- We can either digitize V-V0 or the average of this quantity over ~ 0.1 s.
- Instead of digitizing a signal, we have to digitize a change in a signal. Thus, some differential input is required. Further, because of the high magnitude of the voltage, we will need to de-amplify the signal in the process of subtracting V0.
- Dynamic range is a factor of 200,000 fg/3 fg = 66,666, which is almost exactly 16 bits. To have some error margin, an 18 bit converter would be desirable. Either a successive approximations 18 bit or (less expensive) a Σ-Δ would work well.
- There is already a feedback loop trying to keep current constant. Why not measure current, digitally decide what potential is needed to drive the current to the desired value, and then use an amplifed DAC output to vary V to force i to the desired value? We simply store the set of DAC values, and that's the desired data record. Asset: no high voltage digital electronics. Liability: we need to feed back fast enough that we have a high fidelity rendering of the peak shapes, and we need a high voltage stage after the DAC.