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.

ds100mb203 MODE Control effect on outputs

Other Parts Discussed in Thread: DS100MB203

We are using DS100MB203 and controlling it's operation with the SMBus (not pin mode). We have register 08 set to a C (bit 2 =1) and register 2D set to A8 (bit 6 = 0). As I understand the datasheet this should set the device in PCIEGen3 MODE, which should put the outputs in "transparency mode".

My question is - does "transparency mode" mean that the output controls for VOD (2D bits 2:0) and DEM (2E bits 2:0) have NO EFFECT on the output? The ouput should just follow the signal seen at it's input?

  • Hi Torry,

    Register 0x08 = 0C'h allows SMBus registers to control the device MODE and RXDET.

    Register 0x2D = A8'h Sets the output into "non-limiting" mode and the VOD = 600 mV setting

    For these register changes to have any effect, Register 0x06 must be changed from the default of 10'h to 18'h.

    "Transparency" mode or "non-limiting" mode is driver setting which greatly reduces the voltage gain from input to output.  The device operates more like a linear equalizer.  In this output mode, the VOD will still change the output amplitude and DEM will have some effect as well.  In this mode I recommend keeping the VOD between 900-1200mV and DEM at or less than -3.5dB.

    Regards,

    Lee

  • Lee:

    Thanks for the rapid response.

    We use the DS100MB203 to fanout (1:2) unidirectional signals delivered over cables (DisplayPort up to 8.1Gbps).

    We would prefer to keep one set of outputs as close to the input signal as possible.

    1)  Is there a better way to do this than the one we described above ?

    2) What effect does the "transparency" mode have on EQ settings made at the inputs?

  • Hi Torry,

    1. When you say "keep one set of outputs as close to the input signal as possible," are you referring to maintaining the signal shape? Typically we see this as a desirable trait for systems that required some form of link-training where pre-cursor and post-cursor information are applied to the input signals to the redriver or mux/buffer. If this is the case, then transparency mode will be more helpful in keeping the signal output shape as close as possible as the input shape.

    2. Using limiting mode will act to make the output signals appear quantized, with sharp edges in order to improve the overall signal quality. In non-limiting/transparency mode, the driver will act to apply any VOD gain or de-emphasis linearly. The transparency mode does not have an effect on the EQ settings, as these driver modes are applied after the signal passes through the EQ stages.

    Thanks,

    Michael