Products > Test Equipment
Bandwidth limit on Siglent SDS2000X Plus oscilloscope
<< < (5/8) > >>
pdenisowski:

--- Quote from: TimFox on July 19, 2023, 08:39:53 pm ---One problem with using 0.35/Trise is knowing in advance what the actual rise time of the pulse source is.

--- End quote ---

Another problem is that the "0.35" factor only applies when the scope has a Gaussian frequency roll-off.  A scope with a more flat or brick wall type frequency response would require the use of a higher number factor (usually 0.4 - 0.45).

Since almost all analog scopes have a Gaussian frequency response, 0.35 is commonly cited as the factor to be used when calculating bandwidth from rise time, but care should be taken if the shape of the scope's frequency response is not known.  Also, it's important to know whether "rise time" is being defined as the 10-90 or 20-80 interval.

In my experience, the easier and more accurate approach is to use the "amplitude of a sinusoid" method to determine the bandwidth of a scope.  The bandwidth <-> rise time relationship is best used when you already know the rise time of a pulsed signal and need to calculate the minimum required scope bandwidth to accurately measure that pulse.
Performa01:

--- Quote from: pdenisowski on July 20, 2023, 10:08:34 am ---Another problem is that the "0.35" factor only applies when the scope has a Gaussian frequency roll-off.  A scope with a more flat or brick wall type frequency response would require the use of a higher number factor (usually 0.4 - 0.45).

--- End quote ---
You should avoid the term "brick wall" here, as no serious scope frontend will have a frequency response that is even remotely brick wall.

"Measuring" the bandwidth of a scope with a pulse is rather popular - maybe because the results sound so much better (and more impressive), if the Gaussian factor 0.35 is used for scopes with a more flat frequency response.

Another problem with the fast rise-time pulse (e.g. 40 ps) is that it might ask too much from the average modern DSO with less than 10 GHz bandwidth, where the Nyquist frequency exceeds the bandwidth only by a factor of 1.25 in full channel mode, as is very common today. This results in aliasing artifacts which in turn invalidate the rise time measurements. So care must be taken that the half (or even quarter, in cheap instruments) channel sample rate is utilized to avoid this pitfall.

At the end of the day, the good old levelled signal generator is still the preferred method to get a true and really informative picture. With modern DSOs and their advanced features we can even let the scope plot its own frequency response, see the example for the SDS2504X Plus below. And of course 1 MHz is the most commonly used reference frequency - in some cases like the example below (that goes up to 1 GHz) it can be desirable to have it a bit higher, so I chose 10 MHz back then. But thanks to the frequency plot we can clearly see that there are no significant differences below ~150 MHz, so this choice is legit. It might be different with an artificial bandwidth limit, which is implemented by a first order filter, so amplitude drop starts earlier.


Martin72:

--- Quote from: Martin72 on July 19, 2023, 07:49:56 pm ---
--- Quote ---Otherwise if using the Bodnar pulser, it provides extremely fast edges @ 10 MHz
--- End quote ---

Besides the approximate formula (0.35/risetime), you could also try this:

https://www.teledynelecroy.com/doc/frequency-response-measurements

This didn't really work for me when I had a SDS2000X+, but since then 2..3 firmware updates have passed.
Maybe it works now.

--- End quote ---

Once again I'm too stupid to do it and can't get it to work on my HD, it looks just like it did on the X+.
But since 2N3055 showed it with an HD(didn“t find the thread so far), I assume it's my stupidity. 8)
BillyO:

--- Quote from: Performa01 on July 20, 2023, 04:22:46 pm ---"Measuring" the bandwidth of a scope with a pulse is rather popular - maybe because the results sound so much better (and more impressive), if the Gaussian factor 0.35 is used for scopes with a more flat frequency response.

--- End quote ---
Not quite true.  Obviously 0.35/Risetime is going to be LESS than 0.4/Risetime.

BTW, the SDS2000X-P scopes are definitely in the "flat" response category (at least when un-corked), so you should be using 0.4 which will give you more accurate (and more impressive) results than 0.35.

Performa01:

--- Quote from: BillyO on July 20, 2023, 07:03:04 pm ---
--- Quote from: Performa01 on July 20, 2023, 04:22:46 pm ---"Measuring" the bandwidth of a scope with a pulse is rather popular - maybe because the results sound so much better (and more impressive), if the Gaussian factor 0.35 is used for scopes with a more flat frequency response.

--- End quote ---
Not quite true.  Obviously 0.35/Risetime is going to be LESS than 0.4/Risetime.

--- End quote ---
Thank you for pointing that faux pas out. You can tell that I've never worked with that formula, always used accurate frequency response measurements instead. There were several reasons that got my thinking wrong - ridiculous bandwidth claims being one of them.


--- Quote from: BillyO on July 20, 2023, 07:03:04 pm ---BTW, the SDS2000X-P scopes are definitely in the "flat" response category (at least when un-corked), so you should be using 0.4 which will give you more accurate (and more impressive) results than 0.35.

--- End quote ---
Well, let's see how the formula accords to the data sheet for the various models:

Model       tr [ns]  BW real [MHz]  k [-]
SDS2104X+   3,50     100            0,350
SDS2204X+   1,70     200            0,340
SDS2354X+   1,00     350            0,350
SDS2504X+   0,80     500            0,400
SDS5034X    1,00     350            0,350
SDS5054X    0,70     500            0,350
SDS5104X    0,40     1000           0,400
SDS6054A    0,55     500            0,275
SDS6104A    0,35     1000           0,350
SDS6204A    0,23     2000           0,460

According to the datasheet for the SDS2000X+, the factor 0.35 would be accurate up to 350 MHz, i.e. for all artificially bandwith limited models/modes. According to this, factor 0.4 would be correct for the 500 MHz option. Now we all know that the bandwidth is significantly higher than specified for the SDS2000X+, so we can expect faster rise times too.

It is basically the same for the SDS5000X. The artificial bandwidth limit (which is just a first order lowpass) leads to a Gaussian frequency response, hence factor 0.35 in the data sheet. for the 1 GHz model, the natural roll-off of the input buffer and PGA lead to a more flat response, hence factor 0.4.

It looks very different for a higher bandwidth scope like the SDS6000A. The 0.35 factor fits for the 1 GHz model, but nothing else matches.
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