Since the maximum sample rate is a limiting factor in any Digital Storage Oscilloscope, the shortest time it takes to get an acquisition consisting of a number of samples is going to be limited by the maximum sample rate. Shown below is an example to illustrate this fact.
Using an acquisition memory length of 500 points on an oscilloscope with a maximum sample rate of 100 MS/s, the time taken to take a sample is 10 ns. Therefore the time taken for the entire acquisition, 500 points, is 5 µs. Since one acquisition is ten divisions long then the maximum single-shot timebase must be 500 ns/Div. If the acquisition were now to be set to 50,000 points, the time taken for one acquisition would be 500 µs, giving a fastest single-shot timebase of 50 µs/Div. So the fastest timebase is lower for longer memory lengths even though the sample rate is still the same.
However, since we have the facility for ETS on our instruments (see previous question) we can use this to produce timebases which, on the face of it, require a faster sample rate than the maximum single-shot sample rate possible on this instrument.
Many DSO manufacturers use a very inexpensive technology to achieve higher sample rates. This approach limits memory depth, thus reducing the amount of time a signal can be acquired. Doing this forces the user to compromise the acquisition duration, the information before and after an aberration, and focus only on the small of the signal. Gould DSO's offer a minimal 200 Ksample memory. Some models are expandable to 1 Mword per channel!