Hi,
I'm having some trouble with my I2C bus, especially (but not necessarily the fault of the TCA) when the TCA4311A is connecting two I2C busses.
The I2C bus is as follows:
I2C Master (Analog BF518) protected with 33R0 series-R, 3.3V, 1200 Ohm Pullups
- quite a few 3.3V devices (10-15)
- TCA9406 3.3 to 5 V I2C level converter + two devices (5k pullups on 5V)
- TCA4311A with currently no device (2 are meant to be plugged in), in a partial-powerdown domain (1200 to 8k pullups have been tried) (dubbed TCA Nr 1)
- TCA4311A with one device, in a partial-powerdown domain (5.6k pullups) (dubbed TCA Nr 2)
So it's pretty much one master with a 3.3 to 5V converter with a few devices, and then two seperate power domains (which are usually powered down) seperated with TCA4311A with one / two devices each. All of that on a small 6x9 cm PCB.
The problem is now that the bus is pretty unstable/unusable - the main 3.3V pullups need to be very small (1200 Ohm works), the series-R's (one wouldn't have thought) are critical for bus operation.
Now, when TCA 1.) gets powered on (enable not active) everything is still ok. As soon as enable gets released (I2C busses merge) the bus is pretty much dead - the SCL pulses seem to get truncated, as shown in the scope-"screenshot" below.
TCA 2.) can be switched on, however, instead of gaining the device behind there, I'm loosing one of the main 3.3V devices.
We've tried to vary the pullups, without any success so far.
I'd be happy for any suggestions or comments on this!
Thanks,
Kai
