• State
  • Locked Locked
  • Replies 4 replies
  • Subscribers 47 subscribers
  • Views 1238 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.

DAC38RF82EVM: Problem using DualDAC mode to generate wideband IQ baseband signal

Saad Qayyum
Prodigy 185 points
Part Number: DAC38RF82EVM
Other Parts Discussed in Thread: DAC38RF82, TIDA-01435

Hello everyone,

I am trying to use the DAC38RF82EVM to generate wideband IQ baseband signal. I am using the settings for the Dual DAC mode as described in the User guide available on the product web-page. the output signal of the DAC is captured using an oscilloscope and the constellation (QPSK) is plotted using Matlab. The constellation is extremely scattered with a best EVM (error vector magnitude) of 9%. I have tried to reduce the symbol rate (from 500 MHz to 50 MHz) but the performance becomes even worse (14% EVM for 50 MHz case).

However, the results are far better if a bandpass signal is generated using only the DAC A and the digital mixer of the DAC-EVM. In this case, I use the same oscilloscope and digitally down-convert the signal in Matlab to plot the constellation (with an EVM of less than 2%). In this case, symbol rates as high as 625M could be demodulated with good performance.

I wonder if the DAC-EVM is even intended to be used as baseband signal generator or there is some settings in the GUI or signal path which I am not correctly using.

Best Regards,

Saad

  • 0
    TI__Mastermind 18065 points

    Saad:

    There is a TI Design in the works that uses the DAC38RF82 as an I/Q source to support very wide bandwidth signals.  The TI Design number is TIDA-01435.  This design is not currently released as of 7-10-17 but will be in a few weeks.  In the mean time I am enclosing a preliminary version for your reference.  This may help you set up the device for baseband I/Q operation.

    --RJH

    High Bandwidth Zero-IF Solution for Microwave Backhaul_v1.7.pdf

  • 0
    TI__Genius 9695 points

    Hi Saad,

    I assume you are using a dual channel Oscilloscope to sample the in-phase data from DAC A and the quadrature phase from DAC B? The Evaluation board did not match the trace lengths at DAC B and DAC A output so there will be a phase error. You can use the Output delay feature of DAC to tune out the phase error between DAC A and DAC B outputs. Also note that the DAC A and DAC B outputs are 180 degress out of phase on the eval board so enable the "invert DAC output" on either DAC A or DAC B.

    I will also suggest using a real CW tone to calibrate the delays between the DAC A and DAC B paths before sending the wideband signal.

    Thanks,

    Eben.

  • 0 in reply to Ebenezer Dwobeng
    Prodigy 185 points

    Thanks Eben,

    I will try your suggestions to improve the quality of the baseband signal.

    One further question about the DAC38RF-EVM: is it possible to change the common mode dc level of the DAC output as the IQ mixer following the DAC output requires a non-zero dc level? I could use a bias-T to externally add a dc shift but since the signal extends down to DC, it deteriorates the signal quality.

    Best Regards,

    Saad

  • 0 in reply to Saad Qayyum
    TI__Genius 9695 points

    Hi Saad,

    The DACs output is biased to 1.8V. If this is compatible with the mixers input common mode, you can remove the caps on the EVM, C329 and C399.

    The balun on the EVM will however not be able to pass low frequency signals through and you may see similar distortion as the external bias-T.

    Thanks,

    Eben.