At the beginning of this project, we started by looking at various ways traditional scopes have been implemented. We immediately realized that attenuating the signal and connecting it directly to an ADC was a bad idea. The back of the envelope calculation showed us that in order to maintain the kind of dynamic range and precision that we wanted, we would need to get a high resolution ADC. As we expected, high resolution, high sample rate ADCs were extremely expensive and therefore not viable for this project.
After looking through old oscilloscope datasheets and consulting with Dr. Richard, we determined that we could implement our design using a fixed gain attenuator, a Variable Gain Amplifier (VGA), and a 10-14 bit ADC. We also knew that we should to implement an anti-aliasing filter to remove high frequency components before they get sampled by the ADC. We judged that we could use a 2 or 3 pole filter in order to remove most of the unwanted spectral components. This assumption was largely predicated on the fact that we assume the user will use the device ideally and not try to measure high frequency signals.
After picking out our parts, we ordered them as shown in the figure below. Ideally, we would have liked to place the VGA as close to the signal source as possible in order to prevent extra noise from being added before amplification. This would help to keep the SNR of the system as high as possible. However, due to the differences in signaling between chips, it made the most sense to order them as shown.