E said:
My Rigol (Agilent-branded one) updates about as fast as the lcd can display,
but then it only have 4k memory.
Waveform update rate on oscilloscopes is not the same as display update
rate. The latter is not particularly relevant. Waveform update rate is
essentially acquisition rate, not the sample rate, but the rate at which
triggers may be responded to and result in acquisition records which map
to the display.
Some scope architescures, such as Agilent 6000 and 7000 series, can
update 100000 waveforms/s. This can be done even if the time to fill
the acquisition memory is greater than 10us. For instance, at 4Gs/s, it
takes 1 ms to fill 4MB of acquisition memory. But on a 1us/div
horizontal scale, the screen represents only 10us of time. The scope
can on a setting like this achieve about 50000 wfms/s.
The importance of this is that if you have an intermittent anomaly, such
as a glitch or runt in a signal that occurs at 1ppm on a very fast
signal such as a memory bus transacting at a MHz, for ex., then if your
waveform update rate is only 10wfms/s, you have only a 1/100000 chance
of finding it in one second. Ie, it will take 100000 seconds to ensure
100% probability of finding it. (I may be over simplifying the
probability arithmetic here.)
However, with a scope that can capture 100000 wfms/s, it will only take
10 seconds to find the glitch.
This is one of the things worth paying up for with higher-end scopes.
I just tested how small time difference my 100 Mhz Rigol can detect:
about 0.5ns. If I feed both channels same signal they differ about 0.2ns
But my pulse generator is manual one ie. striking wires into solderless
breadboard damn fast, and that generates bit ugly signals.
Would be interesting to know what the risetime of those pulses is.
I am pretty sure it is under 1 ns but how much, that I don't know.
Anyone with a real scope wanting to experiment?
Interesting, I would have expected much slower response with just a
probe and long ground wire. But I can get 3.5ns rise time plugging my
probe into a 5V breadboard rail, on a 100MHz TDS3014. Since the rise
time of a 100MHz system is 3.5ns, this suggests that the actual input at
the probe tip is relatively impulsive (ie, much faster even) compared to
the scope front end.
I will try this at work with a 500MHz scope. In a few more weeks, I can
try it with the 1GHz scope.
This is clearly convincing me that I may need to go to a full 1Ghz scope
for my hobby electronic projects ;-)