• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Subscribers 32 subscribers
  • Views 1150 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

SN65DP141 AC Gain Setting Datasheet Clarification

Trenton Reed
Expert 3010 points

Hi Team,

The datasheet contains conflicting information. In section 6.6 the typical equalization gain is 15dB. Figure 16 shows the equalization can go up to 15dB. However, Table 2 has a range of 7-17 EQ_AC_GAIN (dB) under 000-111 EQ Setting and RX_GAIN of HiZ. Similarly table 3 has a 7 to 17 dB range. 

Can you clarify if the max AC gain setting is 15 or 17 dB? 

Is there a reason why it's not a linear setting? i.e. 000 = 1 dB, 001 = 2 dB, ... , 111 = 8 dB instead of 1dB to 9dB

Thank you,

Trenton Reed

  • 0
    TI__Mastermind 23815 points
    Hi Trenton,

    You could use 17dB as the total AC gain.
    Total AC gain is not determined only by EQ settings, it is affected by Rx gain too, this is why it changes form 15dB in some places to 17dB in others.

    Gain may not be linear, that's why DS mentions a range instead of gain for each step, besides as I mentioned before it is affected by Rx gain.

    Regards
  • 0 in reply to Moises Garcia
    TI__Expert 3010 points
    What is the length of the defined long input cable in table 3?

    What is the scale for short input swing in table 3?

    There doesn't seem to be an option for long input cable AND short input swing. This is because in practical application a long input cable/trace will always yield a relatively defined large input swing, correct? It will help to understand with some framing.
  • 0 in reply to Trenton Reed
    TI__Expert 3010 points
    Any word on my additional questions Moises?
  • 0 in reply to Trenton Reed
    TI__Mastermind 23815 points
    Hi Trenton,

    Sorry for not replying I'm still looking for detailed information about table 3.
    I'll come back when I have suitable data.

    Regards
  • 0 in reply to Moises Garcia
    TI__Mastermind 23815 points
    Hi Trenton,
    This long and short cable/trace is just for reference it is just that a longer cable/trace will have more loss than a short one.

    EQ_DC_GAIN:
    This is to attenuate the signal if input swing is large, it doesn't matter if cable/trace is long or short.
    Input swing would need to be attenuated to avoid amplifier saturation.

    RX_GAIN<1:0>:
    This will select the EQ range to use, from 1 to 9, 2 to 10 or 7 to 17 dB
    When EQ_DC_GAIN=0 you can have 1 to 9 or 7 to 17 dB AC gain, since this configuration has a higher gain you could use it with a longer cable with more losses.
    When EQ_DC_GAIN=1 you can have 1 to 9 or 2 to 10 dB AC gain, since input signal is not being attenuated AC gain increases very little when you select RX_GAIN=11.

    I hope this clarifies table 3.

    Regards
  • 0 in reply to Moises Garcia
    TI__Expert 3010 points
    Thanks for the added clarity. Can you please confirm the following:
    "There doesn't seem to be an option for long input cable AND short input swing. This is because in practical application a long input cable/trace will always yield a relatively defined large input swing, correct?"
  • 0 in reply to Trenton Reed
    TI__Mastermind 23815 points
    Hi Trenton,

    The input swing will depend on swing from source and the losses in cable/trace.
    You may have a very large source swing, and a very good(short) cable with little losses, so DP141 input swing will be high.
    You may have a small source swing, and a very bad(long) cable with a lot of losses, so DP141 input swing will be small.

    There is no option for long input cable AND short input swing, because a short input swing doesn't require attenuation, so, EQ_DC_GAIN=1.
    If the is no attenuation, RX_GAIN=11b will increase AC gain very little thus a long cable(assuming more losses) couldn't be supported.

    Regards