TXB0104: OE ramp voltage

Hello Matsumoto-san,

If my understanding is correct, then they would like to keep the outputs disabled for a short time on startup, after which it will always be enabled.

This is a common application, but it is not a good idea to have slow rising inputs into a CMOS input, which the OE pin is. 

We would recommend putting a single schmitt trigger device connected at the input to ensure that there is a sharp edge at the OE pin after some time.

I would recommend looking at the SN74LVC1G17 device.

Best,
Michael

Hello Matsumoto-san,

Please see my response below.

<Question1>
This is a common application, but it is not a good idea to have slow rising inputs into a CMOS input, which the OE pin is.
->Is there regulation for OE's rising time?
   Could you let us know "it is not a good idea to have slow rising inputs into a CMOS input"?

Generally non-Schmitt trigger CMOS inputs require a sharp edge to prevent oscillations/increased current consumption. Generally this is specified by a delta(t)/delta(V) in the datasheet and it usually is 20ns/V or 100ns/V. For this part there is a transition rate specification for the A and B, but not the OE input. I would follow the same rise time recommendation on the OE input as the A and B ports of 40ns/V. An RC curve will not meet this specification, so a schmitt trigger buffer device is required to clean up the signal to make a sharp edge.

<Question2>
We would recommend putting a single schmitt trigger device connected at the input to ensure that there is a sharp edge at the OE pin after some time.
->Does it mean a single schmitt trigger device(like as SN74LVC1G17) connected to A1 to A4 of TXB0104? or TXB0104's OE pin?

No need for the schmitt trigger to the A and B ports of the device. This just applies to the OE pin since you have an RC circuit that will cause a slow rising input into only the OE pin. This schmitt trigger buffer will only be required for the OE pin. 

Best,
Michael