Please check this PDF https://www.electronicsworld.co.uk/wp-content/uploads/2019/04/2019ewApr-1.pdf.
I really can't understand the need for 12 V supply connected with reverse bias to D2 and D3. Without it also, the op-amp would saturate to 5 V (on current or voltage > than 20 mA or 5 V respectively).
The authors note is:
Dual series Schottky barrier diodes D2 and D3 divert any overcurrent to power supply or ground. LM358P-A
I think I am missing something here :(
The idea is that if a voltage less than 0V is applied to the input then D3 will conduct and (assuming some source resistance) keep the voltage at the non-inverting input from going so negative that a lot of current flows through the LM358 input to V- pin.
Similarly, D2 would conduct and the 12V supply might sink some current.
In fact this is an amateurish design- D2 will typically do nothing of the kind (because most power supplies can’t sink much current) and both diodes will be quite leaky. D2 is also unnecessary since the input of the LM358 can handle up to 32V even with Vcc=0, and there is an 18V TVS. D3 would be better handled by other means so as not to affect the accuracy (the 1n5819 can leak hundreds of uA up to 10mA depending on temperature and reverse voltage). Just adding a resistor to the input may be sufficient.
