Hi EEVBlog Community,
I've a question concerning an oscilloscope and if it fits my needs. I'm not a professional in electronics (and light years away from Dave's skills) but I "experiment" with AVR µc's and can build basic electronic circuits but mostly digital stuff (and on an Arduino level if you know what I mean). So I think that I'm an advanced beginner. For my newest project I need some kind of oscilloscope since I need to visualize/inspect/debug data on a data bus. The data bus has a clock with 10 MHz and runs also with almost 25 MHz (24-odd-MHz).
Therefore I searched for a reasonable scope and found the "Analog Discovery 2" in a video of Dave (actually v1 is mentioned in the video) and I quite like it: it's cheap (for me, as a student even more) and the software seems to be quite nice. Here you'll find the reference manual. The important specs are:
- Two-channel oscilloscope (1M?, ±25V, differential, 14-bit, 100Msample/sec, 30MHz+ bandwidth - with the Analog Discovery BNC Adapter Board)
- 16-channel digital logic analyzer (3.3V CMOS, 100Msample/sec)
My questions are: 
(1) Can I calculate this way:
If I have a clock at 25 MHz the signal is toggled every 40ns. 100Msample/sec means 100,000,000 sample points per second meaning that every 10ns a sample is acquired. Therefore we get 4 samples per "signal" state. For the 10 MHz signal I'll get 10 samples. But from this I can't really tell if this is good or not.
A good experiment is far better than a bunch of numbers. I just used my AD (original) to sample a 4 signal bus at 25 MHz, 12.5 MHz, 6.25 MHz and 3.125 MHz. The display shows 4000 samples at 100 MHz which would seem to indicate 1000 samples per channel at 100 MHz. But that's the display, the signals are sampled 16 bits wide at 100 MHz. As to whether 10 nS sampling is sufficient, it depends on how fast the signal is changing and, in my case, that's 25 MHz. My FPGA has sloppy rise/fall time so there is a definite slope to these edges
(2) So can I measure the desired signals?
Yes, but you're getting a little thin up at 25 MHz. Generally, logic signals are synchronous so we kind of know the timing relationships. If you're trying to look at setup or hold time, at higher frequencies, this probably isn't the right tool.
(3) For the analog channels: I assume from Dave's video that this is a reasonable spec for a beginner. What does 30 MHz mean?
The bandwidth is the frequency at which a signal is 3 dB down from where it should be. As you approach this point, the displayed level will fall off.
(4) The digital LA is 3.3V CMOS. Does this mean that I can't feed in 5V signals or does it mean that the digital threshold is 3.3V?
It may turn out that the inputs are protected from 5V signals but that's about it. The device is intended for 3.3V inputs so use the appropriate Arduino. This device is too expensive to expose it to extraneous voltages beyond its capability.
Greetings from Germany!
P.S.: It would be quite nice if Dave could make a Fundamental Friday's Video: The specs of a scope and their meanings explained!
I was stationed in Mannheim back in '66-'67 - a very nice tour considering the alternatives.
I think you'll like the Analog Discovery - I certainly like mine. But it isn't a high frequency anything. The first attachment shows those 4 signals I posted above and the second shows the scope image of the 25 MHz signal. Note the '+' marks at each 10 nS grid. Those are the actual samples and, as you can see, it isn't much of a square wave and the voltage levels (0..3.3V) aren't correct. It looks a good deal better on my DS1054Z
I prefer to think of the AD as a device for 1 MHz and below. Not that it doesn't work at higher frequencies, it's just that there are better tools for those frequencies.
Home
Help
Search
Login
Register
