Who has experience with these things?
https://www.ti.com/lit/ds/symlink/txb0108.pdfAt a first glance, this sounds like a no-worries bridge between two devices of different I/O voltage levels, like 3V vs. 5V.
And one doesn't even need to care about directions.
So I got one device with a couple of outputs and inputs, and another one with corresponding inverse function I/Os to be connected to, so some of the "translation channels" on the same level shifter IC would go in one direction, the rest in the other.
And it just works.
Right?
Or are there caveats?
How does this auto-direction-sensing work? Can it be "confused" by suboptimal conditions at a corresponding pair of A/B pins?
Say I have a design where the B device is not always connected, like a PCB test rig. The A device (tester) has inputs which should not be floating when nothing is connected, like would be the case when the device under test is not connected and therefore I pull the OE pin down, to make everything hi-Z.
So I configure those inputs to have a "weak" internal pull-up.
Would that confuse the auto-direction sensing to output something on the counter pin? (which may not be good when, after that, the device to test gets plugged in)
I guess this could all be worked around by a proper sequence of applying power to stuff and toggling enable pins etc.
But it would be nice to know if there even is a potential problem, I guess.