Because of the holidays, TI E2E™ design support forum responses will be delayed from Dec. 25 through Jan. 2. Thank you for your patience.

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.

DS80PCI402: DS80PCI402 AC coupling capacitors

Part Number: DS80PCI402
Other Parts Discussed in Thread: DS80PCI800

Hello,

I have question regarding AC coupling capacitors on PCIe I'm developing two riser cards with redrivers DS80PCI402 connected by 20cm long FFC cable.

Acording t other thread on this forum, I should decouple all TX sides by 100nF cap. Is that realy necessary? I know, that PC Host and Device have capacitors on their TX sides, so adding another capacitors on TX sides of both DS80PCI402 redrivers seems a bit like overkill and I'm worried about signal integrity. Also, what value of AC coupling capacitors should be used to be compatible with gen2 and gen3? I found values from 100nF to 220nF, what is better? 

Thank you for answer

  • Hi Vaclav,

    I would try to use a single redriver in each path.  This tends to be a more optimal solution.

    PC Host Tx  ->  PCIe Riser Card -> FFC cable -> DS80PCI800 -> Device Rx

    PC Host Rx  <-  DS80PCI800 <- FFC Cable <- PCIe Riser Card -< Device Tx

    Do you know the signal attenuation associated with each of the system boards/cables?

    In terms of the AC coupling capacitors, it is per the specification that each link segment will be AC coupled.  They are required for all PCIe components to interoperate with each other.

    For PCIe the typical value used is 0.22uF, this fits nicely into the PCIe specification.

    Regards,

    Lee

  • Hi Lee,

    thanks for quick reply. I dont know the attenuation right now, but when we used version of riser cards without redrivers, we could only achieve gen1 rates.

    I have  DS0PCI402 configured in a way, that it has 0 EQ and de-emhasis in TX direction and it is equalizing only before RX of device or host, so it could be similar to your proposed solution.

    Regards,

    Vaclav

  • Hi Vaclav,

    While some portion of the reduced performance may be due to attenuation, there is a strong possibility that reflections, crosstalk, or other noise is impacting the signal integrity.  The redrivers only work to compensate for attenuation - they cannot fix other types of deterministic jitter.  I would investigate the waveform quality before building another board with redrivers.

    Regards,

    Lee