• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 22 subscribers
  • Views 916 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.

LMX2571: SPI-FAST FM Modulation

ahmet donmez
Prodigy 10 points
Part Number: LMX2571

Dear Support Team;

We use the LMX2571 to do FM modulation. The modulated signal is composed of either a single sinusoidal signal or a composite signal that has a few sinusoidal components.

We demodulate the signal using universal demodulator after analog FM modulation using SPI-FAST. We are sure that the demodulator works perfectly (We have tried various FM demodulation setups including sofware defined radio – GNU Radio to be specific.). But when we measure the phase of the demodulated signal, we observe that the PHASE of the demodulated signal is not constant and changes over time. So although we expect constant and stable phase on the demodulated signal, we keep loosing the phase information.

Our test setup information is as follows:

Our test setup consists of STM32G431KB for the SPI interface and LMX2571EVM for FM modulation. First, the modulated signal is oversampled(80kHz) and then sampling information is written to flash of the STM32G431KB. Then the samples of the composite signal is written to R33 register of the LMX2571 for the FM modulation. The maximum deviation is about +/-20 kHz and the cycle rate is between 100Hz to 15kHz. (DEN=5394304, CH1=4, Ch2=8, fPd=120MHz, prescaler=4, Loop BW=204.8kHz)

Is this to be expected for LMX2571? Can we obtain FM modulation with no or negligible phase corruption?
What might be the possible cause of the problem?

  • 0
    TI__Mastermind 33430 points

    Hello,

    Since the loop bandwidth is very wide, is it possible that the phase of the reference input to the LMX2571 is changing? In other words, is the close-in phase noise of the reference input low enough for the application?

    Does the phase corruption occur across all frequencies in the FM cycle? In other words, if you hold the FM at a specific frequency (e.g. 100Hz or 15kHz) without varying the cycle rate, is there still phase corruption over time?

    Regards,

    Derek Payne

  • 0 in reply to Derek Payne
    Prodigy 10 points

    We have used 20MHz square wave for both the clock of the STM32 and the reference input for the LMX2571. We have repeated the tests using a very clean 20MHz oscillator and obtained the same results.

    Also we have verified that the SPI signal is generated properly. (First we converted SPI signal to analog signal using DAC. Then we sampled the analog signal using ADC and took the complex FFT and verified that the phase of the SPI signal is constant)

    When the modulating signal is composed of five tone sinusoidal signal, after demodulation the phase changes over time are as follows: (linearly proportional to frequency)
    for 2.3kHz, the phase change is 650 degrees/second
    for 7.5 kHz, the phase change is 2150 degrees/second
    for 12.7 kHz, the phase change is 3650 degrees/second
    for 15 kHz, the phase change is 4300 degrees/second
    for 17.3 kHz, the phase change is 4950 degrees/second.
    We wonder if changing the N-Divider value of the PLL effects the phase of the modulated signal?

  • 0 in reply to ahmet donmez
    TI__Guru** 107030 points

    Hi Ahmet,

    When we verified the FSK function, we have never verify the phase performance, our focus so far was on the FM performance such as FSK error, eye diagram, etc. So I don't have any information on phase behavior.

    BTW, when using the FSK function, DEN should be set to 2^24 to allow +/- frequency deviation. Although using a different DEN may work, but there are cases that only + frequency deviation is achievable.