[sdiy] AC coupling caps on MS20 clone

[David G. Dixon]

> If you want switch in front panel, this solution implies connect a cable
> in a high impedance part -the inverting opamp input- when ideal design pcb
> wants short traces at that points.
> I don't see the need.
> 
> It would be better place the capacitor on the input or between two stages.
> So I would say between the sum inverting amplifier and the 30K input
> ssm2164 resistor. (I'm supposing the schematic)

I believe that the impedance at the inverting input of an inverting opamp is
R1 || R2 (where R1 is the input resistor and R2 is the feedback resistor).
Hence, if a blocking capacitor C is inserted between R1 and the inverting
input, then this impedance becomes (-j/wC + R1) || R2 which, at DC (w = 0),
becomes simply R2.

If this were a non-inverting opamp, then you'd be spot on.  (Indeed, that's
the issue with a sample-and-hold cap-and-follower, where the high-impedance
trace should be kept as short as possible, and even surrounded with a guard
trace).  In this case, a resistor to ground would be required to shunt off
any current, and the impedance would then be that of the shunt resistor in
parallel with the cap.

You are confusing the impedance into the inverting input (which is very
large indeed) with the impedance of the trace connected to the inverting
input (which is a parallel combination of all of the impedances leading away
from the trace: the input resistor (or cap), the feedback resistor, and the
inverting input (which can be ignored because it is so large compared to the
other two).

(If I'm smoking crack here, then someone please correct me!)