opAmp decoupling / voltage clipper?

[tester43]

Hi,

I am building a T12 soldering tip controller. It's a heater with thermocouple in series.

- Voltage given to tip will be controlled by mosfet (first prototype was triac controlled by moc3052)
- Voltage applied to tip will be rectified (not filtered) 24V
- I am heating using a whole cycle set because I am detecting z Zero in AC by h11aa1
- When I want to measure the TC then i am simply cutting the power to the heater
- I want to measure thermocouple using AD8495 op-amp

Issue is: i think that putting 24V on opAmp input pin when heater is on will destroy it.
Solution: when AC is on i would like to disable input connection of opAmp.
Doubts: How to disconnect input from heater?
I am thinking about CMOS switch or transistor (mosfet?).
CMOS switch sounds like faulty element because it's a mechanical solution after all.
I am not sure if mosfet will work.

Do you guys have any concept how to do it correctly?

Kind regards,

[capt bullshot]

This amplifier has a built-in over voltage protection.
Just use a series resistor and a suitable clamping device (zener diodes, diodes clamping to GND / VCC, ...) to ensure the voltage at its input doesn't exceed these limits.

[tester43]

Hi,

thank you - and yes, I read about the over voltage protection - but it will fail around 20-24V according to datasheet
I know about zener diode, but here is my second question: if I have a heater working, and when heater voltage gets above zener diode threshold then diode will start dumping energy to GND. I understand that's why the resistor in series before zener and input pin.
BUT: this resistor would be in size of several kOhms - am I correct?

[capt bullshot]

As a general rule, you'd choose the resistor value as low as possible, to carry a current of 1mA ... 10mA while the heater is on. Larger values reduce current and dissipation, but may result in larger measurement errors due to input bias and offset currents of the amplifier. As these appear quite low, I wouldn't be concerned too much with a e.g. 10k or 22k resistor. But do some simpe calculations, how much voltage results from the amplifiers input current across the resistor, a systematic error of 1° ... 5°C (10uV ... 100uV in the world of thermocouples) would be acceptable here. This error is smaller with lower resistor values. Also watch for the clamping devices leakage current (e.g. if you clamp with diode to VCC, a zener to GND, assuming the other end of the thermocouple is at GND, would be better).

[tester43]

So..... resistor seems to be 4.7kOhm or the next one.
Circuit would be: 4.7kOhm resistor -> zenner -> input of opAmp
with zener diode polarised to send everything over a threshold voltage to GND.
Resistor will limit the losses to 5mA in peak of 24V AC.

I just had an "eureka moment" - opAmp measures using bias current coming from input pin (correct me if wrong please). If this current in "pico" level then I can safely ignore any additional resistor because the signal (voltage) loss will be minimal - is it correct?

[capt bullshot]

I just had an "eureka moment" - opAmp measures using bias current coming from input pin (correct me if wrong please). If this current in "pico" level then I can safely ignore any additional resistor because the signal (voltage) loss will be minimal - is it correct?

Yes, it's about that. Still do the calculation, at least the quick and dirty way (didn't read the datasheet again to get the exact figures, so the result I present will be most probably wrong), and use the worst case values:
>> 50pA input current plus 20pA bias current -> 70pA -> 70pA * 4k7 -> 329nV -> OK

[tester43]

thank you - i will build it over the weekend and report results when my ad8495 ic arrives.