TXB0108: SPI Level Issue of TXB0108PWR

Hi Clemens,

We did remove all of pull-up/down and all open on signal pins of VCCA side.

Additionally, we removed all of pull-up/down on all signal of VCCB side

Also trace is about within 40mm~50mm from this part to MCU.

Same issue was observed.

For your information, they are using both uni-direction signal and bi-direction signal and need 8bit product, more small package.

TXB0104 support SPI. also TXB0108 is same scheme like TXB0104.

So I think that there is no issue.

Could you check and give me more debugging point?

If this issue was not cleared, they will change from TI to another vendor.

Pls help me and let us know if you need more information from our customer.

Best regards,

Michael

Please show these signals, on both sides of the TXB, with an oscilloscope.

Hi Clemens,

I think there may be a problem with the internal block or setting of this pin in QN9090THN(MCU+BLE), could you please review it a little more closely?

Could you check customer's schematic on issued pin which is yellow line?

Is there no need pull-up register on these issued pins?

QN9090THN data sheet : chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/www.nxp.com/.../QN9090(T)QN9030(T).pdf

PI1_MISO_QN : PIO1 (Pin No.1)
QN_RDYN : PIO17/ADC3 (Pin No.21)

Best regards,

Michael

Hi Clemens,

I will do debug it on their board today.

Could you please check it and give me feedback asap?

I have some questions.

1. Do I have to check voltage level of only MCU side after disconnecting MCU to TXB? or Should I check voltage level when MCU and TXB are connected?

2. To verify drive strength on issued pin of MCU, how do I test on their board?

3. Is there any way to solve the issue by adding a circuit on the customer's board?

4. I'm wondering about your comment that TXB does not work with pull-up/-down resistors.

    They are using TXB 2point per board.

    Their signal is I2C(external Pull-up), I/O(Internal Pull-up in MCU), SPI(Multi function in MCU).

    There is no issue except issued 2 signal as I mentioned above.

    If we suggest other part (8bit) to change this part, what is your recommendation part?

    Small package will be as good as possible.

Best regards,

Michael

1. The problem is how the MCU's signals and the TXB interact.

2. Look at the resulting voltage levels with a load (in this case, the TXB's output drivers).

3. No. The TXB0104 would have the same footprint as the TXU0x04.

4. The TXB does not work with I²C. The TXS0108E (a passive switch with internal pull-up resistors) might work with you signals. But for fixed-direction signals, a buffered (fixed-direction or direction-controlled) translator would be the best solution.

Hi Clemens,

I saw the following phenomenon while testing today.

I'm wondering how how this could happen.

Could you please review and feedback?

When I removed TXB on their board, SPI_MISO voltage of UWB_Module is 1.8V and SPI_MISO voltage of MCU is 3.3V.

and then when I mount the TXB on the board, The voltage level of both pins of TXB is 0V.

Best regards,

Michael

This sounds as if there are pull-up resistors somewhere. They are probably weaker than the TXB's outputs.

Hi Michael,

In addition to Clemens' feedback, could you double confirm if the phenomenon observed is for the open-drain driver? Thanks.

Best Regards,

Michael.

Hi Expert,

I tried test after floating or pull-down of MISO pin in the MCU side.

but the symptoms were the same.

Here is for your information additionally.

There are UWB_Module, MCU, TXB1080RGYR(VQFN) on NXP reference b'd.

UWB_Module and MCU are same part on customer's b'd like NXP reference b'd.

But The package of TXB1080 is different from NXP reference b'd (TXB1080RGYR) and customer' b'd (TXB1080PWR).

We swapped MCU of NXP reference b'd and customer' b'd.

And then test result is work well on NXP reference b'd. but same issue has happened on customer's b'd.

Is there a different between TXB1080RGYR and TXB1080PWR?

Also, could you give me debugging point on customer' b'd?

Best regards,

Michael