• State TI Thinks Resolved
  • Locked Locked
  • Replies 12 replies
  • Answers 4 answers
  • Subscribers 20 subscribers
  • Views 5120 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.

TXB0108: One-shot circuitry and short input pulses

fdm
Expert 1070 points
Part Number: TXB0108
Other Parts Discussed in Thread: LSF0204

Hi,

How should the TXB0108 respond on input pulses shorter than tW_min (=10ns for the 3.3V<->5V case) - in order of, say, 1-3ns?

Let's consider the following scenario:
The TXB0108 is used for voltage level translation between 5V CPU and 3.3V static RAM data bus.
During a read transaction the CPU releases the data bus and then asserts the OE (Output Enable) signal.
Upon OE assertion the RAM starts driving the data bus.
There is a short interval - say, 1 to 5 ns according to the RAM specification - before these output data become valid.
It means that data lines may change their state one or more times during that time.

How will the edge-rate accelerator within the TXB0108 react on these short pulses?


BR,
   Denis

  • 0
    TI__Guru* 96591 points
    Hi Denis,
    I have notified the appropriate applications specialist and he should be with you soon.
  • 0 in reply to Emrys Maier
    Expert 1070 points

    Hi Emrys,

    Any updates on this?

    Regards,

    Denis

  • 0 in reply to fdm
    TI__Guru* 96591 points
    My apologies Denis. I have reassigned this post and Shreyas will be helping you out soon.
  • 0
    TI__Guru 56110 points
    Fdm ,

    The edge rate accelerator senses the rising edge input on either side of the I/O ports. The input level which triggers the one shot is approximately Vcc/2 ; where Vcc is related to the port which the input is applied to.
    Ideally , OE needs to be set such that the I/O ports are disabled during the interim condition to avoid false glitches.
  • 0 in reply to ShreyasRao
    Expert 1070 points

    Shreyas,

    As far as I understand your answer, it indirectly confirms that the TXB series is not suitable to interfacing such devices as a fast asynchronous SRAM.

    Is there any background info about these edge-rate accelerators - minimum input pulse duration that triggers one-shot circuitry, its retriggerability, dead time etc.?

    BR,
       Denis

  • 0 in reply to fdm
    TI__Guru 56110 points
    Hi Denis ,

    unfortunately , we do not have more information on the one shot timing analysis apart from the approximate duration it is turned on which is about 20 to 30ns. This leads me to believe that the dead time is ~ <10ns. It may cause one or two false glitches based on what you have described. If the CPU is programmed to ignore data for less than <20ns I believe this will work.
  • 0 in reply to ShreyasRao
    Expert 1070 points

    Hi Shreyas,

    According to the SCEA043 Application Report "A Guide to Voltage Translation With TXB-Type Translators" the edge-rate accelerator is turned on "approximately 10 ns or 95% of the output edge, whichever occurs first" and ~4.5ns with no load, isn't it? That 30ns duration seems more relevant to the TXS series.

    I have tight timing budget for read transactions (<10ns) and CMOS input levels at the 5V side in the project I'm working on, so the TXB series of translators seems not suitable for me anyway.
    I suppose the LSF series would be more appropriate, but unfortunately there's neither static nor dynamic characteristics in the DS at 5V, especially for up translation 3.3V to 5V.

    BR,
       Denis

  • 0 in reply to fdm
    TI__Guru 56110 points
    Hi Denis ,

    The TXS and TXB auto bidirectional both have one shots in them. The characterization of the one shot is not fully done and hence this tight requirements will be a good feedback to the product engineering and our system engineering team to define more robust specs.
    The LSF as you mentioned doesnt have the one shot accelerators in them and since this is a passive translator , the switching characteristics will be defined by the pullup resistors.
    are there any particular static specs you are interested in ? I can try to find it in the char data if it was collected.
  • 0 in reply to ShreyasRao
    Expert 1070 points

    Shreyas,

    I am interested in Ron maximum values (DS has only typical ones) and Cio(on)/Cio(off) at 5V.

    BR,
      Denis

  • 0 in reply to fdm
    TI__Guru 56110 points
    Denis,

    Do you want room temp data or across temp ?

    I have the room temp data collected for the following conditions :

    Conditions: VrefA =3.3V ; VrefB= Ven=5V and Vin =0V, I= 64ma
    Ron avg = 3.1ohm

    I do not have the capacitance data at 5V currently.
  • 0 in reply to ShreyasRao
    Expert 1070 points

    Shreyas,

    This  Ron avg = 3.1ohm  value corresponds well to the 3ohm typical value from the LSF0204 DS.
    I'm interested in the Ron maximum values across all PVT variations.

    BR,
      Denis

  • 0 in reply to fdm
    TI__Guru 56110 points

    Denis ,

    The data i have shows the max value of the ron at the condition listed to be 11.85 at ohms at 125C . Although, this is tested across 30 units which will give a good normal distribution , this is not account the process variation.