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.

TXB0108: TXB0108 on Nvidia Jetson SPI ports

Part Number: TXB0108
Other Parts Discussed in Thread: SN74LVC1G34

I need to set up an SPI link to a device whose only interface is TXB0108 (NVIDIA Jetson). The level translator is so weak that it can even hardly drive my oscilloscope probes. In case an SPI port of a PIC is connected the input signal is killed completely. Can a logic buffer be a solution for the issue? Even NVIDIA designer posted a note that since TXB0108 is bidirectional the outputs can be overridden by inputs and this exactly what I observe.

  • Hi,

    I'll have our level translation expert take a look at this when we return to work tomorrow morning.

    Thanks!

    Chad Crosby

  • TXB outputs are designed to drive high-impedance CMOS inputs, and nothing else. You can get problems with larger resistive/capacitive loads such pull-up/-down resistors, long traces, connectors, or cables.

    It would help to put a buffer (something like the SN74LVC1G34) as near the TXB's output as possible.

  • Hi Oleg,

    As Clemens mentioned it is possible to use the TXB device in conjunction with a buffer since SPI signals are unidirectional. He offered a single channel solution, but we also have higher channel count options. Are you needing this for all 8 channels?

  • All, thank you for the quick replies. While the problem is still there I probably found a workaround for the issue: the Jetson board has 2 SPI ports and SPI0 that I used initially is really non-functional, but moving to SPI1 allowed me to use SPI link. My guess is something wrong with the board design (schematic is not available at all, just pinout), at least with SPI0. Their support gave me a brilliant advice, to simply remove the level shifter and use direct connection to the CPU pins :) 

    There is one remaining issue that is still not clear. To verify connectivity I tried to configure 2 TXB buffered pins as an input and an output and see if the input captures a raising edge event. After it failed I hooked up a scope and found that the output is so weak that it cannot pull-up a similar pin configured as input. Is this some known expected behavior for TXB or there is a mistake in design? If not connected the output switches to high and low as expected and if connected to an external microcontroller the input also easily detects raising edge and high level.

  • The TXB is an auto-bidirectional level shifter. Both input buffers and output drivers are always active at all I/O pins; the outputs are deliberately weak to allow other devices to override the voltage level.

    For general-purpose I/Os, TXB is the wrong choice.