• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 2 answers
  • Subscribers 11 subscribers
  • Views 442 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.

TRF2443: Unused components, AFE7522 interfacing recommendations, VCO filter params

Mike Neuman
Prodigy 140 points

Part Number: TRF2443

A few questions on several topics:

1) I have no need for the XPIC capability, nor do I wish to add a SAW filter. The XPICOUT+/- and IFOUT+/- pins usually require a pullup inductor. Can I leave these unconnected, or do I need to provide the pullup?

2) What about XPICIN+/-, IFIN+/-, XPICAGC, and XPICBBCM? Can I leave these floating?

3) I'm using the AFE7522 as my DAC/ADC. Do you have any interfacing recommendations? The AF7522 DAC's compliance voltage (2.7V - 3.8V) won't fit the 1.4V common mode on the TRF2443, so I'm planning on AC-coupling with a 1uF cap and a 100 ohm resistor between the TXBB+/-. Is there a better setup or values I should consider?

4) If possible, I'd like to simplify the VCO filters to use fewer parts. 9 unique parts is a lot for two filters. Is there reduced part count recommendation? Are there guidelines on changing the bandwidth of the filter so I could attempt to create reduced part count? (I assume higher bandwidth gets a faster but less accurate VCO lock, are there other considerations?)

Thanks!

  • 0
    TI__Guru 69984 points

    Hi,

    We will need to investigate these questions and get back to do. We do not have part called AFE7522. The closest we have is AFE7225. Is this what you meant? Thanks. 

  • 0 in reply to Kang Hsia
    Prodigy 140 points

    You're correct--I meant AFE7225. Thanks for looking into it!

  • 0
    TI__Mastermind 18095 points

    Mike:

    The XPIC circuitry is not required.  You can power down that circuitry through SPI.  You can leave these XPIC_Out pins unconnected if not used.  All other XPIC related pins can be floating if the circuitry is powered down.

    The IF path gives you the option to switch to an external SAW.  Even if you use the internal switch, you still need to provide the biasing through the pull-up inductor on IFOUT.  The IFIN pins can be left floating.

    The VCO loop filters have been optimized for best phase noise.  Generally I do not recommend to modify them.  You could reduce the filter order and maintain the 100 kHz loop bandwidth if reducing part count is critical.  I expect that this would degrade integrated phase noise slightly.

    The AFE7225 DAC requires a bias current sink.  I think the best approach is to use a resistor network to provide the proper bias for the AFE and translate the voltage to the common mode of the TRF2443.  Check out this app note: Passive Terminations for Current Output DACs.  The case of a 3.3V current sink DAC plus the ~1.5V modulator interface are most similar to your case.

    --RJH

  • +1
    Prodigy 140 points

    Hi RJ,

    Thanks for your response!

    A few questions:

    1. I cannot replicate the VCO loop filter values from the EVM or datasheet using either public PLL calculators or the TI Loop Filter Calculator. Can you point me in the right direction?

      Using the TI Loop Filter Calculator (1.3)...

      For the TX VCO loop filter (2720MHz, 100KHz BW):
      BW: 100KHz
      Phase Margin: 45 Degrees
      Fref: 20MHz
      Fout: 2720MHz
      R: 8
      Kvco: 45 MHz/V
      Icp: 1mA

      The resulting calculations aren't really even close to the EVM/datasheet.
    2. Does the IFOUT pullup require both the inductor and the capacitor? Is the 82nH value significant, or can it be replaced with the same 68nH that's used from the MIXER load?

    3. I'm planning on connecting the TXBB path through a transformer and using the internal bias of the TRF2443. Aside from the 100ohm resistor between TXBB[iQ]+ and -, are there any other components necessary (e.g. DC blocking capacitors or series resistors to limit current flow between + to -)? 

    Thanks again!

  • +1 in reply to Mike Neuman
    TI__Mastermind 18095 points

    Please see below.  --RJH

    1. I cannot replicate the VCO loop filter values from the EVM or datasheet using either public PLL calculators or the TI Loop Filter Calculator. Can you point me in the right direction?

    RJH>> Understood.  The values in the datasheet were derived empirically, which is why I recommend to keep the same values as it is a proven solution.  Other solutions and/or topologies derived from loop filter tools may work equally as well.

    1. Does the IFOUT pullup require both the inductor and the capacitor? Is the 82nH value significant, or can it be replaced with the same 68nH that's used from the MIXER load?

    RJH>> Yes, the 0.01 uF capacitor is needed for supply bypassing.  The 82 nH was chosen as a suitable RF choke at the IF frequency of operation.  Modifying this value is acceptable but you may see a slight reduction in "choke effectiveness"  with a lower value.

    1. I'm planning on connecting the TXBB path through a transformer and using the internal bias of the TRF2443. Aside from the 100ohm resistor between TXBB[iQ]+ and -, are there any other components necessary (e.g. DC blocking capacitors or series resistors to limit current flow between + to -)?

    RJH>> Yes, you will need AC blocking caps on the differential paths.  You may also need an AC blocking cap on the single-ended side too. No series resistors needed.

  • 0 in reply to RJ Hopper
    Prodigy 140 points

    On #2, I actually meant the 15pF caps in parallel with the 82nH's, not the decoupling caps.

    On #3, I'm actually doing balanced->balanced through a transformer from the DAC to the radio, and I don't think the AC blocking caps would be necessary there, but please correct me if I'm wrong. If there are other pitfalls with the balanced->balance configuration, I'm happy to hear about those now too. :-)

    Thanks for all of your help!

  • 0 in reply to Mike Neuman
    TI__Guru 69984 points

    Hi,

    For #3, you may need to review the transformer/balun topology. We have seen cases where the two primaries are actually shorted due to the architecture of balun design. Also, applying DC directly to the balun will impact the magnetic as there may be DC currents and shifting the EM field.

    Just some general thoughts about this balun/transformer discussion.