Today I noticed that when I center the “Gnd Pos” on the screen, and then use + or - half-near-scale inputs to set the gain (gain adjustment is averaged between the applied + voltage and the applied - voltage), then the offset doesn’t vary with the gain setting adjustments. The “Measurement Vavg” was my measuring tool
I know this is true for 5V/Div where I used + & - 9V to set the gain (with offset=0) (Cal Gain = .43%) and then I used +36.62V to check near full scale accuracy with “Gnd Pos” set to the bottom of the screen. The Nano measurement was 36.3V with a full scale accuracy of 99.19% which is quite satisfactory for me.
Perhaps others would try this procedure on their Nano (using all scales) and report the universal effectiveness of this procedure.