Hi team,
Is there any concern if opamp buffer to GPIO is powered by 5V (which is exceeding 3.6V rating) to buffer a 2.5V DC voltage to C2000? In the case of opamp failure, or 2.5V failure, there is a possibility of having >3.6V at GPIO pin.
How about to use an external series resistor (say >2.5kohm) to limit any potential current injected to the pin <2mA? Is this acceptable?
Hi PY,
I've pasted some a quick TINA-TI simulation below.
In the first case, we only use a 1k external resistor to limit the current to the internal diodes. This does a good job (a little over 1mA flows into the device with 5V applied) but the overall series resistance is 1kohm, which is moderately high for driving an ADC input.
In the second case, an external set of diodes has been added (note: if you are only worried about over-voltage situation, you can omit the low-side diode to VSSA). R3 is added to limit the current as before, but it can be smaller because the external diodes are generally rated for much higher current. R2 is then added to steer current to the external diodes instead of the internal diodes. Overall, the series resistance is now only 200 ohms (and the current going into the device has been reduced as well).
Note that this is just a quick simulation to show that this circuit will reduce the overall series resistance. You'd want to spend a little more time engineering the circuit to ensure that (1) The clamping current to the external diodes and internal device diodes are always within specification (2) the amount of clamping current won't cause issues for your op-amp and (3) the components on the ADC input can allow for a S+H time that is acceptable for the application.
Also note that the control cards use some external diode packs that might be a starting place if you are looking for a way to effectively protect many ADC inputs.
