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.

DS280MB810: De-mux and transparency features

Part Number: DS280MB810
Other Parts Discussed in Thread: DS560MB410

Tool/software:

Hello,

I consider using the DS280MB810 as a de-mux device, in a 100GBase-KR interface design.

The de-mux is required in a midplane scenario, so the link between one side can be switched between 2 selectable

paths (similar to a redundancy scheme).

The most critical feature is transparency so signal characteristics through DS280MB810 is preserved for each of the

de-muxed selected options  – so link training is performed between both link sides, whatever de-mux option is selected.

 

A total bypass would seem to be required to enable the mentioned transparency, however according to datasheet,

It seems that the "bypass" option only bypasses boost stage 1, while the signal is driven via boost stage 2.

Datasheet states that "…EQ boost stage 2 controls. Directly goes to analog. No override bit is needed…"  but it is not clear

why the bypass doesn't include Boost stage 2 as well.

 

There seems to be also DS560MB410 device as option, would that be more suitable for the described application?            

Best Regards,

Arie Menscher

  • Hi Arie,

    Either DS280MB810 or DS560MB410 would be suitable for this application.  These are both linear devices, which enable them to be transparent to the link training process.

    Thanks,

    Drew

  • Hi,

    Thanks for the feedback.

    Can you please advise regarding the "bypass" option in the DS280MB810, because this would be the required function to apply for the transparent transfer of link training and other related actions  -- since the datasheet indicates that the "bypass" option only disables boost stage 1, while the signal is driven via boost stage 2 - why does bypass option not include boost stage 2, for that purpose?.

    I see the indication in datasheet that "…EQ boost stage 2 controls. Directly goes to analog. No override bit is needed…"  but still not clear why and what the impact could be?!

    Best Regards,

    Arie

  • Hi Arie,

    I think it's important to distinguish between the "requirement for transparency during link training" and a "bypass" option.  It is possible for the signal to travel through the CTLE boost stages and still enable transparency for link training.

    The purpose of the bypass option is to provide a method of lowering the CTLE boost.  The lowest boost option using bypass is 2.4 dB @ 12.9 GHz.  The lowest boost option without bypass is 8.7 dB @ 12.9 GHz.

    Even when the signal travels through CTLE boost stages, it is still transparent for link training since the device is a linear device.  A non-linear device such as a limiting redriver or retimer would not be suitable for link training applications.

    Thanks,

    Drew

  • Hi,

    Thank you for clarification.

    1. So, if I want to minimize the impact of the CTLE boost, beyond bypassing boost stage 1, what should I set for boost stage 2, if any?

    2. For the transparency, is it enough to use the default setting?

    Best Regards,

    Arie

  • Hi Arie,

    Do you have access to the programming guide for this device?  If not, please request access here:
    https://www.ti.com/drr/opn/DS280MB810-DESIGN

    1. The programming guide has details on configuring CTLE and the associated boost for each setting.  You will set boost based on how much equalization you want DS280MB810 to handle.
    2. Default setting does not have EQ bypass enabled.

    Thanks,
    Drew

  • Hi,

    Thank you for the information.

    I was looking for "bypass" functionality in terms of real "pass-through", without any CTLE or related effect, to provide the signal at terminal (pin) output as  received at terminal (pin) input. Is such option available?

    Best Regards,

    Arie 

  • Hi Arie,

    This option is not available.

    Would you consider DS280MB810 suitable for your application?

    Thanks,

    Drew

  • Hi,

    Thank you for the reply.

    We indeed consider DS280MB810 - this is our current choice, and our previous discussions were mainly directed at this device, including the last question about bypass option.

    Or have I misunderstood what you meant?

    Best Regards,

    Arie

  • Hi,

    Thank you for the reply.

    We indeed consider DS280MB810 - this is our current choice, and our previous discussions were mainly directed at this device, including the last question about bypass option.

    Or have I misunderstood what you meant?

    Best Regards,

    Arie

  • Hi Arie,

    Thanks for your response, I think you understood correctly.

    Please let us know if you have any additional questions.

    Thanks,

    Drew