• State TI Thinks Resolved
  • Locked Locked
  • Replies 10 replies
  • Answers 1 answer
  • Subscribers 19 subscribers
  • Views 1924 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

SN74AVC4T245: Vcc Isolation Feature

Mark Guastaferro
Genius 11105 points
Part Number: SN74AVC4T245
Other Parts Discussed in Thread: SN74LVC125A

SN74AVC4T245 is being used to translate between 1.8V and 3.3V.  VccA = 1.8V, VccB = 3.3V.  Device is set to translate from B to A on all ports. See attached schematic.  

Noticed that VccB current increases by about 45 mA  when VccA is turned off. There are other devices on the VccB rail, but seems unlikely any of these are source of current increase. 

When VccA is off, VccA voltage falls to about 0.2V.  Is this low enough to place device into Vcc isolation or Ioff mode? All I/Os on B side are pulled up to VccB (even pin 14 which does not show pull up on schematic). 

SN74AVC4T245.pdf

  • 0
    Guru 243980 points

    The datasheet says that VCCA would have to be at GND. Where does that 0.2 V offset come from? Is that voltage floating (which would certainly not work)?

    Would it be possible to set the OEs high when the power is off?

    Please note that for unidirectional down-translation, you need nothing more than a buffer with overvoltage-tolerant inputs, e.g., SN74LVC125A.

  • 0
    TI__Guru 56100 points
    hi Mark ,

    during hi-z , the state of the DIR pin doesnt matter( bto A or A to B),the i/o ports would require defined voltage levels. It needs to be treated as buffers with floating inputs.
    can you try having weak pulldown or pullup on A side when the Vcca is held at 0V ? and as Clemens suggested , why is the 0.2V at the Vcca when turned off ?
  • 0 in reply to ShreyasRao
    TI__Genius 11105 points

    I'm trying to confirm what is causing the 0.2V reading to determine if VCCA is floating.  

    One thing I'm a little confused about regarding VCC isolation is that in App Note SCEA030A on similar voltage translators, it indicates that when a VCC supply is off (= 0V), the inputs on that side are disabled.  To me, this implies that you wouldn't need to worry about floating inputs.  The App Note also has a paragraph talking about floating inputs are OK if the respective VCC for that port is floating as well. 

    Even with VCCA = 0.2V, It's hard to understand how floating inputs cause high bias currents since that doesn't seem to be enough voltage to turn on any internal FETs or diodes.  

  • 0 in reply to Mark Guastaferro
    TI__Guru 56100 points

    Mark ,

    Was there any outcome when trying out pulldown/pullup resistors on the A side ?

  • 0 in reply to ShreyasRao
    TI__Genius 11105 points
    Pull down resistors on the A side did not change the results.
    The translator was removed from the circuit and the VCCB current draw drops considerably, verifying that the translator is the cause of increased current in the partial power-down mode.
    It is still not clear as to the source of the 0.2V on VCCA when this supply is disabled. Next, a pull-down resistor on VCCA will be tried to see if the voltage can be pulled-down and to make sure to provide a resistive path to ground for the Vcc isolation/Ioff feature.
    Would swapping the translator ports help (connect all B side I/Os and VCC to A side and vice versa)? In this case then VCA would always power before VCCB. I didn't think the power-up order mattered, but maybe it does.
  • 0 in reply to Mark Guastaferro
    TI__Genius 11105 points

    A pull-down resistor was placed on VCCA  to make sure this voltage pulled to ground when it was disabled.  The current drawn by the translator is still high. 

    We are not sure what to try next. Would swapping voltages and ports provide any potential improvement? 

    If the translator was replaced with a uni-directional buffer like the SN74LVC125A, would there be any concerns about it's state with the 1.8V VCC disabled and 3.3V inputs still present (pulled up to 3.3V)? 

  • 0 in reply to Mark Guastaferro
    TI__Guru 56100 points
    Hey Mark ,

    I am sorry about the issues you are facing with the product.
    I have not come across this issue before even after having the pulldown resistors on the IO ports. Have you tried different devices to confirm this is or is not repeatable.
    The unidirectional buffer should not have concerns as the inputs are overvoltage tolerant at any valid Vcc.
  • 0 in reply to ShreyasRao
    TI__Genius 11105 points
    Hi Shreyas,
    This issue has been seen on multiple boards, so it seems it is happening with every unit and not a singular device failure.
  • 0 in reply to Mark Guastaferro
    TI__Guru 56100 points

    Mark ,

    To summarize :
    With Vcca at 0.2V, there is ~45mA of current through the Vccb pin.
    There are pulldown resistors on both the IO ports (A and B ports) when the Vcca is powered off to 0V and is being seen on multiple units and boards and not with just a single unit. Also , this device has been isolated from both driver and receiver side and experimented as standalone and the issue persists.
    Please confirm if this is true.

    Please work with the local sales office for initiating an FA on this.

  • 0 in reply to ShreyasRao
    TI__Genius 11105 points

    Hi Shreyas,

    Yes, you're summary is basically correct.  Here are a few clarifications: The 45 mA current increase is seen at the power supply and goes away if the device is fully powered or removed.  The assumption is the current is going into the 245 buffer.

    There are pull-ups on the B-side and pull-downs on the A side. 

    I don't think I would say that the device has been isolated from both the driver and receiver side and run stand-alone. With the current board design, it is not possible to physically isolate the device without cutting traces. 

     I will look to start an FA.