The sample rate speed of the scope's analog to digital converter (ADC) is a very important specification. It is the minimum time between each sample. For instance, a 500 MS/s. sample rate relates to 2 ns. per point resolution. Multiply the number of sample points * the sample period to determine an oscilloscope's maximum recording time @ maximum sample rate.
Modern scopes can sample at 20 to 60 GS/s on each channel. 20 measurements every billionth of a second puts a demand on memory. One (1) million samples is only 50 usec of recording time. Short memory scopes can only sample at maximum rates for short periods of time.
Application Timebase Useable BW Useable BW
32k Mem 8M Memory
=================================================================
Ethernet 10baseT 10 us/div 2.5 MHz 500 MHz
PLL Dynamic Resp. 100 us/div 250 kHz 500 MHz
CDMA-Qualcom 200 us/div 125 kHz 250 MHz
Video and Imaging 2 ms/div 12.5 kHz 25 MHz
Switcher Power 100 ms/div 250 Hz 625 kHz
Is the scope sampling fast enough in real time (single shot) mode on all channels so we can record without aliasing?
Ideally, the sample rate setting should be displayed on screen all the time.
Is the scope acquiring in real-time or equivalent time sample mode?
Real-time refers to single shot. "RIS" or "ET" refers to "random interleaved sampling" or "equivalent time" and can only be used with repetitive waveforms.
The vast majority of digital scopes have 8 bit ADCs. This refers to full scale resolution. An 8 bit digitizer will divide full scale input voltages by 255 counts. Thus the minimum discernible sampled value on a 1 volt full scale 8 bit oscilloscope would be 1 / 256 or .00390 volts per step.
If we have repetitive waveforms, we can increase your vertical resolution with averaging. Remember resolution isn't accuracy.
If we have transient waveforms, we can increase our vertical resolution by low-pass filtering each sweep with FIR filtering for enhanced resolution.
Not all ADCs are created equally. The most common figure of merit is Effective Bits that relate the number of correct bits of a given ADC's actual measurements vs.ideal.