• TI thinks resolved

ISO1541: ISO1541 SDA1 output is not Zero level

Part Number: ISO1541


In customer’s test, when SDA1 is output, the output low level rail is not zero, about 0.7V.

Checked datasheet, there is a diode between the MOSFET source  and GND, so this 0.7V is the diode voltage drop.

Now, customer concern that 0.7V is around the low level threshold of CPU, so the I2C communication maybe wrong in some cases.


Why connect this diode? Why not connect MOSFET source to GND directly?

  • Hi Max,

    Thanks for posting on E2E. The explanation for this is posted in Section 8.4 of the datasheet.  (pages 19 and 20). For convenience, I am quoting the relevant section directly below. 


    Isolator Functional Principle (continued)

    At first sight, the arrangement of the internal buffers suggests a closed signal loop that is prone to latch-up. However, this loop is broken by implementing an output buffer (B) whose output low-level is raised by a diode drop to approximately 0.75 V, and the input buffer (C) that consists of a comparator with defined hysteresis. The comparator’s upper and lower input thresholds then distinguish between the proper low-potential of 0.4 V (maximum) driven directly by SDA1 and the buffered output low-level of B.


    Please let us know if you need additional clarifications.