When trying to find the "best" scope for the money, I stumble over all sort of technical data as sample rade, memory depth, bandwidth and a lot more.
350Mhz 14M wms/s or 200Mhz 140M wms/s and so on, what is important and why (MHz - GSa/s - wfm/s - Mpts or
:-)
Define the money first! There is no way in the world that a $350 Rigol DS1054Z is on the same planet with a $200k Keysight. If you aren't in the market for the Keysight, what does it matter how good it specs out?
Bandwidth is up to you. What do you want to see? If it is a sine wave at 100 MHz, then a 100 MHz scope will do. If it is a square wave at 20 MHz and you think seeing only up to the 5th harmonic is good enough, a 100 MHz scope will do. Remember, that square wave has harmonics from DC to daylight. How many need to be displayed? What happens is that the square wave looks more like a sine wave as the frequency is increased and the ability to display a lot of harmonics is decreased.
Remember the number 2.5, there will be a test later. The industry standard seems to be to sample at 2.5 times the maximum signal frequency - minimum. Technically, 2 would get it done under ideal circumstances but the industry has chosen 2.5
DS1054Z - 4 channels times 100 MHz times 2.5 => 1 GHz sample rate. It all lines up! If all of the channels aren't in use, the 2.5 number can be increased. 1 channel times 100 MHz times 10.0 => 1 GSps on just one channel. Sampling at 10x will produce a better display than sampling at 2.5x but what you see on the screen is really just a drug induced hallucination in any event. There's a lot of interpolation going on down inside that FPGA. On some scopes you can just display the samples as dots without a line and magic interpolation. I don't know what use that is.
Page 10 here for record length and waveforms per second (neither of those specs are particularly important to me):
http://ecee.colorado.edu/~mcclurel/txyzscopes.pdfClearly, more of anything is better than less. The real questions are "Does it matter?" and "Can I afford it?".
There is no such thing as 'best for the money' if money is also a variable. It's better to drive a stake in the ground ("I won't spend more than $xxx, no matter what!") and work back from there. A more expensive scope will almost always be a better scope.
For entry level scopes, the DS1054Z and the Siglent SDS1104X-E seem to be popular. The Siglent is newer and the user interface is said to be much faster than Rigol's. If I didn't already have the DS1054Z, I would be looking at the SDS1204X-E 200 MHz, 4 Channel. Or unlocking a SDS1104X-E to get the bandwidth. See the "Test Equipment" forum here on EEVblog.