A SAR ADC generates voltages through a DAC and then compares that analog voltage to the input voltage. This search takes some amount of time. If the input voltage changes while the search happens, then the search would fall apart. A "sample and hold" amplifier (S/H) holds a copy of the input voltage while the binary search happens. An S/H is a special op-amp circuit that charges up a capacitor and maintains the voltage until it samples again.
S/H and Comparator Schematic
Back when I was waiting for the DAC boards to come back from the fab, I built some initial breadboard circuits. For documentation, I built these circuits on a breadboard and then captured them in KiCad. The LF398 is set but the comparator will likely change when moving to perma-board.
Sample and Hold Amplifier
The sample and hold amplifier is based on an old Texas Instruments LF398 IC. While available in an 8-pin DIP package, it also came in a TO-99. While I do not plan to use the TO-99 in the design, I had to grab one because they look cool.
One down-side to this IC, at least from my perspective, is that it really wants to have a negative voltage supply. The output is limited to about 1.4 volts from the plus and minus rails. For the positive rail, that wouldn't be a problem. But without a negative supply, the output would never go below 1.0 volt.
On the other hand, this circuit was dead simple to use. It only requires the positive and negative supplies, a digital signal for the sample control, and a capacitor to hold the sampled voltage. In operation, you pulse the hold pin, and the IC takes care of the rest. When my friend showed me the IC, I immediately looked at the datasheet to see how much it would drift over time. The datasheet looked promising. And with testing, I found I a 470 nF film capacitor would hold the sampled voltage for at least 30 seconds while only losing about 100 mV.
Despite having a wall of components with boxes that literally describe everything available from a distributor and talking about the essential active components to have on-hand, I am still running into parts I do not have. For example, I do not have a dedicated comparator IC. So, I built a simple comparator with a generic op-amp for the breadboard design, the LM358. There are many negatives to use a generic op-amp, but it would be fine in a circuit like this one. That said, I did pick up some LM339s for the next iteration.
An obvious next step is to build this circuit on something more permanent than a breadboard. After that, the circuit is ready for integration with the DAC bits. It is shocking to me how fast this project went from a crazy idea to parts in progress and partially functional blocks.