Hi A.Fujinaka, My best guess as to why the customer sees improvement with increased S+H duration, but not with reduced RC is that the op-amp BW is the limiting factor. For your initial RC, 1MHz BW is adequate, but for the tests with smaller RC the op-amp BW is limiting according to the linked TI precision labs training. I've tabulated the RCs used in the experiment, their corresponding RC time constant, 1/4 the RC time constant (suggested op-amp tau), and the corresponding op-amp bandwidth 1/(2*pi*tau). The next good experiment would be to find a pin-compatible low-noise and high-bandwidth op-amp, substitute it on the board, and try some of the lower RC values again.
↧