Other Parts Discussed in Thread: TXS0102
I am using the TXS0102 for a 400KHz I2C bus level translation from a 3.3v master (microprocessor, single master) to 1.8v slave devices. Occasionally, the whole I2C bus gets stuck, where the 3.3v I2C-clock is stuck at approx. 1.0v (DC), as is the 1.8v clock side also.
We suspect that the active pull-up of the TXS0102 (approx. 50 ohms) is stuck ON, fighting the pull-down of the master trying to drive the clock low - resulting is approx. 46mA of current (3.3v - 1.0v = 2.3v / 50 ohms = 46mA). At that point, our microprocessor crashes and the clock is stuck "low" (1 volt) permanently (will we reboot the system).
Not sure if it's possible for the TXS0102 one-shot to be stuck on (?) Maybe there are other explanations ?
Disabling the OE on the TXS0102 makes the system work (for the 3.3v I2C devices only). Enabling the TXS0102 results in this behavior. All the 1.8b I2C devices are slaves (3 loads on a small board).
