TCA9617A: Overshoot and Absolute Maximum Rating

Hey Kaedo-San,

"Could you tell me about SCL and SDA's Overshoot?"
You will likely be more concerned about the undershoot which occurs when you pull low on either SCL or SDA. This is due to the quick change in current over time due to a strong pull down. Any parasitic inductance will also worsen this affect. This event is known as inductive kickback.

"If it is possible that Overshoot will exceed the absolute maximum rating, how long can TCA9617A tolerate it?
(ex. Allow -0.6 V for 10 ns)"
-I don't believe these values were tested for a specific amount of time so I do not know (I assume they simply put a negative voltage on the pin for a long time and then test if it works if it doesn't break they increase the voltage). I believe if you violate these max ratings over a long frequency that it could also result in damage. (example: the signal goes -0.6 for 5 ns but the device could break after 10000 times).

"It is not a guaranteed value but a reference value, so could you tell me?"
-My understand here is that a large number of these devices were tested and the break point was found in which case the lowest minimum number was used (the earliest point where a device first broke). We typically leave some room for margin but you should treat the number we provide as an absolute. You may find some break at -1V while others could break earlier at -0.6V.
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------
A solution to clamp the negative swing is to use a schottky diode on the SDA/SCL line tied to GND.

Thanks,
-Bobby