Original question:
My customer meets an issue when they using below design with TCA9517,
When transforming message , Zigzag waveform will occur on SDA pin. Please see below waveform.
Please refer below correct waveform,
And, if we connect SDA to GND directly, please refer to below waveform,
Please help to check why this happen? Any concern or problem , please let me know.
Thanks.
Hey April,
In your schematic, it looks like you have the B sides of the device connected to each other. This is not recommended in the datasheet:
I presume the problem is not present when the master is driving the signal however if one of the slaves drives the signal, then there are some prop delays between the two and the two devices are no longer in sync with each other. There comes a point where A1 drives B1 low which drives B2 low and drives A2 low but when A1 lets go and B1 releases B2 stays low because A2 needs to go above a certain voltage threshold before B2 goes high, this means that B2 then drives B1 low and drives A1 low... After this happens A2 goes high and releases B2 but now A1 needs to go high to release B1 but instead B1 drives B2 low.... It's a cycle. This is the reason why we do not recommend this topology in the datasheet. The difference in prop delays is the issue and this problem cannot be fixed without a board spin.
This design should work if you connect A to A side but not B to B side.
Thanks,
-Bobby
