Suggestion for discrete isolator between input and MCU

[Freshman]

I have the below attached circuit where, the input voltage is voltage divided and given to the ADC input of the MCU. MCU is powered from 3.3V.
3.3V power for the MCU is also obtained by stepping down VIN to 3.3V using buck converter.

Before the 3.3V comes to the MCU, there might a leakage from this voltage divided VIN and through the ADC input pin of the MCU, through it's ESD diode and to the MCU's 3.3V rail.

Can anyone suggest a simple isolation technique to avoid the above?

[ggchab]

Do you have to measure Vin continuously ? If not, why not using lower resistors values and enabling the divider (with a MOSFET controlled by the MCU) only when you measure Vin ?

[Freshman]

Thank you @ggchab for your reply. I need to measure VIN continously.

Can you suggest the circuit?

[ggchab]

Can't you use an op-amp as voltage buffer ? (https://en.wikipedia.org/wiki/Buffer_amplifier)

[Freshman]

thank you @ggchab. So, I have checked. I will use MIC6211 op-amp. Will connect the voltage divided input to the non-inverting terminal of the op-amp and the inverting terminal will be shorted to the output.

I have 2 questions.

1. How will this prevent the issue that I described earlier?
2. What power rail should I connect my VDD and VSS of the op-amp. 3.3V only right?

[ggchab]

The leakage current can change the voltage at the output of your resistive divider. The ADC may also not work very well if the impedance of the source is too high. Adding the buffer is equivalent to lowering the impedance of the resistive divider. So, the leakage current has almost no influence on the voltage and the ADC can also work better.
You could also reduce the resistors of the voltage divider but it will draw more current !

[PGPG]

Before the 3.3V comes to the MCU, there might a leakage from this voltage divided VIN and through the ADC input pin of the MCU, through it's ESD diode and to the MCU's 3.3V rail.

I'm using solution like your since years without any extra isolation.

1.
Why you get such high voltage at output of your divider?
I was using only 2 MCU families and both had internal voltage references for their ADC in a range of 1V that are better references than using VCC as reference.
So my divider is 100k+(3k3 || 1n) to be measured in 0..1V range for VIN up to 30V.

2.
Even you get at divider output voltage comparable with 3V3 VCC I would not be afraid of powering MCU by 470k resistor.

3.
I understand that VCC comes from the same VIN.
The voltage at SEN_ADC11 rises with time constant around 10ms. Do the step down is not faster than this?

[Freshman]

@ggchab, i do not understand your point.

1. Will my op-amp connections work?
2. What power rail should I connect my VDD and VSS of the op-amp. 3.3V only right?

[ledtester]

Before the 3.3V comes to the MCU, there might a leakage from this voltage divided VIN and through the ADC input pin of the MCU, through it's ESD diode and to the MCU's 3.3V rail.

Can anyone suggest a simple isolation technique to avoid the above?

With R1 = 470K, Vin needs to be more than 470V in order for 1mA to flow through the ESD diode. Will that be a problem?

Your max sensing voltage of 18V suggests that the ESD diode can handle 15mA before burning up - is that correct?