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.

TXB0104-Q1: Could this damage the TXB0104 in such application?

Part Number: TXB0104-Q1
Other Parts Discussed in Thread: TXB0104, , TXU0304-Q1, TXU0304

Hi team,

     Urgent Support need.

     The customer use TXB0104-q1 in such application.MCU1, slave1 and slave2 are communicating via 4-wire SPI with date rate at 2.5Mbps.They meet two questions about TXB0104.

      Q1:Here is how they test: when MCU1 I/O is set high, the slave1 and slave2 get communication with each other via SPI. They find that when IMU_SPI_MISO send data, the SER_SEP_MISO would be pulled down about 0.6V lasting 20ns at 25℃. So they want to know whether this situation would damage the TXB0104 or not. And they find that the Slave1 and Slave2 would get CRC error via SPI communication at -20℃. Could you please have a look on this?

     

  Q2.They also get confused about the timing. Slave1 power up faster than the TXB0104, which means the Slave1 send date to TXB0104 while the TXB0104 would not get power up. Could this will damage the TXB0104?

In such situation, how much voltage of spi will damage the chip?、

YOURS

NAN

  • The TXB has extremely weak outputs (guaranteed drive strength is 20 µA), so it cannot be damaged by bus contention.

    However, the TXB is always bidirectional, and has edge accelerators, so all edges that it sees on any pin will be replicated on the other side. And its low-impedance inputs do load the bus.

    You should replace the TXB with the TXU0304-Q1. When the MCU does not communicate, disable the translator with the OE pin.

  • HI Clemens,

         Thanks for your reply.

         The weak drive strength means huge value of output resistors which could protect the TXB ,even though the port A is set high and port B send data. But the TXB is bidirectional, will the sending data on port B with port A being set high damage the device?

    The TXB has extremely weak outputs (guaranteed drive strength is 20 µA), so it cannot be damaged by bus contention.

    YOURS

    NAN

  • HI Clemens,

         So we can explain that the wave tested by customers is pulled down because  the O.S.  test the falling edge so that the port A would have same falling edge. But i do not know why it is pulled down lasting 20ns? (In my thought, I find the O.S. would last 10ns and the tpd is 10.6ns, which seems like 20ns.Am i right?

    However, the TXB is always bidirectional, and has edge accelerators, so all edges that it sees on any pin will be replicated on the other side. And its low-impedance inputs do load the bus.

    YOURS 

    NAN

  • Hi Clemens,

        Could you please tell me why customer pull down OE pin to diable the TXB1040?

    You should replace the TXB with the TXU0304-Q1. When the MCU does not communicate, disable the translator with the OE pin.

        

    YOURS

    NAN

  • As shown in the functional block diagram, all outputs have 4 kΩ series resistors. These limit the current (and allow other devices to override the voltage at the TXB's pins; this is how the bidirectionality is implemented).

    The O.S. timings are not guaranteed. 20 ns is a plausible value.

    I do not understand the question about OE. It has the same function with the TXB and the TXU.

  •      Why  should i replace TXB by TXU? 

    You should replace the TXB with the TXU0304-Q1.

          In addition, do you have some thought about why the temperature could effect the communication between SLAVE1 and SLAVE2 via spi while the PORTA is set high?

    yours

    NAN

  • Apparently, the TXB's edge accelerators or its inputs impair your signals. The TXU does not need edge accelerators, and has high-impedance inputs.

  • Hi Nan,

    In addition to echoing Clemens' feedback; note that the TXU has a higher drive strength with stronger outputs and without your one-shot concerns (i.e amplifying  reflections and false edge triggers due to large capacitance), as auto-sensing devices are intentionally designed to be weaker for such auto-sensing capabilities. Per your fixed directional SPI application, you seem to be in need of a different device such as the TXU.

    Also, lower temperature typically means lower output impedance. Hence, more prone to reflections from impedance mismatch per the PCB traces, thanks.

    Best Regards,

    Michael.

  • Hi Michael and Clemens,

         Thanks for your kind and professional reply.

         Let me do a summary

    1. In the picture above, when the A port is set high by MCU internally, the SLave1 and Slave2 communicate via spi. In this situation, the TXB0104-Q1 could not be damaged.

    2. While the Port A is set high, the port B voltage changing would pull down the voltage level temporarily  because the TXB0104 would replicate the port B 's falling edge on port A. Pulling down the OE of TXB0104 or using TXU0304 to avoid this.

    3.The  bidirectional TXB0104 would influence the SPI communication because the low input resistor which could cause the reflection on SPI bus, especially in low temperature.

    YOURS

    NAN

         

  • Hi Nan,

    1. Yes, setting A ports high will not damage the TXB especially if within the absolute max ratings.

    2.  Yes, inputs are replicated on the outputs.

    3. No, not the low input resistor but the output impedance of the TXB (except if you are referring to additional resistors for the MCU / peripherals; which are impacted by temp as well). 

    However, the output impedance of the TXB can be impacted depending on impedance of the traces and reflections can occur if mismatched, thanks.

    Best Regards,

    Michael.

  • Hi Michael,

         Thanks for your kind reply. I will update with customer. If any update, I will align with you ASAP.

    YOURS

    NAN