• State TI Thinks Resolved
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 46 subscribers
  • Views 76 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

DAC38J84: QMC configuration for qmc-phaseab

Rannie Zhou
Intellectual 1350 points
Part Number: DAC38J84

Tool/software:

Hi team,

My customer is using DAC38J84 QMC function. They configure the register like below. When set the config16(qmc_phaseab) to 0x7ff or 0x800, the QMC don't work, and the IQ waveform is abnormal. Could you please help to elaborate how to configure to QMC and check why the waveform is abnormal?

register configuration:

Config0: Enable bit 12/13

Config16: 

0x000: yellow=I blue=Q 132mV

0x7ff (“011111111111” = 0.49975.) Abnormal: the I signal is distortion.

0x800: (“100000000000” = –0.5.) Abnormal: the I signal is decrease to half.

BRs,

Rannie

  • 0
    TI__Guru 67904 points

    Hi Rannie,

    Could you please advise the end customer's name and application?

    QMC is only 11bit. Only 0x7FF is valid. 0x800 exceeds 11-bit

    Also, the customer is basically multiply the baseband sine and cosine wave by 2x. This is saturating the baseband signal, if the signal is already a full-scale.

     Note that with a full-scale complex baseband signal of cosine and sine wave (i.e. amplitude of unity for both waves), the resulting RF signal is also a full-scale cosine wave that has been upconverted by the digital LO frequency.

    For the case of baseband signal is a DC signal, the resulting RF signal is the following:

    The combination of cosine and sine wave is a trigonometry identity, which results a sine wave with amplitude of √2A with phase shift of -45°. The amplitude exceeds unity, which may cause digital saturation and will need to be attenuated in the digital domain. So the amplitude back-off of the baseband DC signal is needed for the constant data path.

  • 0 in reply to Kang Hsia
    TI__Intellectual 1350 points

    Hi Hsia,

    The EE is ATE tester and the customer is HYC.

    Kang Hsia said:
    QMC is only 11bit. Only 0x7FF is valid. 0x800 exceeds 11-bit

    In the DS, the qmc_phaseab(11:0) is 12bit. so, I think 0x800 should also be valid.

    As in the DS, qmc_phaseab(11:0) controls the phase imbalance between I and Q. Why is the gain of Q changed but not the phase when setting the qmc_phaseab(11:0) in this case?

    The description of saturation in your reply is for amplitude but not explain why the phase imbalance doesn't work. Could you please help to provide how to set up the register to set phase gap between I and Q?

    Best regards,

    Rannie

  • 0 in reply to Rannie Zhou
    TI__Guru 67904 points

    Hi Rannie,

    Please help check if the customer is updating the QMC circuit logic with the proper sync signal (synchronization of the QMC circuit) with the following programming value.

    If the logic is not synchronized or initialized, then it is possible that the written value of phase is not applied to the signal chain.