This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

ISO1641: Question regarding VOL output level of Side 1

Part Number: ISO1641
Other Parts Discussed in Thread: INA237, , ISO1640, TCA9803

Hi,

We are using INA237 with ISO1641 in our application. Both sides are operating at 3.3V. The side 2 of ISO1641 is connected to INA237 and this part is fine. However, the side 1 of ISO1641 is connected to the I2C master, which is a (non-TI) processor. 

The concern is about SDA1 VOL level, which is about 0.57V to 0.71V. Since the processor VIL is 0.66V max. (0.2 x Vcc), we don't have enough margin. The pull-up resistor on the I2C bus is 2.2K. I could increase the pull-up resistor further, but I think the best case margin here would be 0.66V - 0.57V = 0.09V, which may especially not be enough at extreme operating temperature range. 

Is there a way we can resolve this issue without introducing any additional part in between?

Any suggestions would be appreciated.

Regards,

Kaitav 

  • Hello Kaitav, 

    Thank you for reaching out. 

    Side 1 and side 2 have different VoL and VoH specs, this difference is needed for the bidirectional functionality of the transceiver. Also, side 1 of the isolator is designed to be connected to the I2C master and side to is usually the bus side. If possible, best solution would be to switch the isolator around, and connect side 1 to the I2C master. 

    Best,
    Andrew

  • Hi Andrew,

    Thank you for your reply and suggestion. Unfortunately, we cannot reverse the isolator sides, as it would then create the same issue with the INA237 part. I assume using ISO1640 (bidirectional SCL and SDA) isn't a solution either as it would then cause VOL issues for both SCL and SDA signals. 

    I was thinking about the possibility of adding TCA9803 level translator on the Side 1 of ISO1641 (so, between ISO1641 and processor), but we have one more device on the same I2C bus (a TI temperature sensor part), so since I would have to use B side of TCA9803 for the processor side (since B side cannot work with Side 1 of ISO1640 due to the voltage offset; section 10.1.2.2 of TCA9803 datasheet), which cannot have external pull-up resistors, it would be a problem to have the temperature sensor on this side of the bus. In this case, one suggestion I received from our local FAE is to move the temperature sensor part to the A side of TCA9803, together with ISO1641, and have B side connected only to the processor. I am not sure if this would be practically feasible for us, but I am thinking further on the same. 

    Any further suggestions would be appreciated.

    Regards,

    Kaitav

  • Hello Kaitav, 

    I think using a level translator would be your best bet. Unfortunately, the voltage difference is a common feature that is the result of the bidirectional function and is common in I2C buffers and isolators. 

    Also, side 1 of the isolator is designed to be connected to the I2C master and side to is usually the bus side. If possible, best solution would be to switch the isolator around, and connect side 1 to the I2C master. 

    To clarify here, the concern is that multiple connections to side 1 will exceed the capacitive load recommended in table 6.3 Recommended Operating Conditions (side 1 has a lower recommended capacitive load). The FAE's suggestion is a possible solution as long as the capacitance is not too high for either side. Multiple I2C nodes will add bus capacitance, the isolator will be able to drive the bus as per the datasheet values as long as the load capacitance is kept below these recommendations. 

    Best,
    Andrew

  • Hi Andrew,

    Okay, I understand. 

    Thanks for the suggestion. I have thought about the bus capacitance and since we only have two devices (with 10 pF load for each), the total bus capacitance on the side 1 of ISO1641 will be 20 pF (the side 2 only has one device, so no problem). For TCA9803 also there's no problem since there's no such limit mentioned for bus capacitance on A side and for B side we only have the processor, so it is expected to be around 10 pF. 

    I will accept your previous post as a solution in a couple of days, just in case I have any further queries during this time. 

    Regards,

    Kaitav 

  • Hi Kaitav, 

    Thank you for letting me know. I will go ahead and close this post for now. Feel free to reply to the post to open it again. 

    Best,
    Andrew