• State TI Thinks Resolved
  • Replies 9 replies
  • Answers 1 answer
  • Subscribers 19 subscribers
  • Views 120 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

SN74AXC8T245EVM: SN74AXC8T245EVM

Owen Wang
Prodigy 10 points
Part Number: SN74AXC8T245EVM
Other Parts Discussed in Thread: SN74AXC8T245, TXB0108, SN74AVC16T245

Tool/software:

Hi there,

We bought 2 SN74AXC8T245 Evaluation Modules, but unfortunately the channel 1 of the 8 channels does not work for both EVMs.

We set DIR1 and DIR2 both to low, which means the level shift is from B to A, but channel 1 doesn't work (when pulling high or low in the B side, we didn't see the high or low on the A side), but all other 7 channels does work. 

BTW, we put both VCCA and VCCB to 3.3V as we mainly target to enhance the driving force of our original GPIOs.

One side question, from the info of SN74AXC8T245 datasheet, it seems the output driving force is max. 12mA, strong enough to directly drive a LED. the question is if we don't put a resistor cascaded with the LED from the A side GPIO, or say directly driving a LED with GPIO on A side, will that cause any potential issues for SN74AXC8T245? Thanks a lot!

BR,

Owen

  • 0
    TI__Mastermind 25040 points

    Hi Owen,

    It should not cause the issue. Except if the ch 1 input is not fast enough for device to switch but it sounds like you are not toggling the input. 

    Since all other inputs work, i presume OE is configured correctly.
    Is the same issue observed with a known good evm and unit?
    Is the input measuring high? Do you have block diagram to show the setup as well as measurements/observations?

    Best Regards,

    Michael. 

  • 0 in reply to Michael Ikwuyum
    Prodigy 10 points

    Hi Michael,

    Thanks a lot for your quick reply! much appreciated!

    As you mentioned, we are not toggling it, just normal GPIO high/low control to drive a LED.

    Yes, I guess OE should be working well, so as to DIR1 and DIR2 I believe.

    We have 2 same EVMs, both have the same issue, where channel 1 doesn't work at all. (that means if we input a high (3.3V) in B side, A side seems always be 0V for channel 1, but other 7 channels seems working well).

    The setup we have is like below:

    - OE to GND.

    - DIR1 and DIR2 to GND.

    - VCCA and 2 VCCBs are all connected to 3.3V.

    BTW, from B side, what connected is Nvidia Jeston Orin NX carrier board with TXB01018 inside the carrier board. Or say the GPIO we controlled (connected to channel 1 and other 7 channels) are output from TXB0108. Not sure if this info is helpful.

    BTW, regarding directly driving a LED from A side of SN74AXC8T245, it seems working well without a protection cascaded resistor, but I'm not sure if this would produce other potential issues of SN74AXC8T245? As the voltage drop of the LED is kinda fixed (like 1.7~1.9V), if directly drive a LED from any channels in the A side  from 3.3V,  I'm not sure if SN74AXC8T245 will internally absorb the extra voltage drop (3.3V - 1.9V)? And if so, any potential damage would be caused? I didn't find internal circuit description of SN74AXC8T245, so hope could get some hint from you. Thanks a lot! 

  • 0 in reply to Owen Wang
    Guru 257000 points

    At 3.3 V, AXC outputs have a specified worst-case output impedance of about 58 Ω (see [FAQ] What is the output voltage (VOH or VOL) when the output current is X or the supply voltage is Y?); the typical value is roughly half of that. And LEDs (and other diodes) do not have a constant forward voltage.

    12 mA is the maximum current for which a voltage drop is guaranteed. You can try to draw higher currents (up to the absolute maximum rating of 50 mA), but then you do not know what the voltage drop might be. (The MOSFETs will eventually saturate.)

    On the EVM, A1 and B1 are also connected to the SMB connectors. It's possible that you have a short to GND somewhere, or that the pin got damaged by ESD.

  • 0 in reply to Clemens Ladisch
    Prodigy 10 points

    OK. Sorry, I'm still a little unclear what this sentence means "12 mA is the maximum current for which a voltage drop is guaranteed", could you please explain a little further? thanks a lot!

    If my understanding is correct, that means if we directly drive a LED from the GPIO, it could draw up to 50mA of current with inside-chipset 58Ohm impedance, right? if that's the case, is there any protection inside the chipset to make sure it will out damage the chipset? or say, when directly driving a LED, is it safe at least from the chipset side? thanks!

    Regarding the potential short to GND for A1 or B1, we will check further and let you know. Thanks! 

  • 0 in reply to Owen Wang
    Prodigy 10 points

    We've double checked, both A and B of channel 1 doesn't got shorted to GND. thanks!

  • 0 in reply to Owen Wang
    TI__Mastermind 25040 points

    Hi Owen,

    12 mA is the maximum spec characterized for the VOH / VOL levels listed in the data sheet. Also, no protection internally and beyond 50 mA is specified to damage device. 

    When you say both A and B are not shorted, do you mean SMB connector works instead?  I.E. ch1 did not  previously work, but ch8 worked. However both channels were configured similarly, thanks. 

    Best Regards,

    Michael. 

  • 0 in reply to Michael Ikwuyum
    Prodigy 10 points

    Thanks for the reply Michael!

    Good to know that there is no protection internally and >50mA may damage the device.

    When I said channel 1 of A and B are not shorted, I meant I measure the channel 1 pin of both A and B side, they are not shorted. I didn't check for SMB connector which we will not use. Yes, I also believe channel 1 and other channels should be configured the same way, but just not sure why just channel 1 doesnt' work for both of the 2 EVMs we have. Thanks!

    BR,

    Owen

  • 0 in reply to Owen Wang
    Prodigy 10 points

    BTW, do you have a EVM for SN74AVC16T245 (16 channel)? thanks!

  • 0 in reply to Owen Wang
    TI__Mastermind 25040 points

    Hi Owen,

    Unfortunately, we do not have 16 ch EVM.

    I would further recommend re-soldering a new good unit ensuring no ESD damage and also verifying the same setup works for all channels while toggling each channel for high and low with an input frequency as I suspect damages per the LED setup, thanks.

    Best Regards,

    Michael.