• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Answers 1 answer
  • Subscribers 32 subscribers
  • Views 280 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.

DS90UB954-Q1: Coax Mode PCB Layout clarify

alexpioner
Prodigy 140 points
Part Number: DS90UB954-Q1

Hi, all.

We have some problems with FPD channel layout at our card, so I was asked to investigate this issue, and I have some questions.

The FPD channel is operated at COAX mode.

1) Why should the "RIN-" trace be loosely coupled to the "RIN+" trace? Why can't we route the FPD as a 100ohm differential microstrip/stripline between the deserializer and the connector? After all, the RIN of the deserializer is designed to be differential, isn't it?

2) Alternatively, why can't we place the termination resistor of 50ohm close to the deserializer right after 47nF capacitor? Some tutorial says, it should be placed close to the connector?

3) In case we route the FPD at the outer layer, and the next layer is ground, should we moat more layers below the FB1? (The stack-up is L1(sig) - L2(gnd) - L3(sig) - L4(gnd))

4) Is it so critical to route the "RIN+" trace at the opposite side from the connector? The wavelength at 4Gbps (2GHz)  using FR-4 is about 75 mm, while the stub is about 2-3 mm. I understand we want to do the best, but will this stub significantly degrade the signal?

Sorry for my English and many questions.

Best regards

Alex

  • 0
    TI__Expert 7935 points
    Hello-

    1. Yes, the deserializer input is a differential input. However, the signal coming from the TX is a single-ended signal. A single-ended signal needs to see a 50-ohm impedance. The coax cable is 50-ohms, so the RIN+ trace needs to match that impedance as well.

    2. There may be a little benefit of routing the RIN- trace close to connector as it would provide some shielding to the RIN+ trace. Any noise coupled to the RIN+ would also couple to RIN- and the differential RX wouldn't "see" the noise. However, due to space constraints, most customers place the 50-ohm termination right after the ac-coupling capacitors. This is acceptable, too.

    3. Yes. The goal is to have impedance of the PoC filter network as high as possible (or at least >1kohm within the frequency of interest). Minimizing the parasitic capacitance of the FB pads helps as the capacitance reduces the impedance of the PoC filter. It is suggested to moat under all FBs in the PoC filter.

    4. Eliminating the connector stub is one of the most critical recommendations. Connectors, especially thru-hole connectors, generally have the poorest impedance control among other transmission channel elements (cables, PCB traces, etc.).

    Regards,
    Davor
  • 0 in reply to Davor Glisic
    Prodigy 140 points

    Hello, Davor.

    Thanks a lot.