the boards could cause the Tx signal to be outside the filter bandwidth.

    If you use 1 kHz or lower RBW on your spectrum you should be able to see if you send a modulated signal on the expected frequency.
  • 0 in reply to TER
    Intellectual 675 points
    Hallo TER,

    - I have routed all the GPIO signals that are actually routed on 'original' TI CC120xEM 420-470MHz ..
    - no but I will arrange a trial with your suggested approach and reply with results.

    Thanks,
    Mike
  • 0 in reply to TER
    Intellectual 675 points
    Hallo TER,

    we tried using 4.8 OOK RX BW 128 KHz (434Mhz) preset and this worked.
    Starting from such BW we tried again with 1.2kbps 2-FSK 12.5kHz preset changing just
    BW parameter.
    We found that the lowest BW that 'works' is around 48kHz (32kHz is NON working).

    What sensitivity drop we can expect using this BW setting?

    Best Regards,

    Michele Santucci

    bye by{t}e[s]..TuX!
  • 0 in reply to Michele Santucci
    TI__Guru**** 317180 points

    Have you measure the exact frequency the Rx and the Tx DUT transmit on? I suspect that you have a frequency offset on one or both units that are a bit large. Try trimming the xtal load caps to get a more accurate frequency and/ or enter the offset into the FREQOFF register.

  • 0 in reply to TER
    Intellectual 675 points

    Hallo TER,

    I don't know what you mean by DUT frequencies (I'm quite a newbie in RF) but anyhow I cannot measure the exact frequency anymore since the Spectrum Analyzer have been taken off the office and I think will not have it back before a couple of weeks.

    By "try trimming xtal load caps to get a more accurate frequency" you mean trim C302, C301, C311 values to get X1 freq as close as possible to 40MHz?

    Reading more accurately the X1 datasheet (I'm using a Kyocera CX2520DB/CX2520SB family crystal unit) I got that the load capacitor of this crystal should be 8pF/12pF while the board mounts 15pF caps... do you think this can be source the the offset ?

    Reading FREQOFF register description I see that there's a FREQOFF_CFG register that enables feedback to the PLL to "increase the RX filter BW". Is it a good approach to enable this feature or you suggest to 'play' strictly on FREQOFF register?

    Best Regards,

     Michele Santucci

  • 0 in reply to Michele Santucci
    TI__Guru**** 317180 points
    DUT- Device under test, in this case the 2 board you are trying to get link between.

    Without a spectrum it's difficult to tune the RF frequency since you need a accurate reference to do that. But if the xtal you are using is spec'ed with different load caps (the two caps going from Q1 and Q2 to ground) your frequency is probably off (40 MHz * some ppm is easy 10 -20 kHz on 868 MHz)

    FB2PLL is a nice feature and you should look into it, some more info could be found here: processors.wiki.ti.com/.../Category:Sub-1GHz
  • 0 in reply to TER
    Intellectual 675 points

    Hallo TER,

    'DUT' ... thanks! We'll never stop learning! 

    Good news! Even if we miss the spectrum analyzer we discovered that the board has an offset of about 18KHz. Shifting the center frequency of the proper value made all the things works... suprisingly this can be done almost  'automatically' from SmartRF Studio using the Advanced RF features subwindow. Now the minor issue is to clearly identify what caused the offset but I bet the problem originate from the wrong capacitance values used on X1. Today we are going to change such values with correct ones (whatever they are) and check again offset. I will update you with results.

    Best Regards,

     Michele Santucci

  • 0 in reply to Michele Santucci
    TI__Mastermind 39860 points

    If you measure a positive (negative) frequency offset you need a higher (lower) capacitive loading. 

  • 0 in reply to Sverre
    Intellectual 675 points
    Hallo Sverre,

    thanks for your hint we placed a 24pF capacitor and we achieved a 1kHz offset ... everything works now!
    Of course a BIG thanks to TER for his support. We never get off this without you guys!

    Thanks again!