• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 20 subscribers
  • Views 188 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.

LMX2572LP: RFout enable without SPI programming

Andrew Loblaw
Prodigy 50 points
Part Number: LMX2572LP
Other Parts Discussed in Thread: LMX2594,

I would like to configure the RFout to be transmitting for short intervals of time, ideally configured by an external input (e.g. SYNC pin?) or configured via register settings.

The sequence is as such:

1. Calibrate for frequency 1

2. Send for a pre-defined interval of time (e.g. 500ns)

3. Stop transmitting

4. Calibrate for frequency 2

5. Send for pre-defined interval of time

6. Stop transmitting

7. etc.

Is this possible, or will I basically require SPI communications at each step? 

  • +1
    TI__Mastermind 29010 points

    Hi Andrew,

    Toggling the CE pin would theoretically put the device in the lowest power state, but I'm guessing that the internal LDO startup time on power-up would be prohibitively expensive (millisecond range). Consequently, the only quick ways to handle stop/start transmission would be through SPI programming.

    As for frequency hopping, the only practical way to support this requirement is by SPI programming or through I2S, since the VCO frequency must change and therefore the N-divider and potentially the channel divider must change as well. If you are going to a limited number of frequencies, the FSK feature may be usable. On the other hand, if you have a multitude of frequencies, the fastest frequency hopping will involve building a lookup table and using full-assist, but this may not be feasible on smaller or simpler systems. Partial assist calibration is next fastest, and although the VCO would be re-calibrated each time you frequency hop, the calibration time could be minimal.

    To stop transmitting, my recommendation is generally to disable the output buffer with OUTx_PD bit, since this is fast (100s of ns) and does not disturb the VCO, PLL, or other critical elements required to ensure lock.

    Regards,

    Derek Payne

  • 0 in reply to Derek Payne
    Prodigy 50 points

    Hi Derek,

    Thanks so much, this detailed response is incredibly helpful. So it sounds like SPI programming with OUTx_PD bit configuration is the way to go to achieve the time gating we need (~100's of ns resolution assuming we can hit the SPI clock frequencies required)


    My follow-up question is: how quick can we realistically expect to calibrate the VCO between each frequency step? I imagine the calibration time depends on whether an assist feature is used, and the frequency step width. I'm not familiar enough with the assist features, before exploring them in more detail, it would be good to understand which option is required based on my requirements:

    <20us calibration time (output buffer will be off)

    ~1-2us transmitting (output buffer will be on)

    5MHz frequency steps

    Any insight here is very much appreciated. Thank you in advance!

  • +1 in reply to Andrew Loblaw
    TI__Mastermind 29010 points

    Hi Andrew,

    It sounds like you have a large frequency range over which you will be generating your 5MHz frequency steps at the output, so I think this would not work well with FSK feature where the step size is usually expected to be small enough to fit within the variation of the N-divider fractional numerator without changing the integer N-divider component. Selecting a new N-divider value and recalibrating is probably the only way to achieve your full frequency range.

    You should be able to achieve <20µs calibration time with partial assist or full assist calibrations. Please take a look at the app note linked below for a primer on using calibration assistance features. While the app note is written with reference to LMX2594, the same set of features is implemented in LMX2572LP, so the same principles will apply and the timings achieved will be comparable. Figure 38 and Figure 39 in particular suggest about 18µs for frequency and amplitude calibration with partial assist - full assist would be even faster, but would require per-part calibration tables.

    https://www.ti.com/lit/an/snaa336/snaa336.pdf

    Also worth pointing out: the OUT_MUTE bit can be used to ensure that the outputs are disabled while the VCO is calibrating.

    Regards,

    Derek Payne