• State TI Thinks Resolved
  • Locked Locked
  • Replies 13 replies
  • Answers 4 answers
  • Subscribers 21 subscribers
  • Views 3639 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.

I'd like to design '4-SPI' & '1-Reset' using level shifter. Recommend a suitable part.

Inho Jeon
Expert 1452 points
Other Parts Discussed in Thread: SN74AVC4T774, SN74LVC1G34, TXB0108

Hi All,

I'd like to design '4-SPI' & '1-Reset' using level shifter.

And, in that web page ;  TXB-series is proper to SPI & SD. TXS-series is proper to I2C.    

But '1-Reset' is concerned. 

Recommend a suitable part me, please, ASAP.

Best Regard,

Inho Jeon

  • 0
    Guru 317000 points
    Is the reset signal push/pull or open-drain?

    Are all signals unidirectional? In which direction?
  • 0 in reply to Clemens Ladisch
    Expert 1452 points
    Hi Clemens,

    Reset signal is push/pull.
    And, all signal is unidirection.
  • 0 in reply to Inho Jeon
    Guru 317000 points

    For unidirectional signals, you should consider using unidirectional or direction-controlled voltage translators.

    How many in each direction? And what are the voltages?

  • 0
    TI__Guru* 97046 points
    When you say 4-SPI, do you mean "quad SPI"? I know this is becoming popular, and the data rates are extremely fast. I don't know of a translator that can support the full bandwidth.

    For traditional 4-wire SPI, I would recommend the SN74AVC4T774. Each channel (4) has its own direction control pin.

    Depending on the reset bit's direction (high to low or low to high) I would recommend different parts for it.

    There's probably a dozen other ways to solve this - it all depends on your system's specs. You really haven't given us any information to go on.
  • 0 in reply to Emrys Maier
    TI__Guru* 97046 points
    Hi Inho,
    I found your schematic in another thread, along with a repeat of these questions. In the future, please post only once. You would have gotten an answer much faster if you had included the questions in this thread along with the schematic you posted in the other thread.

    I see that you're trying to start the OLED_RST\ line low for a short time. Is this a requirement of your system?

    The capacitor/resistor chosen will produce a very slow edge - about 30ms to rise ~85%. Is that necessary, or can the reset be faster?

    Will there be a device controlling the OLED_RST\ line after the initial reset, or will this line be left high forever until power is lost?

    Which direction is the OLED_RST\ line going? From 1.8V to 3.2V or from 3.2V to 1.8V?
  • 0 in reply to Emrys Maier
    Expert 1452 points

    Hi Emrys,

    Thank you for your reply.

    I'm sorry for writing two questions in same subject.

    I answer your questios below this.

    On my schematics, R10 & C41 is need to delay time for system power on. 

    'OLED_RST\'  is for optional event of OLED display. So, On stable conditions 'OLED_RST\' is high until power is lost. 

    'OLED_RST\' is from 3.2V(MCU) to 1.8V(OLED_Display). 

    And, I wondering about TXS & TXB.

    I think TXB is best for my schematics, because TXS is designed for SPI, Is that right?

    Recommand right part for my schematics.

    Best regards,

    Inho Jeon.

      

  • 0 in reply to Inho Jeon
    TI__Guru* 97046 points

    Hi Inho,

    Thanks for the additional information.

    Down translation is actually very easy and opens up the number of parts available to achieve it.

    So, to achieve your desired startup reset of the OLED_RST, I would recommend eliminating the RC circuit and using the SN74LVC1G34. Here's a schematic of my idea:

    The buffer is the SN74LVC1G34. The inputs can be driven at 3.2V while the device is powered from 1.8V.  Note the 10kohm pull-down resistor. This will ensure that the reset is held LOW until the MCU is powered on and can control the RST\ pin itself.

    For the other pins, I would recommend using the SN74AVC4T774. This device allows for individual channels to be going different directions (A to B or B to A) but all channels are buffered and unidirectional, which I believe is what you need here.

    Together, the SN74LVC1G34 (0.8 x 0.8 mm) and the SN74AVC4T774 (2.6 x 1.8 mm) won't take up much space in your wearable device

  • 0 in reply to Emrys Maier
    Expert 1452 points
    Hi Emrys,
    Thank you for your reply.

    You mean TXS & TXB are right for my schematics, Is that right?

    And, You mean your recommaned parts(SN74XXXXXXXX) is replaced to RC delay circuit, Is that right?
    I think for issue of size & cost, I'll do not apply your recommanded parts for OLED_RST/.


    Best regards,
    Inho Jeon.
  • 0 in reply to Inho Jeon
    TI__Guru* 97046 points

    Hi Inho,

    You mean TXS & TXB are right for my schematics, Is that right?

         No, I never said that, nor do I mean it.

    From the schematic, it looks like you are connecting the translator to some kind of connector - I'm not sure of the capacitive load there, but a unidirectional solution will be far more likely to work in that situation.  TXS or TXB might function, but they are definitely not the optimal solution here.

    AVC4T774 isn't the only solution - it's just kind of a 'catch all' that will definitely work for your system. Unidirectional buffers are some of the smallest and lowest cost components available... so I'm not sure what you mean by 'issue of size & cost'

    You mean your recommaned parts(SN74XXXXXXXX) is replaced to RC delay circuit, Is that right? 
    I think for issue of size & cost, I'll do not apply your recommanded parts for OLED_RST/.

         Yes, that's what I mean.

    It's fine if you don't use my design, but I don't think you understand why I suggested it.

    You are connecting an RC circuit directly to a CMOS input. That will cause internal oscillation and shoot-through current during every power cycle, and will significantly reduce the lifetime of your semiconductor devices. You might want to read this application report: Implications of Slow or Floating CMOS Inputs

    If many of these devices fail once customers get them, how much will it end up costing the company when a few cents could have entirely avoided the issue?

  • 0 in reply to Emrys Maier
    Expert 1452 points

    Hi Emrys,
    Thank you for your reply.

    I mean 'issue of size & cost' is just for working samples. I know about your meant.
    But it is no time to show my working Spls for end-customer.
    In pre-prodution, I apply to your comments for my products.

    And, I have a question.
    On my schematics, for level shifting of line 5 ports ; 4 ports of SPI & 1 port of RESET.

    I saw your application resport on Ti web page below this ;
    "The TXB push-pull buffered type architecture uses autodirection sensing to remove the requirement for a
    DIR control signal to establish the direction of data flow. This architecture is designed to exclusively be
    connected and interfaced with a push-pull CMOS driver and is capable of driving capacitive or high
    impedance loads in applications such as Secure Digital (SD) or Serial Peripheral Interface (SPI).
    The TXS translators also use autodirection sensing and offer system designers a good solution to remedy
    mixed-voltage system incompatibilities when interfacing with open-drain drivers in applications such as
    I2C."
    So, I think 'TXB0108' is right to 5 ports of level shift on my schematics, Is that right?


    Please, Inform me that question, ASAP.


    Best Regards,
    Inho Jeon

  • 0 in reply to Inho Jeon
    TI__Guru* 97046 points
    Hi Inho,
    The TXB won't work with pull-up/pull-down resistors on the line. It has _extremely limited_ drive strength (20uA max).
  • 0 in reply to Emrys Maier
    Expert 1452 points

    HI Emrys,

    You mean TXB is not possible SPI interface, right? 

  • 0 in reply to Inho Jeon
    TI__Guru* 97046 points
    No, I didn't say that either.

    I mean what I said -- you can't have pull-up or pull-down resistors on the line with TXB - and your design clearly has the need for a pull-up resistor on at least one of the lines.