The TI E2E™ design support forums will undergo maintenance from Sept. 28 to Oct. 2. If you need design support during this time, contact your TI representative or open a new support request with our customer support center.

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.

AFE7225EVM: Rationale for filter to TRF370333?

Part Number: AFE7225EVM
Other Parts Discussed in Thread: TRF370333, AFE7225

Can you explain the rationale behind the filter chosen between the AFE7225 and the TRF370333? I've referred to SLA399, section 4.1.3, and it has a substantially different solution. I've also used the SLA399 included spreadsheet (Matched_Impedance, for example) and get different values.  I'm hoping to understand the methodology so I can increase the order of the filter and understand if there are impedance implications when routing the board.

  1. What is the purpose of the 37.4 serial resistors? (R117, R125, R130, R140). 
  2. What is the purpose of the 4.7pF to ground right before IN? (C95, C96, C108, C109) Are these to provide a bypass path to ground at the driven end after the filter?
  3. Doing a little reverse engineering, the 2nd order filter inductor and capacitor values chosen would imply a 50 ohm filter input SE impedance, and a 260 ohm filter output impedance, but R155 doesn't match. Also, those are kind of...unique...values. What is the expected input and output impedance for the filter?

Using SLA399, 4.1.3 (which has a nice explanation of deriving various values). I would expected to eliminate the 37.4 resistors, the 4.7pF to ground capacitors, and change R154/R155 to 200 ohms. A differential 2nd order Butterworth with input impedance of 100 ohm (50 ohm SE) and output of 200 ohm would have inductor values of 215nH and a capacitor of 5.6pF (using rf-tools.com/lc-filter and converting for differential).

Any insight would be much appreciated. Thanks!

  • Hi Mike,

    Can you please pass the document over, SLA399? I don't see that in our databases. So it is difficult to help answer your questions.

    Regards,

    Rob

  • Hi Rob, Sorry about that--it's SLAA399. www.ti.com/.../slaa399.pdf

  • Thanks Mike,

    I am finding the right person that covers the TRF370333. The filter you are asking about would be more related to this product.

    I will get back with you.

    Thanks,

    Rob

  • The AFE7225 can provide up to 2Vpp differential signal output.  That is too high for the modulator input to keep it within its linear range.  The 37.4 ohm resistor in conjunction with the 130 ohm shunt resistor provide a voltage divider signal attenuation to reduce the signal into the modulator.  Those values also provide a 200 ohm differential load to the filter.

    The 4.7pf caps to ground right at the modulator input pins is intended to provide a high frequency bypass/filtering at the baseband inputs.  This is a precautionary approach to ensure high frequency spurs and/or noise do not get into the modulator input.  With a more substantial BB filter, these caps can be eliminated.

    The filter design should be configured for 200 ohm differential input/output impedance per the values on the EVM.  There are many viable options for the output topology.  The approach on the EVM was to drive the DAC at its maximum to achieve best DAC performance metrics and then attenuate down after to keep within the modulator's linear range.

    --RJH

  • This is great information. I wish this rationale were in the datasheet! Thank you!

    Two quick questions on the resistor network:

    • P1db of the TRF is 9dBm. 1V at 100ohm is 10dBm, so we want at least 2dB attenuation per end. So, we want a voltage divider that gives .8V and has a total impedance of 200ohm. This ends up being almost perfectly the 37+37/130 chosen in the EVM. Does this all sound like the correct thought process?
    • Why not eliminate the voltage divider and set the BIASJ to 1.2k on the AFE to give 16 mA max current? This would make the pullup 120, pulldown 555, shunt 200 and eliminate the divider. If my thinking is flawed, please let me know.
  • Actually, you likely want to drive the modulator more backed off from compression to keep the device operating in its linear range.  The voltage divider provides 20 * log ( 0.5*(130) / (37 + 0.5(130))) = ~ 4 dB of attenuation.

    Yes, you can opt to change the DAC load impedances or to change the DAC bias current to reduce the swing from the DAC and then eliminate the voltage divider pad at the output.  That is a viable strategy.  Some of the DAC parameters like SNR are best when driving the DAC full scale.  The EVM design opted to drive the DAC at full scale and then passively attenuate afterward.

    --RJH

  • Awesome feedback, RJ. Great catch--my math was off by 2dB. Also, the pointout of the NSD vs Iouts graphs is spot on.

    Thanks!