• State TI Thinks Resolved
  • Locked Locked
  • Replies 16 replies
  • Answers 3 answers
  • Subscribers 34 subscribers
  • Views 670 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.

Original question:

TCA9517A: Datasheet specification inquiry

TCA9517A: TCA9517A: Datasheet specification inquiry

Tommy Tuu
Prodigy 150 points

Part Number: TCA9517A
Other Parts Discussed in Thread: TCA9517

Tool/software:

Hi Tyler

about thr TCA9517A Bside

Has this been repeated across multiple TCA9517A's?

Yes, it does

I just want to make sure in the simulation and the measurement different

in the simulation the result shows the VOLB is about 0.51V, which is in the SPEC 0.45V~0.6V

but in the measurement result shows 0.39V which is lower than the simulation result and SPEC

what cause the different, is it the probe side? like at pin or at die, because the probe side we measure isn't at die, but the simulation is.

Thanks

Tommy

  • 0
    TI__Guru 55568 points

    Hi Tommy,

    The simulation result makes sense. VOLB = 0.51V when driving a LOW from A-side to B-side sounds correct and is within the datasheet spec. 

    Tommy Tuu said:
    what cause the different, is it the probe side? like at pin or at die, because the probe side we measure isn't at die, but the simulation is.

    It could be different depending on where the voltage is being measured from. Is it possible to measure VOLB near the pins of the DUT? Is the probe calibrated properly - Any error induced by the probe? What is the loading condition on B-side - i.e. how strong are the pull-up resistors? 

    Any scope captures to show of the measurement? 

    Regards,

    Tyler

  • 0 in reply to Tyler Townsend
    Prodigy 150 points

    Hi Tyler

    I'm checking the probe side, it do measure VOLB near the pins of the DUT, but it still like picture below

    but I have to check one thing first

    I saw the datasheet below

    Is that means the IOL shuld be 100uA or 6mA, not 100uA to 6mA? because we use in output is 3.3V and pull up 1K ohm
    and what does it means VILA = VILB = 0 V?

    Thanks

    Tommy

  • 0 in reply to Tommy Tuu
    TI__Guru 55568 points

    Hi Tommy,

    IOL can be either 100uA or 6 mA. Another way of saying this is that the PU resistance is very weak (100uA), or the PU resistance is very strong (6mA). 

    VILA = 0V means that the input signal on A-side is biased to 0V or GND. I don't think it should say VILA = VILB, since this is the output for VOLB on SDAB and SCLB. 

    What is the "base" measurement measuring? I am not familiar with that scope measurement. 

    Can you try using the scope cursors to see if they match the base measurement? Measure the low period using the cursors to compare. 

    If not of these work, can you bias SDAA = GND and measure SDAB? Holding the input LOW is the same way it is measured in the datasheet. 

    Regards,

    Tyler

  • 0 in reply to Tyler Townsend
    Prodigy 150 points

    Hi Tyler

    we use in output is 3.3V and pull up 1K ohm so it means IOL will be 3.3mA. 

    Is that means it doesn't apply to this VOL SPEC?

    our measure is like picture below, one is probe on the output side(at pin) and on is on GND

    But I still confused why the simulation result is different to measurement result

    Thanks

    Tommy 

  • 0 in reply to Tommy Tuu
    TI__Guru 55568 points

    Hi Tommy,

    That is fine. ~3.3mA is a good current load within datasheet specifications. The SVO should still be within the datasheet spec of 0.45V to 0.6V. 

    Have you tried any of my above suggestions? 

    Regards,

    Tyler

  • 0 in reply to Tyler Townsend
    Prodigy 150 points

    Hi Tyler

    we try to use SDAA = GND and measure SDAB, but the waveform looks bad, please see the picture below, I'm not sure how to let VILA = 0V , because our supply in Aside is 1.8V, we can't turn it off.

    I'm not pretty sure the error between simulation and measurement, maybe is the probe affect?

    or something I miss

    Thanks

    Tommy

  • 0 in reply to Tommy Tuu
    TI__Guru 55568 points

    Hi Tommy, 

    VILA = 0V means that you GND one of the terminals. Either SDAA or SCLA. Once the input is LOW (GND), then you can measure a clean DC voltage on the output either SDAB or SCLB. 

    I2C signal should not be running at this time. This is just to test the output of the buffer. 

    It could be variations in the probe. Have you tried different probes on the same channel to see if they yield a different voltage measurement? 

    Also, what does the BASE(3) measurement actually measure? Is it an average of the LOW signal? 

    Regards, 

    Tyler

  • 0 in reply to Tyler Townsend
    Prodigy 150 points

    Hi Tyler 

    We tried the different board but same design
    the result shows is can up to 0.45V, that's make sence 
    the Base3 is the avg VOL

    So it means different board will lead to different result?

    but what's the reason?

    Thanks

    Tommy

  • 0 in reply to Tommy Tuu
    TI__Guru 55568 points

    Hi Tommy,

    This scope capture looks like the B-side is being externally driven by a target device which is why the base(3) voltage is 310 mV, way less than the static voltage offset output from the buffer which is > 0.45V. 

    In the second scope capture, the buffer is producing the low voltage at ~0.52V. A low is being driven on A-side, the buffer re-drives this signal on the B-side output and has a voltage around ~0.52V. You'll notice that there is no pedestal voltage step when the buffer releases high. 

    Are we certain that we a measuring the same data and same direction when comparing the output voltage to the datasheet? The direction of the data being buffered matters here. 

    Regards,

    Tyler

  • 0 in reply to Tyler Townsend
    Prodigy 150 points

    Hi Tyler 

    we do use the same design and end device is same either, it's the same board but different films

    and the other probe way is all the same.

    So I get little comfuse, what cause it.

    BTW where is the 0.52V?



    Thanks

    Tommy

  • 0 in reply to Tommy Tuu
    TI__Guru 55568 points

    Hi Tommy,

    When you say "different films" are you talking about different tape and reels? Do you have lot information to search this? 

    The 0.52V is in the datasheet. It is the voltage offset, on B-side when driving a LOW on A-side. On the customer's scope captures, I believe it is located in red. 

    Regards,

    Tyler

  • 0 in reply to Tyler Townsend
    Prodigy 150 points

    Hi Tyler

    different films means the same design but different board like 1&2


    located in red means here?

     board1

     board2

    the picture is measurement of board1 and board2, my question is how cause the different, 1 pass and 1 not pass. And what's mean the target Low?

    Thanks

    Tommy

  • 0 in reply to Tommy Tuu
    TI__Guru 55568 points

    Hi Tommy,

    When an I2C target device on B-side pulls the SDAB line LOW, it pulls to a different VOL that should be < 0.4V. 

    In the first capture that you copied, BASE(3) = 310 mV, which I assume is the VOL of a target pulling LOW. 

    In the second scope capture, I see that the BASE(3) = 540 mV which suggests that a device on A-side of the buffer is pulling LOW, which would cause an output voltage on B-side ~ VOLB = 0.52V. The scope measured 540mV, so this suggests that this is the buffer output voltage. 

    These two scope captures don't look to be the same I2C frame? The 1st capture looks to be communicating from B-side to A-side, with the ACK bit occurring from A-side to B-side. 

    The 2nd scope capture looks to be communicating from A-side to B-side, with the ACK bit occurring from B-side to A-side. 

    See this FAQ

    In short, the I2C looks fine. I think it is the direction across the buffer that is confusing. 
    When you say "a board fails" do you mean that the I2C host and target devices cannot communicate with each other? I.e. they don't recognize I2C data because of the buffer in between? 

    Regards,

    Tyler

  • 0 in reply to Tyler Townsend
    Prodigy 150 points

    Hi Tyler,

    These two scope captures don't look to be the same I2C frame? 

    -> It's the same I2C frame, same design but different board.

    "a board fails" means the VIL didn't come up to 0.45V~0.6V



    Thanks

    Tommy

  • 0 in reply to Tommy Tuu
    TI__Mastermind 47731 points

    Hi Tommy,

    The team is out on US holidays today and responses will be delayed. Thank you for your patience.

    Regards,

    Jack

  • 0 in reply to Tommy Tuu
    TI__Guru 55568 points

    Hi Tommy,

    I am back from being ooo. 

    Have we tried A-B-A swap? What I mean is to take the "bad unit" (The TCA9517 showing 310 mV offset), and soldering that device onto a known good board to see if the issue follows the IC replacement? Do we have the resources to conduct this type of test? This would help to eliminate if this is an IC related issue rather than a board issue. 

    Did my previous reply make sense? Are we aligned on the direction of communication across the buffer? It matters for measuring the offset voltage. We are certain we are driving a signal from A-side to B-side? If communication is from A-side to B-side, the 0.52V offset makes complete sense. However, if we are driving from B-side to A-side, then the voltage on B-side will be lower <0.4 V. 

    Regards,

    Tyler