• State
  • Locked Locked
  • Replies 2 replies
  • Subscribers 21 subscribers
  • Views 703 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.

TXB108 got connected through 3ft flat cable

robin gangopadhya
Intellectual 855 points

Other Parts Discussed in Thread: SN74LVC8T245, TXB0108, SN74AVC8T245

We have a board where TXB108 level translator is seen driving a high density 16-wire ribbon cable of about 3ft. This cable is routed along with host of other flat cables.

The receiving side has SN74LVCxxx series driving similar flat cable.

During system integration, turns out that the cable position affects proper functioning of the interconnected boards ...some signals seem to be corrupting logic levels. By separating the cables & placing along a specific path inside the chassis seems to make the system functional.

So it is a case of such chips not being able to drive cables we have.

We do not have high speed signal at all in any path. Max may be 1MegHz clock.

There is really no level translation going on. All sides are 3.3V referenced to DGND on either side.

One way to debug would have been to remove/replace with a chip capable of driving such lengths. But I have not found any.

If I made a tiny pcb as an interposer board,  which chip would be most optimum to use in this situation?

We clearly want the cabling to be insensitive to where/how it is routed. Since I suspect drive capability  itself, adding shielding here would be making the situation even worse.

Looking for a clean solution.

thnx

robin

  • 0
    TI__Expert 6335 points

    Robin,

    For driving larger capactive loads, I would consider replacing the TXB0108 with SN74AVC8T245 or SN74LVC8T245. Keep in mind that would need to configure the DIR pins since the AVC/LVC family of traslators do not support auto-direction sensing.

    TXB translators support up to 70 pF of line capacitance. Do you know the line capacitance of the ribbon cable?

    Do you have any oscilloscope screenshots or any schematics that you can share with us?

    robin gangopadhya said:

    During system integration, turns out that the cable position affects proper functioning of the interconnected boards ...some signals seem to be corrupting logic levels. By separating the cables & placing along a specific path inside the chassis seems to make the system functional.

    We clearly want the cabling to be insensitive to where/how it is routed. Since I suspect drive capability  itself, adding shielding here would be making the situation even worse.

    1. What do you mean by 'cable position'? Can you illustrate this with an example?

    2. By 'seperating cables', do you mean increased spacing between the data channels, which helps to reduce crosstalk?

    Best Regards,
    Nirav

  • 0
    Intellectual 855 points
    thnx for the prompt reply, Nirav
    yes, I had them take oscilloscope pix & they have seen oscillatory waveforms.
    Capacitance of the flat ribbon cable 3ft...have not measured but you also have significant inductance plus mutual coupling among signals
    If they resist changes to eliminate root cause, I will go the extent of characterizing the impedance of the ribbon. But can be found in 'net as well.

    considering 74LVC540A because neither direction nor level shifting is being used in the circuit. If I get samples, I can almost fit it in the same footprint with some minor pin 1, GND & VCC isolation to test out the hypothesis. We have several such chassis with this issue & several non-driver buffers connected to several cables.
    what do you think?
    r