Electronics > Beginners
Oscillo-confusion MHz GSa/s wfm/s Mpts
<< < (8/13) > >>
hamiltont:
The problem I had in your situation was I had no idea what my base MHz requirements were. I understood that bandwidth was a core consideration, and more bandwidth lets me scope faster signals, but....how fast are the signals I am interested in?!?!?!

To solve this I looked into common bit rates for digital protocols (my main interest area) and used a trick from stackexchange to turn them into megahertz. Here is a good table - https://en.wikipedia.org/wiki/List_of_interface_bit_rates

For example, USB 1.0 in full speed mode runs at 12Mbit/s. It's digital, so a 1 followed by a 0 would look like a square wave e.g    |▔▔|___|   so 12/2 (two bits per entire "wave") means it runs at about 6MHz. So if you want to play with USB 1.0, you need a scope that can reliably handle a wave of 6MHz - typical advice is to get a scope with more bandwidth than you need due to attenuation(see image belwo). Once you know what MHz your signal of interest will require, other posters can give you lots of guidance on how to choose the best scope.

Some other examples:

USB 2.0 High speed - 480Mbit/s e.g. 240MHz
USB 3.0 Superspeed - 5Gbit/s e.g. 2.5GHz
Ethernet 10-Bx - 10Mbit/s e.g. 5Mhz
Ethernet 100-Bx - 100Mbit/s e.g. 50MHz
I2C - 3.4Mbit/s e.g. 1.7MHz
HDMI 1.0 video - ~5Gbit/s e.g. 2.5GHz 


tggzzz:

--- Quote from: rstofer on January 02, 2019, 03:15:46 pm ---
--- Quote from: tggzzz on January 02, 2019, 01:52:31 pm ---
--- Quote from: rstofer on January 02, 2019, 12:37:05 am ---High bandwidth will never be an issue for those interested in audio. 

--- End quote ---

There bits/resolution and linearity are probably more important. The Analog Discovery is pretty good in that respect.

--- End quote ---
In my view, the Analog Discovery is unappreciated.  The lab courses associated with Digilent's Real Analog course demonstrate many of the features.  It truly is an electronics lab in a small box.

--- Quote ---

--- Quote ---My interests are up to around 50 MHz on uC pins and usually not that high.  Maybe something in the 2-5 MHz range.

--- End quote ---

Ahem. That 2Mb/s signal could have components above 1GHz, depending on logic family. But you know that!

Personally I'd use a cheap and nasty LA to capture/decode such signals.

--- End quote ---
Whether the LA can display setup and hold times will be a function of its sample rate.  I have a LA that will sample at 200 MSa/s which is pretty good but before I went for one of the $8 units, I would want to look carefully at sample rate.  My scope will sample at 1 GSa/s which will more accurately display the time difference between events.

--- End quote ---

A two channel scope is better for measuring setup and hold times than an LA, and will also do the other necessary signal integrity measurements.

Only after that is it worth flipping to the digital domain with an LA.


--- Quote ---ETA:  Logic Analyzers will have much better triggering options.  I sometimes have to add logic just to create a trigger when using a scope. 

--- End quote ---

Exactly. LAs are better at allowing you to ignore irrelevant stuff, and concentrate on the rarer important information.


--- Quote ---Note how the sample rate drops when the device is used for state analysis and the device under test provides the sample clock.  State analysis is my primary use of a LA.  I don't care what the state of logic signals is relative to some arbitrary internal LA clock, I want to know what their state is relative to the internal logic clock.

--- End quote ---

Agreed on all counts (pun intended).

At high data rates in general purpose LAs, usually the synchronous state capture rate is only 25% of the asynchronous capture rate, since 4 samples are used to detect two edges. The memory length may or may not be reduced.

Some LAs, especially those operating at low speed or special purpose LAs, have separate clock inputs which clock the LAs logic directly. They can operate at full speed.

Often a high speed LA has "strange" speed/size limitations. These can often be understood if you know the input structures in modern FPGAs, especially the SERDES blocks.
tggzzz:

--- Quote from: hamiltont on January 02, 2019, 06:01:31 pm ---The problem I had in your situation was I had no idea what my base MHz requirements were. I understood that bandwidth was a core consideration, and more bandwidth lets me scope faster signals, but....how fast are the signals I am interested in?!?!?!

To solve this I looked into common bit rates for digital protocols (my main interest area) and used a trick from stackexchange to turn them into megahertz.

--- End quote ---

Stackexchange is worth what you pay for it. I'm afraid your understanding is simply and completely wrong.

For digital signals, the only thing that matters is the transition speed, tr; the period is completely and utterly irrelevant. The usual rule-of-thumb is that the signal bandwidth is 0.35/tr. There are a few nuances, but at this level they can be ignored.

If you don't believe standard theory and practice and my statement, perhaps the measurements shown at https://entertaininghacks.wordpress.com/2018/05/08/digital-signal-integrity-and-bandwidth-signals-risetime-is-important-period-is-irrelevant/ will convince you. It shows the very different frequency content of three different 1kHz signals:

* a 1kHz square wave with 120μs rise and fall times – the baseline for comparison
* a 1kHz square wave with 120ns rise and fall times – representative of a 1kHz digital clock signal
* a 1kHz signal with a width of 10μs (1% duty cycle), 120ns rise and fall times –  representative of a general digital signalIn reality the digital signals won't have tr=120ns. Instead 12ns, 1.2ns and 0.12ns can all be found in modern jellybean logic. Even a 74LVC1G04 gate is 250ps with frequency content above 1GHz: see https://www.eevblog.com/forum/testgear/show-us-your-square-wave/msg1902941/#msg1902941

Hence if you have, say, an Arduino which changes the output once per hour, and it takes 5ns to go from high to low, then the maximum bandwith in the signal is 70MHz (i.e. 0.35/5e-9) - and you scope needs to be faster than that.
james_s:
I don't think that's necessarily true. Yes you need a lot of bandwidth to display a really square looking square wave accurately, but the thing is most of the time you don't really need to display it accurately, you just need to understand the limitations of your test gear. Thousands of hobbyists have gotten by with scopes of 20MHz or less debugging digital circuitry. Occasionally seeing the rise time of a signal matters but most of the time you just need to see the pattern of 1s and 0s or see when a signal is changing states relative to another signal. To say someone *needs* x amount of bandwidth is somewhat subjective, more bandwidth is always nice, but you don't need a 400MHz scope to debug an Arduino, a much less exotic instrument will do the job, even if it rounds off the edges of your square waves. Even by the time you get to a 100MHz scope probes and probing technique starts to become an art and IMO it's simply not something a beginner needs to worry too much about.
mvs:

--- Quote from: FriedMule on December 31, 2018, 07:29:01 pm ---I am now looking at the
Siglent SDS1204X-E 200Mhz 1 GSa/s

but may be willing to buy the
Siglent SDS2352X-E 2Ch 350MHz 2 GSa/s

--- End quote ---
This specs are not quite correct.
SDS1204X-E has actually two 1GSa/s ADCs, each serving a pair of inputs (1+2, 3+4). It can sample at 1GSa/s in dual channel mode, if inputs are chosen from different ADCs.
SDS2352X-E has only one 2GSa/s ADC and it will reduce sampling speed to 1GSa/s in dual channel mode.
Navigation
Message Index
Next page
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod