# Mesure sensor voltage by 328p and OP Amp in parallel

I am designing a circuit to measure the temperature via a thermistor via the ADC pin on a 328p and I am looking for some advice. As a failover I am also using an OP Amp as a comparator, to see whether the voltage exceeds a certain threshold. My question is the following. What can I do to protect the analog voltage line of the thermistor from interferences from the 328p? Can I connect the thermistor out directly to the 328p and to the MCU? My ideas are the following:

• Using an in-line diode before the 328P adc pin, so that it can measure voltage but not interfere with it.
• Using an OP Amp voltage follower to provide the voltage to the 328p pin

Is there a better way to do so? Are my possibilities feasible or are there any problems? What are the advantages and disadvantages of one or the other solution?

I am thankful for any help or advice.

simulate this circuit – Schematic created using CircuitLab

• The ADC input is very high impedance, so unlikely to interfere with other measurements on the same pin. You don't need either the diode or the separate op-amp. – Boggyman Jun 14 at 12:19
• @Boggyman however if the output impedance of Thermistor and R1 is around 50k, and you are measuring also something else, there will be much bigger crosstalk than if you have lower output impedance (recommendation from datasheet is up to 10k). However if he's measuring single slowly changing signal, then it won't be affecting it much – KIIV Jun 15 at 7:09
• @KIIV I am quite new to this, R1 is 100k and I am measuring the voltage via 328p and the OpAmp. What exactly do you mean by crosstalk? Why should the impedance be lower? Hope I am understanding correctly that impedance is the measure of resistance in an AC circuit? – jack Jun 22 at 13:53
• @jack There is a sampling capacitor (few pF), if you are charging it through 100k resistor, it'll take much more time to charge/discharge it from different voltage level, than if you have few kiloohms. And if it takes more than sampling time, you'll be measuring nonsenses. Mainly if you are measuring something near 0V and something with high impedance near 5V, it'll be really far from it's real value. – KIIV Jun 22 at 14:00
• @KIIV thanks a lot for your explanation. I am sampling rather slowly on the 328p (one sample every 5s or so). Do you know where I can find the sampling rate for the OpAmp (LM324DT), I am that I have to ask, I am not even sure an OpAmp has a sampling rate – jack Jun 22 at 14:11