• State TI Thinks Resolved
  • Locked Locked
  • Replies 11 replies
  • Answers 2 answers
  • Subscribers 20 subscribers
  • Views 3617 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.

TXS0108E: (TXS0108EPWR) I2C Problem

Arnold Braun
Prodigy 60 points
Part Number: TXS0108E
Other Parts Discussed in Thread: LSF0108, 2N7001T

Hello, 

i try quit simple to use TSX0108EPWR for level translation. 

For Testing VCCA = VCCB = 3V3 

OE = 3V3

Capacitor 100nF on VCCA,VCCB

With out this part everything is ok. but if i use it, SDA and SCL have random values. 

Pullup 50K have no effect. 

Does it realy work. 
Levelshifting with simple FET works fine. 

I tried already 3 diffrent. All have the same result

Thanks Arnold

  • 0
    Guru 316440 points

    The description of the circuit sounds OK.

    Please show the schematic, and ideally a photo of the circuit, and tell us how you have measured the voltages.

  • 0
    TI__Guru 61870 points

    What is the operating freq? Why is the TXS used with both voltages at 3.3V, so no translation done?

    You can also consider using the LSF device, but needs Vccb > Vcca +0.6V...

    please show scopeshots, schematics of your setup.

  • 0 in reply to Clemens Ladisch
    Prodigy 60 points

    Hello, 

    the voltage is mesuared with Osciloscope Hantek and with I2C Logic analyser. 

    Just now i have problem to upload the schematics. VCCA, VCCB have 100nF to Ground. VCCA = 3V3,VCCB tested with 3V3 and 5V
    OE = 3V3 

  • 0 in reply to ShreyasRao
    Prodigy 60 points

    Frequenz = 100K for now. later will be 400k
    Sorry i can not Change to LSF because my prototype PCB the foodprint may not fit. 
    If there would be some comment in the datasheet that it don´t work with I2C then i would have used diffrent part. 

    Testing Voltage 

    VCCA = 3V3 
    VCCB 3V3 and 5V productiv: 5V

    Thanks Arnold

  • 0 in reply to Arnold Braun
    TI__Guru 61870 points

    Arnold,

    The TXS is capable of 3.3V <> 5V translation for the I2C interface. I would suggest not to have external pullup resistors as the TXS internally has pullups.

  • 0 in reply to Arnold Braun
    Guru 316440 points

    The internal pullups of the TXS will conflict with any external pulldowns, such as the one on the ESP-INT line.

    Are there any other pulldowns on some data line?

    The TXS works fine with I²C on many other boards. The difference must be in the rest of your circuit.

  • 0 in reply to ShreyasRao
    Prodigy 60 points

    Hello, 

    I also tried without pullup´s. same result

    thanks

  • 0 in reply to Arnold Braun
    Prodigy 60 points

    Measuring resistor from SCL, SDA without TXS

    SCL-GND  18M

    SDA-GND 18M

    SCL-5V 10M

    SDA-5V 10M

    SCL-3V3 8M

    SDA-3V3 8M

    The pullups in the schematic are not soldiert

  • 0 in reply to Arnold Braun
    TI__Guru 61870 points

    Pls remove the pullups and other external components tied to this and measure.

    Whats 18M, 10M and 8M here? pls share the scopeshots if possible.

  • 0 in reply to ShreyasRao
    Prodigy 60 points

    Thanks, 

    I already mentioned that i tried without pullup. without success. And i´m wondering where is the image from my scope. 
    I already published once. Should be in the same post as the schematic

    I already changed the prototype to use simple BSS138. I just wanted to save space, because i need to translate 8 bit. 

    I will throw out this part and make a new PCB with BS138. At least i know this works.
    The whole solution have a few Microprozessor. they all sharing I2C or SPI.  
    I read already in a few Groups that this part have problem with I2C. They can not all be wrong. 
    I expect from a Level translator that it is relaiable and works. 

    Thats the scl. 

  • 0 in reply to Arnold Braun
    TI__Guru 61870 points

    Hi Arnold,

    I am sorry to hear that the TXS0108 couldnt work for you. I see from the scopeshots that the first 3 pulses have issues but subsequent pulses are fine. I know that there is clock stretching in I2C wherein the slave device can try to hold the clock low if the data communication is fast. would this be a possibility here?

    Since the clock is unidirectional, you could consider using the 2N7001T which is 1 bit level translator for 1.8 <> 3.3V.  The LSF0108 is another 8 bit auto bidirectional device similar to the NFET being used in your system currently which provides additional ESD protection, and integrated solution.

    Please let us know what you decide.