Author Topic: Siglent SDS800X HD Review & Demonstration Thread  (Read 213506 times)

0 Members and 2 Guests are viewing this topic.

Offline RAPo

  • Frequent Contributor
  • **
  • Posts: 673
  • Country: nl
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #225 on: March 16, 2024, 02:19:56 pm »
Thanks, that is a clever method (if and when there is a dot mode on the xy-plot window).

For measuring how fast the average is calculated one can use this method:
https://www.eevblog.com/forum/testgear/scope-with-fast-waveform-averaging/msg4340002/#msg4340002

For X-Y plot, I would use two sine waves with 90 degree phase shift at low frequency (say 1 Hz) and set the scope to a fast time scale (say 10 nsec/div). If the display is in dot mode, you would expect a series of dot blotches appearing around the circle. The number of blotches tells you the number of updates in 1 second. Depending on persistence, one could also set a finite number of triggers, so the blotches don't run around the circle and start to overlap.

Edit: should still be visible in line mode if the number of blotches around the circle is small, so frequency of sine wave can be increased.
 
The following users thanked this post: egonotto

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 4130
  • Country: fi
  • Born in Finland with DLL21 in hand
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #226 on: March 16, 2024, 03:20:27 pm »


1) I wonder what happens if an averaged trace is displayed in dots mode? Are the averaged dots now evenly spaced with 1/sample_rate interval, or do they still retain fractional horizontal positions? [Particularly when a large number of traces are averaged.]

Here, Dots, Persistence and Average 256 so there can see that samples have been in all possible fractional positions (here sample interval 2ns)
Ch3 and Ch4 on only for drop samplerate (traces out of screen)math

Note that math trace in wfm/s is slower than channel direct wfm/s


« Last Edit: March 16, 2024, 04:02:05 pm by rf-loop »
BEV of course. Cars with smoke exhaust pipes - go to museum. In Finland quite all electric power is made using nuclear, wind, solar and water.

Wises must compel the mad barbarians to stop their crimes against humanity. Where have the (strong)wises gone?
 
The following users thanked this post: Performa01, egonotto, gf, core

Online Martin72

  • Super Contributor
  • ***
  • Posts: 6292
  • Country: de
  • Testfield Technician
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #227 on: March 16, 2024, 04:56:41 pm »
Obviously the signal was implicitly up-sampled/interpolated to a higher rate. I wonder if the same applies to the SDS800X?

Unlike the SDS2000xplus (at that time, this may have been fixed), the 800X HD "sticks" to the entered interpolation coefficient, in this example 20 (40GSa/s).
"Comparison is the end of happiness and the beginning of dissatisfaction."
(Kierkegaard)
Siglent SDS800X HD Deep Review
 
The following users thanked this post: egonotto

Offline gf

  • Super Contributor
  • ***
  • Posts: 1302
  • Country: de
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #228 on: March 16, 2024, 05:01:58 pm »
Obviously the signal was implicitly up-sampled/interpolated to a higher rate. I wonder if the same applies to the SDS800X?

Unlike the SDS2000xplus (at that time, this may have been fixed), the 800X HD "sticks" to the entered interpolation coefficient, in this example 20 (40GSa/s).

Thanks! Good that it honores the requested upsampling factor :-+

EDIT:
But I wonder how it can do a 16k point FFT @40GSa/s if the captured record length is only 200 points? 20x interpolation gives only 4000 point, which is less than 16k.
[ Theoreticaly it would be possible to window the 4000 points and then zero-pad to 16k before  invoking the FFT. This would result in an interpolation in the frequency domain. But that's just speculation :-// ]

But I actually meant FFT(Average(Cx)), without d/dx or explicit interpolate in the formula.
IIRC, it did interpolate, too, although not explicitly asked to do it.
« Last Edit: March 16, 2024, 09:12:44 pm by gf »
 
The following users thanked this post: egonotto

Offline Performa01Topic starter

  • Super Contributor
  • ***
  • Posts: 1686
  • Country: at
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #229 on: March 16, 2024, 09:40:01 pm »
In message #175 I asked if you accept “beginners” asking you questions about “obvious things”.
In your message #180 you replied: “don’t mind people asking questions here, even newbie questions 😉”
Yes, you are right. I might have just been a bit surprised. Since this is a “Review & Demonstration” thread for a specific digital oscilloscope, I expected a newbie-question in this context to be from someone unfamiliar with Siglent scopes – or even digital scopes in general, so I was prepared to deal with DSO specifics, but not with basics of signal theory and Fourier series for imaginary signals that can never exist in the real world.


On YouTube, on the R&S channel and on the Teledyne LeCroy channel, there are “theoretical videos” on the subject of “square wave excitation frequency response”, in a different way from your presentation. They just don't have the practical part. In summary, they state that the BW of the DSO must be greater than 5x the frequency of the fundamental wave signal.
Well, I don’t think much of such videos. Firstly, we need not tell serious users of an oscilloscope what bandwidth they require for displaying a square wave, and even more importantly, there’s no use in telling some street number like “5 times the square wave frequency”. Because it is simply not correct.


2. In my opinion as a “beginner”, for a non-SIGLENT user (not yet), the fact that it presents the “Bandwidth” of a sinusoidal signal, and “Pulse Response” does not allow us to conclude that the frequency is tolerable with deformations, for the square wave SDS800X, is f=80MHz (photo: SDS824X HD_Square_3.5ns_80MHz).
Based on the “Bandwidth” of sine signal, and “Pulse Response”, how did you reach the conclusion of BW for square wave?
Rise time is the key and I’ve demonstrated it. Here is how I would deal with it on a theoretical basis:

We were talking about the SDS824X HD, an oscilloscope whose bandwidth has been measured to be about 245 MHz. There is the well-known formula to estimate rise time from bandwidth:

tr = 0.35 / 245 MHz = 1.43 ns.

The SDS824X HD is specified for 1.8 ns (which it probably has if more than two channels are active), yet with just a single channel, I had estimated its rise time from my measurements as ~1.5 ns and the formula tells us that it might actually be even a tad lower than that. For the ease of use and the sake of this exercise, let’s just continue working with the initial assumption of 1.5 ns.

We can calculate the total rise time tr from the signal rise time trs and the scope frontend rise time trf:

tr = √(trs² + trf²);

The error would then be: err = (tr / trs - 1) * 100 [%];

Rewriting this formula to get the permissible signal rise time for a given error margin leaves us with:

trs = √(trf² / ((err / 100 + 1)² - 1));

How big an error are we willing to accept?

1%? Then the signal rise time should not be faster than 10.6 ns.
2%? Then the signal rise time should not be faster than 7.5 ns.
5%? Then the signal rise time should not be faster than 4.7 ns.
10%? Then the signal rise time should not be faster than 3.3 ns.
15%? Then the signal rise time should not be faster than 2.65 ns.
20%? Then the signal rise time should not be faster than 2.26 ns.

Or we don’t care? Then everything is fine, as long as at least the fundamental frequency does not exceed the bandwidth of the DSO.

I’ve always used the spectrum of a signal to decide how much bandwidth is needed to get a faithful reproduction on the DSO, just as I’ve demonstrated it in my earlier post with the 3.5 ns rise time square wave. Once again, it’s up to the user to decide what level of harmonics needs to be within the acquisition bandwidth. And of course we cannot accurately determine the expected error from this, but with some imagination and experience, we can know what we really need, also depending on the task on hand.

I have already pointed out that in digital communications we never had any bandwidth to waste, so all the digital modulations always really transported more or less distorted sine waves, which only get converted back into square waves within the receiver.


3. As you demonstrated that for the SDS800X, BW = 245MHz for sinusoidal signal, I just divided it by 5, and asked you to start the test at f=50MHz, and increase the frequency, and you arrived at f=80MHz with a “square wave” (“photo: Ref-Spec_Square _1ns_80MHz”), which 80 x 3 = 240MHz…
I’ve shown the two extremes. First, the benign and still fairly accurate 3.5 ns, where we can expect an error of max. 8.8%. Second, aggressive 1 ns rise time with up to 80% error. The actual errors were lower, partly because the rise time of the SDS824X HD is a little faster than the 1.5 ns that we’ve assumed here. Anyway, we could define a “square-wave-bandwidth” only if we have a clear error margin in mind. The 80 MHz square wave with 3.5 ns rise time was rendered perfectly okay by the SDS824X HD, even though it looked pretty much like a sine. Yes, there are irritations, because some folks probably expect to get a trapezoid waveform instead of a sine, and that is actually where the bandwidth limited square wave collides with our imagination. But how do you define aberrations in the expected shape of a signal edge? As a consequence, I would say the signal rise time is the more readily available and less complicated metric.


4. As for the video by “Professor Michel van Biezen”, for experts, the issue comes down to just the photo at the beginning of the video… the equation and the Fourier Series Coefficients.
In my “beginner” analysis, it doesn’t matter much if the theoretical equation doesn’t support a square wave signal of 1ns risetime…
I have to admit that I have difficulties understanding what you actually want. In a previous post you wanted me to show how the Fourier coefficients of the DSO measurements match the theory shown in that video, now you claim that accuracy doesn’t matter for you.


5. As for the Fourier Series, you learned another formula that I had never seen. This Fourier Series formula has many ways to learn it.
Well, in practice me at least do not deal with instantaneous time-domain values like cos(ω * t), but with harmonic signal levels from the frequency domain.


Having observed your comments, it is still not possible to understand whether the “Fourier Series Coefficients” should be placed in Vp, Vpp, Vrm, or something else.
In my example it has quite obviously been a mixture of all. The formula, which originally includes 4 times the peak value (which is actually half the amplitude) as first term, has been simplified by me by using double the amplitude instead. It goes without saying that the amplitude of an ideal square wave is identical with its peak-to-peak value, hence Vpp. The Fourier coefficients are still related to half the signal amplitude; hence they are peak values, which I have clearly stated in my article (Vp). Finally, I have converted the peak values to dBVrms. Yes, I’ve just written dBV, because dBVrms happens to be the only dBV unit available in Siglent DSOs (and many others).


Request: Even if it contains errors greater than 20%, as there is no perfect square wave, if you can assemble and present an equation of the Fourier Series, with f=80MHz, of the square wave, I will be very grateful. As in the formula presented in the video by “Professor Michel van Biezen”.
Sorry, I honestly don’t know what you want. The equation for a Fourier series looks always the same, no matter if it is for 1 kHz or 80 MHz.

Maybe someone else can chime in, because I personally am at my wit’s end. All I can do now is strongly recommend that you stop fixating on that video.


6. In the photo: “Ref-Spec_Square _1ns_80MHz”, in the “Peak Search Table”, in “Marker 6” it has f=880.00000MHz. I said here that it has “8 digits”, but you repeatedly state that the SDS800X has “7 digits” in the Counter. Do you count from “0 to 7”, or from “1 to 8”?
Yes, the math might show more digits in order to provide more resolution for detailed analysis, but that is not necessarily accurate in absolute terms.

What I am referring to is the always visible trigger frequency counter in the top right corner of the screen. This is the closest thing to a real frequency counter and it has class-leading 7-digits resolution, albeit only pretty average 25 ppm absolute accuracy.


7. After your “class”, I agree with you that the “Risetime” is the starting point for buying an AWG, and “the good one is the 1ns”, but it is the top of the line, and is far above of my hobby budget.
Once again, I’m not sure to understand you. While it is obvious that it’s quite nice to have a professional grade AWG that can deliver fast rise time signals, this is a completely different topic. I was always trying to get the message through, that a square wave should be primarily judged by its rise time and not frequency, even though – and this once again is something I’ve already stated – the two are somewhat related. A square wave at 240 MHz just cannot have a rise time significantly slower than 1 ns.

The actual message is: on a SDS824X HD you can faithfully reproduce a 50 MHz square wave with 3.5 ns rise time, while you are not able to do the same with a 1 kHz square wave that has 1 ns rise time - even though the scope-bandwidth is 245000 times the square wave frequency!
 
The following users thanked this post: Electro Fan, egonotto, tautech, Bad_Driver, BRZ.tech

Offline ebastler

  • Super Contributor
  • ***
  • Posts: 6676
  • Country: de
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #230 on: March 16, 2024, 09:51:38 pm »
7. After your “class”, I agree with you that the “Risetime” is the starting point for buying an AWG, and “the good one is the 1ns”, but it is the top of the line, and is far above of my hobby budget.
Once again, I’m not sure to understand you. While it is obvious that it’s quite nice to have a professional grade AWG that can deliver fast rise time signals, this is a completely different topic. I was always trying to get the message through, that a square wave should be primarily judged by its rise time and not frequency, even though – and this once again is something I’ve already stated – the two are somewhat related. A square wave at 240 MHz just cannot have a rise time significantly slower than 1 ns.

The actual message is: on a SDS824X HD you can faithfully reproduce a 50 MHz square wave with 3.5 ns rise time, while you are not able to do the same with a 1 kHz square wave that has 1 ns rise time - even though the scope-bandwidth is 245000 times the square wave frequency!

I think I know where the misunderstanding lies. Let me try to paraphrase:

Performa01's message is: When considering how well you can observe a given square wave, it's the risetime which counts, i.e. which is the critical parameter (and not so much the frequency). Whereas BRZ might have misunderstood it as: When considering how to best generate a square wave, aim for the generator with the fastest risetime.

The latter is a misinterpretation. In fact, I would say one should avoid faster-than-necessary risetimes in many cases, due to the problems in signal propagation and detection which the high-frequency components bring. (As discussed by Performa01 in the context of the challenges they bring when observing square waves on an oscilloscope.)
 
The following users thanked this post: Performa01, egonotto, BRZ.tech

Offline Performa01Topic starter

  • Super Contributor
  • ***
  • Posts: 1686
  • Country: at
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #231 on: March 16, 2024, 10:03:09 pm »
1) I wonder what happens if an averaged trace is displayed in dots mode? Are the averaged dots now evenly spaced with 1/sample_rate interval, or do they still retain fractional horizontal positions? [Particularly when a large number of traces are averaged.]
Yes, they retain fractonal positions, as rf-loop has nicely demonstrated. Yet as we already know, Average as a math function is not so fast and we have to make do with about 19 averages per second, whereas the original signal trace is orders of magnitude faster.


2) I think to rememer that you (or was it somebody else?) did demonstrate in a different thread that FFT(average(Cx)) resulted in a higher FFT sample rate than the original sample rate of Cx, on either the SDS2000 or SDS6000 (don't remember which one). Obviously the signal was implicitly up-sampled/interpolated to a higher rate. I wonder if the same applies to the SDS800X?
I personally have never used FFT(Average(Cx)), because it doesn't make sense to me. Tests have shown that it makes matters much worse in case of weak signals, because removing all the noise also prevents all the implicite or explicite resolution enhancement techniques from working. Not to mention the fact that we could only watch static signals with that anyway.

I've just tried it now and quite unsurprisingly, it does not change any sample rate and I'm pretty sure it never did.

Also Interpolate() always worked correctly. It was the Derivative() in conjunction with Interpolate(), which resulted in too high numbers as sample rate in the FFT - but that was only observed in the SDS2000X Plus, not in any other Sglent SDS model.
 
The following users thanked this post: egonotto

Offline BRZ.tech

  • Regular Contributor
  • *
  • Posts: 51
  • Country: br
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #232 on: March 17, 2024, 01:07:20 am »
...
Once again, I’m not sure to understand you. While it is obvious that it’s quite nice to have a professional grade AWG that can deliver fast rise time signals, this is a completely different topic. I was always trying to get the message through, that a square wave should be primarily judged by its rise time and not frequency, even though – and this once again is something I’ve already stated – the two are somewhat related. A square wave at 240 MHz just cannot have a rise time significantly slower than 1 ns.

The actual message is: on a SDS824X HD you can faithfully reproduce a 50 MHz square wave with 3.5 ns rise time, while you are not able to do the same with a 1 kHz square wave that has 1 ns rise time - even though the scope-bandwidth is 245000 times the square wave frequency!

Dear @Performa01,
I am very grateful for your explanations. And I apologize for taking up your precious time.
I realized from your comments and the performance of the SDS800X HD, and other SIGLENT instruments, that they are much better and more reliable than they appear to be, and seek perfection in providing their customers with the possibilities of carrying out Measurements and Tests in general Extremely Reliable.

In fact, when it comes to DSO/MSO I have not reached the stage of stating what is extremely correct, in relation to what is not correct. Perhaps debugging will occur with time and the practice of making correct measurements and having the fundamental concepts well founded.

Really, for an expert, my questions and doubts of “newbie” can cause confusion in understanding what I asked. But I see that you and the other colleagues treated it with great respect, and tried to explain things in the best possible way.

As for "these videos", I still don't have to make a value judgment, what I have done is watch and follow educational material from sources that appear to be safe. But it seems that they are not that so reliable and safe.
TKS.
73
 
The following users thanked this post: Performa01

Online Bad_Driver

  • Frequent Contributor
  • **
  • Posts: 373
  • Country: de
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #233 on: March 17, 2024, 08:39:51 am »
Dear Performer01

thank you very much for the many very insightful explanations and tips on using the SDS800.
I have to admit to my shame that I am also doing many things for the first time and understood that I have never found or used on my SDS2000X+, which has been on my desk for 4 years.  |O
 
The following users thanked this post: Performa01, egonotto, 2N3055

Offline Performa01Topic starter

  • Super Contributor
  • ***
  • Posts: 1686
  • Country: at
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #234 on: March 17, 2024, 10:58:06 am »
FFT-Setup

Like all Siglent DOSs, starting with the venerable SDS1202X-E, the SDS800X HD have a very useful FFT implementation. To get the most out of it, we should be able to set it up correctly.

The FFT-length in the SDS800X HD can be up to 2 Mpts. This enables low resolution bandwidths and low noise, but it's still “only” a 12-bit DSO, hence results below -73 dBFS shouldn’t be trusted blindingly. In many cases, the usable dynamic range with decent accuracy can be up to 100 dB though.

Some hints for proper setup of the FFT on Siglent DSOs in general and the SDS800X HD in particular:


FFT-Bandwidth and RBW

This is quite different to a traditional SA (spectrum analyzer). There is no menu for the resolution bandwidth and also no direct setting for the FFT-bandwidth, even though we have a menu item for the horizontal display parameters, i.e. start/stop frequencies for wideband measurements and center frequency/span for narrowband applications. But this is just for zooming into a longer FFT result; for best speed and lowest RBW (resolution bandwidth) we need to make sure that no high zoom factor is required to get the desired display. The following rules apply:

•   The analysis bandwidth (FFT-BW) is always half the FFT sample rate (FFT-SR).
•   The frequency step Δf is the sample rate divided by the number of FFT points.
•   In the Acquire menu, a constant sample rate can be set in order to also limit the FFT-sample rate.
•   The resolution bandwidth (RBW) is the frequency step multiplied by a factor specific for the window function in use.

The maximum number of FFT points can be up to 2 Mpts, but it also depends on the record length, which increases with slower time base settings, which in turn might be limited by the maximum memory as defined in the Acquire menu. Apart from that, the number of FFT points can be further limited by the corresponding setting in the FFT-Config menu.

RBW = Δf * k, where k is the 3 dB bandwidth factor in bins, depending on the window function:

k: Rectangle 0.99, Blackman 1.74, Hanning 1.62, Hamming 1.64, Flattop 3.73.

Blackman and especially Flattop are the most universal and useful window functions in practice. You definitely should stick to these two as long as you cannot prove that some other window would actually work better in your specific application.

Thus: Δf = RBW / 4 (rounded) in case of the flattop window or RBW / 2 for Blackman.

To get the proper settings for any given FFT-BW and RBW pair, proceed as follows:

Determine the FFT sample rate: SR = FFT-BW * 2 [Sa/s];
Determine the number of FFT points: FFT-pts >= SR / Δf [pts];
Determine the time base: TB >= FFT-pts / SR / 10 [s/div];

As mentioned earlier, you can lower the FFT-sample rate by setting a constant acquisition sample rate; this can be useful when you want really low FFT-sample rates but do not want to use very slow time base settings, which would slow down the acquisition considerably.


Setting up the FFT

Even from the best FFT implementation, we can only expect good results as long as the scope has been set up properly for that specific task. How many so called “reviews” have we seen where FFT has been engaged and some scope settings randomly altered just to get a halfway plausible but actually not very meaningful FFT graph, which was then either praised or criticized?

Of course we can get away with some quick & dirty setup if the requirements are low, but even then we should never ignore the most important parameters like FFT bandwidth, which should always cover the full signal spectrum, otherwise aliasing artefacts could easily spoil our measurement results.

For optimal speed, frequency resolution and dynamic range, we need to put a little more effort into a proper setup, which has quite different requirements compared to the usual Y-t view (aka time domain). Below there is a complete checklist how to properly set up the DSO for analysis in the frequency domain (most of these topics should be obvious, but still listed for completeness):

1.   Set acquisition mode to normal. Use ERES only for a good reason and stay away from average. Avoid Peak Detect under all circumstances and without any exception!
2.   Use edge trigger in auto mode to make sure signal acquisition doesn’t stop even when the signal amplitude drops below the trigger sensitivity. FFT doesn’t require a stable trigger, so you can also use AC-line trigger for that.
3.   Determine the lower bandwidth limit for the FFT analysis. If it is >10 Hz, use AC-coupling for the input channel to ensure maximum dynamic range even with large DC offsets and/or high input sensitivities. If DC-coupling has to be used, use the vertical position control to compensate for any DC offset, so you can optimize sensitivity, hence get the highest dynamic range.
4.   Determine the upper bandwidth limit for the FFT analysis. In order to avoid aliasing artifacts, this should not only cover the desired analysis bandwidth, but include the highest expected input frequency. In general, it’s best to start with a higher upper bandwidth limit and reduce it only after it has been confirmed that there is no significant signal content above the desired final limit.
5.   Choose the frequency step size according to the explanations given earlier in this article, which would be about one quarter of the required resolution bandwidth when using the Flattop window.
6.   Find an appropriate set of horizontal time base setting and the number of FFT points; refer to the explanations given earlier in this article. You should watch the displayed FFT parameters while altering the time base and double check that they match your expectations. Be aware that the desired resolution bandwidth might not be achievable due to the limited choice of sample rates and FFT lengths and/or the maximum specified FFT length of 2 Mpts.
7.   Engage FFT mode, select the correct source channel and start with Split Screen mode.
8.   Set the vertical gain so that the peak amplitude of the input signal is between ±2 to ±4 divisions.
9.   Set the horizontal FFT display parameters according to the bandwidth you want to display and select linear/decade mode for the frequency axis. Decade is advantageous for wideband measurements, whereas linear is best for narrowband applications.
10.   Set the vertical FFT display parameters, i.e. the desired level units (dBV or dBm, forget volts!) and make sure the external load impedance matches reality whenever working with power levels, i.e. dBm. Set the reference level and vertical scale so that the FFT amplitude range of interest makes best use of the available screen space.
11.   Setup (at least) automatic peak-peak measurement for the input channel. During frequency domain analysis, especially in Exclusive mode, keep an eye on the Vpp measurement for the input channel to make sure no overload occurs.
12.   Make sure the desired window function is selected.

Hint: stay in Split Screen mode until the amplitude setup is finished and the levels are reasonably stable, then switch to Exclusive mode. By keeping an eye on the peak-to-peak measurement of the input signal, you can still detect an overload condition instantly; the scope indicates that by displaying > instead of = in front of the measurement value.

Example: Pk-Pk >851.875mV instead of Pk-Pk  755.000mV.



FFT Window Functions

For the ones who try to understand the consequences of certain settings in the FFT analysis – this is about the window functions.

Why are there so many different windows (only few of them available on the SDS oscilloscope)? What is the best window to use?

There have been times when processor systems haven’t been nearly as powerful as today. Non-RISC architectures with just 1 MHz clock frequency and less than 1 kB RAM were not uncommon during the seventies of last century. Instruments that could compute a FFT at all have been rather exotic, and FFT-lengths like 64 points were quite common. In the light of this, there is no wonder that less than ideal FFT-window functions optimized for certain tasks were popular.

Sometimes there are descriptions about the benefits and drawbacks of the various window functions, yet most folks would rather not care and want a universal setting that works for them every time. Just like with a traditional SA (spectrum analyzer) with analog RBW (resolution bandwidth) filters in the final IF (intermediate frequency) path. And fortunately, there is one…

There are several features of a window function, and two of them are amplitude accuracy and resolution bandwidth. If we look at just these two properties, then the rectangle window would have the narrowest resolution bandwidth but the worst amplitude error, whereas it’s just the opposite for the Flattop window.

So, whenever we need the best frequency resolution, we just sacrifice a bit of accuracy and use the Rectangle window?

It’s not that simple. An FFT divides the entire analysis bandwidth into frequency bins. If, for instance, we have an FFT-length of 32768 points, then we get 16384 such frequency bins and at an effective sample rate of 2 MSa/s, each of them will be 61.04 Hz wide. In this case, 61.04 Hz is the bin width and the bin spacing at the same time. The center of a bin will always be an integer multiple of the bin width.

Now FFT-windows behave differently, depending on the offset of the input signal frequency from the bin center. I did a selectivity test for the various window functions available on an SDS800X HD and used the before mentioned parameters:

FFT-sample rate = 2 MSa/s
FFT-Length = 32768 pts
Bin-width = 61.03515625… Hz

The test will be for amplitude accuracy and the -20 dB, -40 dB and -60 dB selectivity. I define the latter at the frequency distance for a -20, -40 or -60 dBc signal to still produce a visible 3 dB peak in the spectrum (and not drowned out by the leakage of the neighboring o dBc reference signal).
 
The -3 dB bandwidth might be most important for characterizing the passband of any two-port network, but for a filter, where the selectivity is the main concern, the bandwidth at a useful attenuation is even more important.

The metric of a filter shape factor exists, which is usually defined as the ratio of the filter bandwidth at -60 dB and -6 dB. The shape-factor might even be the most important property of any RBW filter at all, because it ultimately defines selectivity. Consider what a proper BP (Band-Pass) filter, as it might be found in any traditional SA, looks like:


Ref-BP_Cheby9_0.2dB

This is a 9th order Chebyshev BP filter with 0.2 dB passband ripple. It has a -3 dB bandwidth of about 10 kHz and less than 16 kHz bandwidth at -60 dBc. The 3/60 shape factor is thus ~1.57. The lower the shape factor, the higher the selectivity.

What does it mean in practice? We can distinguish a strong signal at 0 dBc together with a weak signal at -60 dBc next to it, as long as the distance is at least ~13 KHz, i.e. 1.3 times the RBW. That is excellent and makes for a useful analysis in the frequency domain.

An FFT is quite different to a classic continuously swept spectrum analyzer. If we feed a stable sine wave into a swept analyzer, we’ll get the frequency response plot of the RBW-filter.

The FFT doesn’t plot a continuous filter response, but simply shows the outputs of the individual frequency bins, and ideally only the one bin covering the input frequency responds with the correct amplitude (and phase, but that’s not displayed). It is like a huge filter bank with a bunch (16384 in our next example) of filters, all working in parallel. What we see is different from the RBW-filter response; we rather see the “leakage”, i.e. signal outputs from neighboring “filters” (bins), sometimes even far away from the signal frequency - and there are amplitude errors for the main bin.

A 0 dBm test signal of 499.99342 kHz has been used for the first test, which is equivalent to precisely 500 kHz if the time base of the SDS800X HD were accurate. this is precisely 8192 times the bin width, hence the exact center of a bin.

What do we get?

Window     Ampl.  Selectivity [Hz]
            [dBm]   -20 dB   -40 dB    -60 dB
Blackman     0.0     236.58   236.58    236.58
Flattop      0.0     366.58   416.58    416.58
Hamming      0.0     168.58   168.58    168.58
Hann         0.0     171.58   171.58    171.58
Rectangle    0.0      98.58    98.58     98.58

Look at the rectangle window, with two signals, the 0 dBm reference signal (carrier) and the 2nd signal at -40 dBc (with an enormous amplitude error of 2.8 dB), which creates just a 3 dB peak at a distance of 98.58 Hz, as an example:


SDS824X HD_FFT_Rectangle_0.0_S40dB_500092Hz

With a distance of just 1.615 bin widths, selectivity is quite good.  Yet real-world signals will usually not be an exact integer multiple of the bin-width, so we need more tests.

In any practical application where the FFT of a general-purpose oscilloscope is to be used, we cannot freely define the signal frequencies, hence they will be more or less off center. Even if we could select a frequency, most related signals like intermodulation (mixer) products and spurs can still have any frequency offset with regard to the bin spacing. Consequently, we need to take the worst case into consideration, that is a frequency offset of half the bin-width.

A 0 dBm test signal of 500.02394 kHz has been used for the following test, this is precisely 8192,5 times the bin width, hence the exact bin-border for my individual sample of the SDS824X HD.

Window     Ampl.  Selectivity [Hz]
            [dBm]   -20 dB   -40 dB    -60 dB
Blackman     -1.0    206.06   266.06    426.06
Flattop       0.0    386.06   386.06    451.06
Hamming      -1.8    196.06   196.06    -
Hann         -1.3    203.06   326.06    586.06
Rectangle    -3.8    381.06  2780.06    -                   

This looks very different, doesn’t it? All of a sudden, the rectangle window has 3.8 dB amplitude error and its selectivity isn’t all that good anymore. In fact it is unbelievable >45 bin-widths for the -40 dBc selectivity! By contrast, the Flattop window hasn’t changed at all: the amplitude error is effectively zero as it was before and also the selectivity has only marginally changed. That means more than -60 dBc attenuation at 300 Hz (less than 5 bins) distance from the center.

Hamming does not have a 60 dB attenuation within a reasonable bandwidth – in fact it is so wide that I could not be bothered to measure it. The same applies to the Rectangle window, where the -40 dBc selectivity test didn’t reveal anything useful:


SDS824X HD_FFT_Rectangle_0.5_S40dB_502800Hz

What do we want for a proper RBW filter for spectrum analysis, in order to get serious and accurate measurements?

1.   High dynamic range. We have no use for a RBW filter that has no proper stopband attenuation, hence picks up all the garbage from the neighborhood.
2.   Fast transitions into the stopband, which is equivalent to a low shape factor.
3.   A reasonably flat passband without massive amplitude errors, as soon as the signal frequency gets off center a bit.
4.   And most importantly, all these parameters shall be constant and independent of the exact input frequency.

The only window that appears to be perfect in almost all regards is the Flattop window. It has a very high dynamic range, low shape factor, a totally flat passband, and even more important, its properties remain constant and do not depend on the signal frequency. Only downside: it has the widest bandwidth of all candidates. Yet in modern equipment, where at least 1 Mpts FFT have become standard, we need not desperately look for a windows function that sacrifices a lot of good properties just for a little narrower RBW.

Blackman is the only alternative that I can recommend from the selection in the Siglent SDS800X HD. It has less than half the RBW of the Flattop window and the shape factor is still reasonable. It works down to -60 dBc even with the worst-case frequency offset of half a bin width, and the passband flatness, hence also the amplitude error, is not too high. This is generally true for all window functions of the Blackman-family, especially Blackman-Harris. Siglent only implements the original Blackman window though.

You can take a look at all the remaining window functions in the attachment, albeit only for worst case frequency offset, indicated by “0.5” in the file name.
« Last Edit: March 19, 2024, 03:01:37 pm by Performa01 »
 

Offline ebastler

  • Super Contributor
  • ***
  • Posts: 6676
  • Country: de
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #235 on: March 17, 2024, 11:28:30 am »
Brilliant, thank you Performa01!

You should really think about converting this material into a (PDF?) booklet -- an "Application Guide to Digital Oscilloscopes" or such. Maybe something Siglent could fund, and distribute as either a separate document or merge into the user manuals?

As a short-term step, would it make sense to reorganize the "table of contents" in the initial post? It is currently organized chronologically. But since you have re-visited some topics over time, and the total amount of material is becoming too much to grasp even the whole TOC at a glance -- could the links to the individual posts be grouped into a few larger topic areas? I am not sure what that meta-structure would best look like. Maybe start with basic horizontal and vertical settings and performance, move on to measurements/cursors/math, then specific application areas?

In any case -- thank you again for the huge effort you have and continue to put into this thread. I have not seen anything like it, anywhere and on any scope!
« Last Edit: March 17, 2024, 11:40:54 am by ebastler »
 
The following users thanked this post: Performa01, egonotto, hansibull, Jacon, myf, pdenisowski, mwb1100, core, radiohomebrewer2000

Offline iMo

  • Super Contributor
  • ***
  • Posts: 4896
  • Country: vc
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #236 on: March 17, 2024, 11:52:05 am »
I wonder why that hints above cannot be somehow suggested to the user by the o'scope itself. It cannot be such a rocket science to implement those into the o'scope's system. Simply navigate the user with short messages with above hints when he/she messes up with the settings in a suboptimal or even wrong way..
 
The following users thanked this post: egonotto

Offline Grandchuck

  • Frequent Contributor
  • **
  • Posts: 683
  • Country: us
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #237 on: March 17, 2024, 11:59:12 am »
Many thanks to professor Performa01 for bringing us oscilloscope university.
 
The following users thanked this post: Performa01

Offline gf

  • Super Contributor
  • ***
  • Posts: 1302
  • Country: de
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #238 on: March 17, 2024, 01:58:46 pm »
As we can see, the rectangle window is fantastic. It has no amplitude error at all and an ultra-narrow -3 dB bandwidth of just 6.4 Hz. First doubts should arise though, when we look at the shape factor, which is a rather catastrophic 18.44! Let’s take a look at that:

SDS824X HD_FFT_Rectangle_0.0_3dB

Well, the -3dB bandwidth of a rectangular window is always about 0.884*Δf, which is about 54 Hz for the given parameters.

The root cause why you measure a horribly wrong value of 6.4 Hz is the linear interpolation between adjacent frequency points, which is not appropriate here. With a proper sinc interpolation you would see the true shape of the window function's frequency response and measure the correct bandwidth. In the time domain, you would not sample a sine wave signal with (say) 2.3 points/period either, interpolate the samples linearly and expect that you can still recognize it as sine wave. But the same happens here in the frequency domain. The scope fools you by connecting the frequency points (which are correct) with straight lines. It is important to be aware and not to trust these straight lines blindly.
« Last Edit: March 17, 2024, 02:07:26 pm by gf »
 
The following users thanked this post: Performa01, egonotto, 2N3055, Martin72, mawyatt

Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 3519
  • Country: us
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #239 on: March 17, 2024, 04:37:50 pm »
As we can see, the rectangle window is fantastic. It has no amplitude error at all and an ultra-narrow -3 dB bandwidth of just 6.4 Hz. First doubts should arise though, when we look at the shape factor, which is a rather catastrophic 18.44! Let’s take a look at that:

SDS824X HD_FFT_Rectangle_0.0_3dB

Well, the -3dB bandwidth of a rectangular window is always about 0.884*Δf, which is about 54 Hz for the given parameters.

The root cause why you measure a horribly wrong value of 6.4 Hz is the linear interpolation between adjacent frequency points, which is not appropriate here. With a proper sinc interpolation you would see the true shape of the window function's frequency response and measure the correct bandwidth. In the time domain, you would not sample a sine wave signal with (say) 2.3 points/period either, interpolate the samples linearly and expect that you can still recognize it as sine wave. But the same happens here in the frequency domain. The scope fools you by connecting the frequency points (which are correct) with straight lines. It is important to be aware and not to trust these straight lines blindly.

Good catch, it "looks" like 64Hz as we didn't spot the decimal point!!

Linear interpolation of sparse data, one should expect errors, some cases significant with interpolated results. This was the limiting dynamic range in SPICE like Time Domain simulations even tho SPICE utilized quadradic interpolation, and why in 80s we developed a version that "forced" the simulation at exactly the proper times with a-priori knowledge of the expected post FFT implementation. Doing 2-Tone IMD of ultra-linear systems reveled the SPICE limitations, later Cadence picked up on this.

However, this certainly doesn't take away from the superb efforts of Performa01 which really shows the incredible versatility of these Fine Instruments when in the hands of a Master  :clap:

And we "discredit" these as just DSOs :palm:

Best,
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 
The following users thanked this post: Performa01, egonotto, gf

Offline radiohomebrewer2000

  • Contributor
  • Posts: 39
  • Country: us
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #240 on: March 17, 2024, 04:56:15 pm »
If someone comes up with a guide for the SDS800X HD scopes like the material that Performa1 has been provided with explanations on how to do them as well, I would buy this book.  Seriously. 

I have looked at the documentation for the SDS800X HD series ahead before it is delivered so I can some idea of what to expect.  But, the documentation is very terse.  Maybe I am just used to the manuals (spoiled) for my commercial ham radios from companies like Kenwood, ICOM, Yaesu that provide very good manuals with their ham radios.

Again, I am willing to buy a book for US$50 for more.

 
The following users thanked this post: Performa01, egonotto

Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 3519
  • Country: us
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #241 on: March 17, 2024, 05:02:27 pm »
Didn't you already "Buy the Book", and the SDS814 was just the "Special Promotion" ;D

That's how we "Explained" it to our "CFO/CEO"  ;)

Best,
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 
The following users thanked this post: radiohomebrewer2000

Offline tszaboo

  • Super Contributor
  • ***
  • Posts: 7583
  • Country: nl
  • Current job: ATEX product design
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #242 on: March 17, 2024, 05:25:25 pm »
OK, so I had a go at the Bode plot module. This is a SDS800HD together with a SDG1000X measuring a current transformer-opamp module called ZMCT103C. I was interested in it's transfer function (the current transformer is 1:1000 turns )
The setup only requires a BNC-BNC cable and a BNC-Crocodile cable and a resistor, everything else comes with the instruments. CH2 on the AWG is set to tracking mode, no other settings are necessary on the AWG. The scope is connected to the AWG with a USB cable, it recognized it automatically, CH2 of the AWG is connected to the scope. Other than that, I just had to enter the frequency range, and the number of points, set up the channels.
One thing I noticed that the Bode plot will measure 3 outputs if set up, that's a nice touch that doesn't really cost anything. It would be nice if the save button would work in the bode plot menu, but now it does nothing and the save location has to be entered manually. Overall I'm very satisfied with this setup, and the capabilities of both the scope and the AWG so far.
« Last Edit: March 17, 2024, 05:48:58 pm by tszaboo »
 
The following users thanked this post: Performa01, egonotto, 2N3055, KungFuJosh, Martin72

Offline Performa01Topic starter

  • Super Contributor
  • ***
  • Posts: 1686
  • Country: at
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #243 on: March 17, 2024, 05:57:09 pm »
Well, the -3dB bandwidth of a rectangular window is always about 0.884*Δf, which is about 54 Hz for the given parameters.

The root cause why you measure a horribly wrong value of 6.4 Hz is the linear interpolation between adjacent frequency points, which is not appropriate here. With a proper sinc interpolation you would see the true shape of the window function's frequency response and measure the correct bandwidth. In the time domain, you would not sample a sine wave signal with (say) 2.3 points/period either, interpolate the samples linearly and expect that you can still recognize it as sine wave. But the same happens here in the frequency domain. The scope fools you by connecting the frequency points (which are correct) with straight lines. It is important to be aware and not to trust these straight lines blindly.
Thank you for pointing this out. Something isn’t quite correct and I will definitely look into this and try to come up with a test that documents the practical usefulness of the various window functions and also passes under the strict eyes of our Guru “gf” at the same time 😉
 
The following users thanked this post: egonotto, 2N3055, Mortymore, pdenisowski, KungFuJosh, Martin72, core

Offline electronics hobbyist

  • Regular Contributor
  • *
  • Posts: 157
  • Country: cn
    • sds800x-hd-review-demonstration-thread
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #244 on: March 19, 2024, 08:57:42 am »
Thanks a million to professor Performa01.

This review book is so great, so I translated it into Chinese and published it on Zhihu (similar to Quora), and I have only completed 20% so far.
The main purpose is to learn for oneself, and it would be even better if other people could also learn about these oscilloscope knowledge.
If you feel it's not suitable, I will hide it.   ^-^
Thank you again.

The following English image is translated from Chinese to English through Google.
« Last Edit: March 19, 2024, 09:03:23 am by electronics hobbyist »
 
The following users thanked this post: Performa01, egonotto, Mortymore

Offline Performa01Topic starter

  • Super Contributor
  • ***
  • Posts: 1686
  • Country: at
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #245 on: March 19, 2024, 11:52:24 am »
I have revised the FFT-Window article, which should now convey my points without major inaccuracies.

You should really think about converting this material into a (PDF?) booklet -- an "Application Guide to Digital Oscilloscopes" or such. Maybe something Siglent could fund, and distribute as either a separate document or merge into the user manuals?

As a short-term step, would it make sense to reorganize the "table of contents" in the initial post? It is currently organized chronologically. But since you have re-visited some topics over time, and the total amount of material is becoming too much to grasp even the whole TOC at a glance -- could the links to the individual posts be grouped into a few larger topic areas? I am not sure what that meta-structure would best look like. Maybe start with basic horizontal and vertical settings and performance, move on to measurements/cursors/math, then specific application areas?
Thank you very much! I have planned to make this available as a PDF document sooner or later. In fact, I’m already working on it and this will of course be better organized. Since I don’t want to create updates all the time, I’ll wait until I’ve got the feeling that I’ve reached a consistency point, the inspiration for new articles fades a little, and this thread has settled a bit.


EDIT: forgot to answer the 2nd part of your suggestion:

Regarding the reorganization of the current TOC, this makes not much sense, because I’ve initially reserved a number of posts where I cramped as much content as possible into every single posting, and these cannot be grouped in a different way anymore. The idea was to have all the information all in one place so that it can be found easily.

When I ran out of reserved postings, I realized that it doesn’t matter much where the postings are within the thread, as long as we have an overview with that table of content in the OP.
« Last Edit: March 19, 2024, 02:43:31 pm by Performa01 »
 

Offline gf

  • Super Contributor
  • ***
  • Posts: 1302
  • Country: de
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #246 on: March 19, 2024, 01:34:05 pm »
An FFT is quite different to a classic continuously swept spectrum analyzer. If we feed a stable sine wave into a swept analyzer, we’ll get the frequency response plot of the RBW-filter.

The FFT doesn’t plot a filter response, but simply gives an output, which of course should ideally look like an isolated square, one frequency bin wide and the height depending of the amplitude.

Sorry if I disagree again, with this statement. There is not really a difference. Like the SA, the FFT also shows the filter resonse, centered at the frequency of the sine wave stimulus frequency. But the FFT does not plot the continuous filter resonse, but it samples the filter resonse at discrete frequency points (-N/2...N/2-1)/N*sample_rate. And the result of the FFT are only the frequency domain samples. And that's where all the confusion comes from. Mentally we are not able to reconstruct the  continuous filter response from the samples (although the reconstruction can be done mathematically), and connecting the samples with straight lines gives a wrong impression either.

The blue curve in the attachments is the continuous frequency response of a rectangular window function. And the three plots also show the resulting samples, when the continuous curve is sampled at three different frequency offsets. Note that the samples look quite different, although the shape of the continuous curve is the same. As said, from each set of samples, the continuous (blue) curve can be reconstructed - but not "mentally" :scared:

EDIT: The samples at the bottom of figure1.png are actually located at -Inf dB, since they hit the zeros of the frequency response. For clarity, I just wanted to keep them inside the plot. In practice they would be at the noise floor.
« Last Edit: March 19, 2024, 04:21:52 pm by gf »
 
The following users thanked this post: Performa01, egonotto, newbrain, Mortymore, pdenisowski, core, orzel

Offline Performa01Topic starter

  • Super Contributor
  • ***
  • Posts: 1686
  • Country: at
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #247 on: March 19, 2024, 03:11:44 pm »
An FFT is quite different to a classic continuously swept spectrum analyzer. If we feed a stable sine wave into a swept analyzer, we’ll get the frequency response plot of the RBW-filter.

The FFT doesn’t plot a filter response, but simply gives an output, which of course should ideally look like an isolated square, one frequency bin wide and the height depending of the amplitude.

Sorry if I disagree again, with this statement. There is not really a difference. Like the SA, the FFT also shows the filter resonse, centered at the frequency of the sine wave stimulus frequency. But the FFT does not plot the continuous filter resonse, but it samples the filter resonse at discrete frequency points (-N/2...N/2-1)/N*sample_rate. And the result of the FFT are only the frequency domain samples. And that's where all the confusion comes from. Mentally we are not able to reconstruct the  continuous filter response from the samples (although the reconstruction can be done mathematically), and connecting the samples with straight lines gives a wrong impression either.
Thanks again for being alert – I’ve rephrased this paragraph one more time and now tried to describe the reality properly instead of just making a sloppy statement…
 
The following users thanked this post: egonotto, 2N3055, Mortymore

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 4130
  • Country: fi
  • Born in Finland with DLL21 in hand
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #248 on: March 20, 2024, 12:28:55 pm »
Here measured Rectangular window  -3dB BW.

Based to Keysight,  Uniform window -3dB width is 0.8844 x Δf


Input 125kHz (128xΔf), dBm (external 50 ohm).    (f input matched with oscilloscope reference)


Input freq rised up 0.8844 x Δf/2
Result as expected -3dB drop. (lack of enough level measurement resolution)

As can see oscilloscope FFT display also RBW. But what kind of "RBW" it is? Least it is not main lobe -3dB BW. 

(Mostly least I think RBW as normal spectrum analyzer RBW)

There is  Δf 976.56Hz and then RBW 966.80Hz. This looks like based to 0.99
But perhaps  863.7 Hz is more right value for RBW(3dB)?

Rectangular (aka Uniform, aka Boxcar) window is nice for noise measurements because its noise BW is just Δf.
BEV of course. Cars with smoke exhaust pipes - go to museum. In Finland quite all electric power is made using nuclear, wind, solar and water.

Wises must compel the mad barbarians to stop their crimes against humanity. Where have the (strong)wises gone?
 
The following users thanked this post: egonotto, RAPo

Offline Performa01Topic starter

  • Super Contributor
  • ***
  • Posts: 1686
  • Country: at
Re: Siglent SDS800X HD Review & Demonstration Thread
« Reply #249 on: March 31, 2024, 10:30:32 am »
I have promised to provide a condensed PDF version of my Review & Demonstration articles, and today I present a revised document which also includes additional content and a complete list of bookmarks for easy navigation.

Due to space restrictions for forum attachments, I had to split this into two parts. Here comes the first part, including the following main topics:

Document History
Introduction
Basic Information

Acquisition
Display
Trigger
Measure
Cursors
Math
Analysis
Probes
« Last Edit: March 31, 2024, 10:35:33 am by Performa01 »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf