LM66100: ST signal state when power supply disconnected

Besson,

This behavior is correct. ST pin will be High-z if channel is conducting, so that output will pulled-up.

When 3V6_REM is disconnected, U23 is unpowered, which means that that ST pin is no longer controlled, and continues to be High-z, So it is not suitable to be use ST pin for your need, especially if you are only using 1x LM66100.

One possible solution is as follows. Use a comparator against VIN and 50% (or your selected level based on R ratio) of VOUT from a resistor ladder, for example, below.

You can power this COMP from 3V6_REM_OR.

There are other solutions which are mentioned in the datasheet which rely on using either a second LM66100, or an external FET to accomplish ORing of two supplies. If you used these, you would be able to rely on ST pin to relay channel status back to the MCU but since you are placing the supercap on OUTPUT these are not suitable for your application.

It is not the best idea to try to power LM6610 from the output-side, even through diodes in order to use the internal /CE comparator to change ST status.

doing this can causes reverse current which may be unacceptable to you.

However, If 3V6_REM line doesn't have any other circuit on it  which you wish to remain unpowered when the line is not present, you should be able to use this solution as the Reverse current case is not a possibility.

This can be accomplished with 1+ diodes fro VOUT to VIN as follows. This will be able to power-device off 3V6_REM_OR when 3V6_REM is not present, but because of the 1+ diode forward voltage drop, /CE internal comparator will change ST pin to LOW when 3V6_REM becomes floating. 

Like i said, external comparator solution appears to be the best solution for this case.

Please let me know if this is OK for you, and if you want to discuss further please dont hesistate to let me know here. If this answers your question, could you let me know by pressing the green button?

Best

DImitri

Hopefully this answers your question,

Besson,

Besson said:

But actually I tested the floating Vin of LM66100 (without pull-down resistor to GND) when battery is disconnected from it, it's still at 3.6V for a while after the disconnection. so how to understand this phenomenon?

From when CE>VIN threshold is reached, the delay from this to LM66100 turning off is ~2us as listed in d/s (toff). This does not include fall time.

VIN voltage may not drop immediately, especially if you have large capacitance on 3V6_REM How long are you seeing VIN dwell?

Besson said:

is that means the PMOS is still ON when Vin is disconnected?

PMOS could potentially be ON for a very short time. FOr esxample, if you have large cap on 3v6_rem it will take extra time to exhaust the charge there before the internal comparator triggers the device to turn off.

Your circuit can also work but expect the rise/fall to be slowed by your large pulldown resistance at "IO of MCU", and it will limit your VIH/VIL. Using the resistor divider on 3v6_rem_or along with comparator is what i prefer because that rail is always powered, so the resistors can be large to minimize quiescent current without impacting rise/fall time.

With a comparator, you dont even need the resistor divder at al actually.

Other option is to use a second LM66100 with supercap at VIN. This may or may not work for you on the systemlevel

Let me know if you have additional questions on this.

Best

Dimitri