• State Resolved
  • Locked Locked
  • Replies 11 replies
  • Answers 2 answers
  • Subscribers 54 subscribers
  • Views 271 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.

CC1101: Common Shared Crystal with MSP430

Chris Yorkey
Genius 15205 points
Part Number: CC1101
Other Parts Discussed in Thread: MSP430F5359, CC1190

TI sub-1GHz Team,

Current by GE Lighting is trying to develop a common RF module platform with our CC1101 and MSP430 (F5359) using a common, shared 26MHz crystal if that is possible/recommended.

They desire to do so to save cost and BOM space.  They have reviewed our DSs extensively and are very interested in doing this.

In doing so, they are seeking guidance as to which path to follow:

1) Let the CC1101 drive the crystal (and act as "Host" so to speak)

2) Let the MSP430 drive the crystal (and act as "Host" so to speak)

3) Use the MSP430 Clock Output feature @ 26MHz output with appropriate harmonic filtering (less likely)

Do we have a TI recommendation per above?  GE is very interested.

Let me know if you need further information.  Looking forward to input please.

TY,
CY

  • +1
    TI__Mastermind 31620 points

    Hi,

    Recommend to use a 26 MHz crystal for the MSP430, then output the reference crystal signal on a GPIO pin and feed this to the CC1101 XOSC_Q1 input. XOSC_Q2 line must be left un-connected.

    Regards,

       Richard

  • 0 in reply to RGW
    TI__Genius 15205 points

    Thank you Richard, much obliged!

    -Chris

  • 0 in reply to Chris Yorkey
    TI__Genius 15205 points

    Richard,

    GE just asked if we are sure that the 3V signal won’t overdrive the xtal input, and cause RF performance issues ?  I'm not that concerned myself, but let me know your thought.

    As a precaution, they would probably want to add a small LCR harmonic filter to reduce the p-p voltage and soften the edges while removing the higher harmonics.

    Finally, do you know what is the optimum input wave-shape and p-p voltage on the CC1101 Xtal input ? They can use that to tweak the filter to match.

    TY,
    Chris

  • 0 in reply to Chris Yorkey
    TI__Mastermind 31620 points

    Hi Chris,

    The reference clock signal to the CC1101 can either be a full swing digital signal (0 V to VDD) or a sine wave of maximum 1 V peak-peak amplitude.

    It's important to note that the crystal to the CC1101 is the main reference for the RF so this signal has to be a clean 26 MHz signal without any spurs, glitches or any other harmonic content otherwise the disturbances will be reproduced and mirrored to the RF section of the radio.

    If the output from the MSP430 GPIO is not a clean reproduction of the buffered crystal signal, then this will not work. Please verify that the output from your chosen MSP430x is adequate enough on an evaluation board to be a clean reference signal for the radio prior to making a board spin. Recommend to connect the buffered clock reference from the MSP430 evaluation board to the CC1101 EM XOSC_Q1 input together. Then generate a static unmodulated Tx carrier on the radio. Any possible disturbances on the clock reference will be seen in the RF spectrum.

    If any noise is on the reference signal, then it is better to use two crystals instead.

    Regards,

       Richard

  • 0 in reply to RGW
    TI__Guru**** 317180 points

    Is it possible to clock the MSP from the CC1101 using a GDO (from Table 41):

  • 0 in reply to TER
    TI__Mastermind 31620 points

    Hi Chris,

    Please try the method provided from TER as a first option since this will introduce less noise in the radio section providing that the MSP can be clocked in this manner.

    2nd option: configuration mentioned previously in thread

    3rd option: two crystals.

    Let us know the outcome.

    Thanks.

  • 0 in reply to RGW
    TI__Genius 15205 points

    Good info thanks again Gentlemen!

    From customer directly, unfortunately, using GDO to drive the 26MHz clock out to the MCU is also problematic, as the 2 pins are already used for other purposes so that rules that out at least for now.  ;-(

    And the warning on GDO output clock config: ” To optimize RF performance, these signals should not be used while the radio is in RX or TX mode.”

     We agree it would be best to let the CC1101 have its crystal and drive the MSP430 from it.

     Can you elaborate what is the *minimum* sine wave  p-p voltage that works reliably with the CC1101?  

    And the same for the MSP430F5359’s high-frequency oscillator/clock input?

    We suspect there should also be a sweetspot p-p voltage that minimizes both jitter and harmonics. It would be good to know what that is.

    We are also thinking that taking the crystal driver output from the CC1101 and either just capacitively coupling it directly into the oscillator input pin of the MSP430 (to minimize the extra loading on the crystal driver) would work best.

    [Or worst case use a small single gate logic buffer or low-gate charge small signal mosfet inverter to buffer that to the MSP430 would work out best with least risk.]

     Lastly, we might design the first prototype to allow clock propagation in either direction to do some testing and check out the impact on signal purity on the spectrum analyzer as suggested in one of the appnotes. But we would be more concerned about effect on sensitivity (harder to test) than output harmonics as the design already has a high performance saw filter between the cc1101 and CC1190.

    Comments indeed welcomed.

    TY,

    Chris

  • 0 in reply to Chris Yorkey
    TI__Mastermind 31620 points

    Hi Chris,

    I would keep the minimum to 0.7 V pk-pk for the CC1101 clock reference voltage and maximum to 1.0 V pk-pk.

    I do not have this info for the MSP430, so please ask the MSP430 support team.

    Thanks,

       Richard

  • 0 in reply to RGW
    TI__Genius 15205 points

    Much obliged, Richard!  Will do.  Thanks, Chris.

  • 0 in reply to Chris Yorkey
    TI__Guru 56685 points

    Chris, 

    From an MSP perspective, you would not be able to output a 26MHz clock  from the MSP430F5359. This device is limited to 20MHz operation. You can certainly input a higher frequency clock that s divided down to a max of 20MHz, but any output would be limited to this max frequency. 

    For the input to the MSP430, the XT2 is meant to be used with Xtal, so its an active drive circuit to do this. You can utilize this input in bypass mode, but you need a digital clock input in that mode. 

    Another option if the precision timing is not needed for the MSP430 device, would be just to utilize the internal DCO. They could use a LF Xtal for FLL reference, or the internal REFO at the cost of additional current. 

  • 0 in reply to Jace H
    TI__Genius 15205 points

    Thank you Jace !