ISO1540: Vol for ISO1540

Hi Koteshwar,

Because this is PoE pplication, we still need to use the isolator for the I2C bus. but signal quality is also what we will concern.

From the MCU spec the max VIL is 0.8V. I think this is fine.
However, I see ISO1540 datasheet and discover the following things,

1. When the signal is from side 1 to side 2, there is 0.4V step (Figure 20 & 33). When we measurement the signal quality like this, customer will ask how come the waveform is so strange. Do you have any suggestions which we can explan to our customer?
I also check with other vendors, Silicon lab Si8600AC-B-IS, looks like this step is exist on their part.

2. When signal is from side 2 to side 1, the Vol2 is 0.4V (Maximum, page 9 section 6.9), I checked other vendors' part, like Silicon lab and ADI, all of them cannot actually output to 0V. But we really want to know is this the defect for the I2C isolators? How is this caused to explain to our customers.


By the way, would you please also suggest which side is to connected to MCU (I2C Master) and which side is to connected to I2C slaves for ISO1540. Thank you!


Thanks,
Steve

Hi Steve,

Thank you for sharing additional details about your application and the requirement, it helps me understand the situation better.

Glad to know that MCU has a max VIL spec of 0.8V.

Since I2C isolator requires bidirectional channels, an isolator needs to use two unidirectional channels and connect them back-to-back to achieve a single bidirectional channel. Connecting two unidirectional channels back-to-back directly would lead to a lockout situation of both channels in LOW. Hence to avoid such a lockout situation, a diode has been introduced at the output of Side1 to make the LOW output of output channel on Side1 looks like a HIGH for the input channel of Side1 thereby avoid the lockout loop creation. The diode placement in shown in the block diagram below.

The thresholds are carefully designed to make sure the bidirectional channel operates smoothly as long as the VOL specs are acceptable and VIL specs are met for ISO154x. I understand that it is not desirable to have a step on the port at Side1 but please note that this is perfectly an expected behavior and the voltage levels are in acceptable levels. I hope explaining why there is this step at Side1 might help your customer to consider the behavior acceptable.

Like you have already observed with other vendors, this has been the common practice in the industry for achieving bidirectional function. Now to answer the questions you have listed,

1. The step output cannot be eliminated as this is a necessity for the bidirectional operation of the channels.

2. Since the low-side transistor on the output is going to have a finite voltage drop [V(CE)] across it, a true 0V output is difficult to achieve. This is not just with the I2C isolator, most digital channels in isolators, MCUs, ADCs and others will have a finite output voltage.

3. Side1 of ISO1540 is meant to be connected to the I2C node (master or slave) which needs isolation while SIde2 should be connected to the I2C bus.

Let me know if any of the above points are not clear to you, thank you.

Regards,

Koteshwar Rao