SYSTEMRESOLUTION (overall)= SensorResolution X LensResolution / (SensorResolution + LensResolution)
Sometimes it is shown as:
1/SYSTEMRESOLUTION (overall) = 1/SensorResolution + 1/LensResolution
Speaking as a physicist, that formula looks wrong to me.
I believe that the problem is equivalent to combining two sources of error on a single measurement, which is a very common occurrence in physics. The correct method of calculating the combined error is to add the individual errors *in quadrature*:
(1/SYSTEMRESOLUTION)^2 = (1/SensorResolution)^2 + (1/LensResolution)^2
So, if the lens and sensor both have an individual resolution of 50 lppm, you ought to get a combined resolution of 1/sqrt((1/50)^2+(1/50)^2), which equals 35 lppm.
S
