Back to back diodes at inputs of opamps

[Mark Johnson]

But how about non-linear junction capacitances?

DataBOOK says 5pF at Vreverse=0 and 1.6pF at Vreverse=10 . Capacitance Change = (5 - 1.6)/1.6 = 213 percent

Now suppose you connect a 180pF ceramic capacitor with NP0/C0G dielectric, in parallel with the diode.

Now it's (180+5 = 185pF) at Vreverse=0 and (180+1.6 = 181.6pF) at Vreverse=10. Change = (185 - 181.6)/181.6 = 1.9 percent

You've reduced the capacitance change by 41dB.

 

[Kirchhoff]

Now suppose you connect a 180pF ceramic capacitor with NP0/C0G dielectric, in parallel with the diode.

Assuming we're still talking about the diodes across op amp inputs, TI has some good assessments of the outcome:

http://www.ti.com/lit/pdf/slyt087

http://www.ti.com/lit/an/sloa013a/sloa013a.pdf