• State
  • Locked Locked
  • Replies 5 replies
  • Subscribers 19 subscribers
  • Views 898 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.

TXB0104DR Latchup?

Terje Froysa
Intellectual 330 points
Other Parts Discussed in Thread: TXS0104E, SN74AVC4T774

Dear Forum,

I have just started testing a cape for the BeagleBone Black. I have equipped the GPIO lines with TXP0104DR level translators. At the test, VCCA=VCCB=3.3V.

Setting the I/O lines as outputs and applying static high/low results in correct 3.3V/0V ouput levels.

However, If I connect one line to the gate of a MOS transistor and a 180K pull-up, odd things happen.

When connecting the I/O line while high, it remains high. But when setting it low it settles on 1.604V and seem to latch up there. Setting the line high again rises the voltage to 1.628V.

The resistive load > 50KOhm and capacitive load <70pF should be OK for the translator.

Please advice (Schematics attached).

Best regards
Terje Froysa

2350.LevelTranslator.pdf

  • 0
    Intellectual 330 points

    Added information to my questions above:

    Having a TXB0104DR configured VCCA=3.3V VCCB=5.2V.

    Input to TXB: (A-side) goes from 0.003V to 3.31V

    Output from open-ended TXB (B-Side): goes from 0.002V to 2.457V (Measured with Fluke multimeter)
    But if I take the metering probe away and back a couple of times I suddenly read: 5.199V

    This level is stable until I set the A-side low. Then output level drops to 2.499V
    If I remove Fluke probe and puts it back, the output level changes to 0.002V.

    Hence the B-output stops midway between high and low an must be exited by the light impact of my Fluke probe to fullfill the transaction.

    Please advice.

    Best regards
    Terje Froysa

  • 0 in reply to Terje Froysa
    TI__Expert 6430 points

    Hi,

    50cm cable is too long. Is is possible to decrease cable length and test again? if there is an external pull-up resistor, it is better to use TXS0104E.

    Thanks

  • 0 in reply to Wei Lang
    Intellectual 330 points

    Hi,

    Thanks for your answer. I will consider using TXS104E.
    If I need to extend the cable to 1.5m, what pull-up resistor value would you reccomend?
    Is it feasible to use the TXS104E in a 400KHz SPI (single port-to-port) connection over 1.5m of cable?

    By the way, my observations of the TXB104 was without cable, using interactive python commands to set lines high and low. When I wrote a python script toggling the lines at 1Hz, the "latchup" dissapeared. The rise-time should be equal to toggeling the port interactively. I find this difference in behaviour quite odd.
    In my opinion, there should not be any difference in toggeling a line once a minute and at 1Hz.
    If you have a theory about this I would like to hear.

    Best regards
    Terje Froysa

  • 0 in reply to Terje Froysa
    TI__Expert 6430 points

    Hi,

    If there is 1.5m cable, I think TXS0104E is not suitable device. For SPI application, please consider SN74AVC4T774. For your test results about 'a minute and 1 Hz',  I have no idea about it. But for TXB010X, the I/O is clamped to mid-voltage could be caused by transmittion line effect, output load (pull-up,pull-dwon resistor), and input signal transition edge.

    Thanks

  • 0 in reply to Wei Lang
    Intellectual 330 points

    Thanks for your feedback,

    In the meantime I have tried out the TXS01xx devices and find no problems in driving cables of 7m even without adding pull-up resistors on the far side.

    Best regards

    Terje Froysa