[sdiy] differential subtractor vs summer with an inverted input

[David G. Dixon]

> > In other words, given E1 and E2, I want:
> >
> > EO = G(E1 - E2/4)
> 
> A differential amp with an attenuator/pad on the E2 input?  One op-
> amp, handful of resistors.

Yes, that would basically be my Option 1.  The two resistors on the + input
make an 80% divider for E2, and the - input and feedback resistors set the
gain at -25% for E1 and 125% for E2 (thus giving 100% overall).  The
voltages sources must also be buffered, however, because neither input will
have high input impedance in this configuration, and in this case (which I
neglected to mention before) the sources are two separate R/2R ladders, so
they must be buffered.  Finally, because of the constraint on resistor
values for the differential subtractor, it is not feasible to apply the gain
at that stage (I tried, but some of the resistors would have to be
negative), so a separate gain stage would be required.  That brings the
total opamp count to four for either Option.  I think the "invert one and
sum algebraically" option is better.