Rigol 5074 vs Siglent SDS2104X Plus vs Keysight DSOX1204G

[2N3055]

Do this with these Rigol + Migsig together. Alone they can not, not even close and together if you have purchased both because can get with same price as one Siglent.

Zooming out is much more useful feature. But the Siglent can't do that. The Anritsu network analyser I bought for $400 OTOH can also tell me the equivalent circuit parameters to actually model the crystal.

Are you serious?
Again with this... You have problems.. Really...
You cannot go more than 2 weeks without "I need zoom out, please let me have zoom out... Seriously, man just one zoom out please.. I need it.."
 
And no, it is marginally useful thing very few use on daily basis.

And whats up with " Well, peasants should buy VNA if they want to measure something"...

There is need for good frequency resolution in FRA apart from crystals. Crystal is just used to to show good resolution and dynamic range...

[nctnico]

Do this with these Rigol + Migsig together. Alone they can not, not even close and together if you have purchased both because can get with same price as one Siglent.

Zooming out is much more useful feature. But the Siglent can't do that. The Anritsu network analyser I bought for $400 OTOH can also tell me the equivalent circuit parameters to actually model the crystal.

Are you serious?
You cannot go more than 2 weeks without "I need zoom out, please let me have zoom out... Seriously, man just one zoom out please.. I need it.."

Just like rf-loop can't go without a day of 'history' and 'FRA'... just providing balance. Fix your sarcasm detector; it is broken 

And whats up with " Well, peasants should buy VNA if they want to measure something"...

There is need for good frequency resolution in FRA apart from crystals. Crystal is just used to to show good resolution and dynamic range...

If you need to measure frequency responses then buy the NanoVNA. Much more useful than FRA on an oscilloscope. Goes from 10kHz to 1.5GHz which a large dynamic range for about $100. The NanoVNA really is a nice bit of kit to have.

[2N3055]

If you need to measure frequency responses then buy the NanoVNA. Much more useful than FRA on an oscilloscope. Goes from 10kHz to 1.5GHz which a large dynamic range for about $100. The NanoVNA really is a nice bit of kit to have.

FRA serves mostly frequency range from almost DC to 10s of MHz. Also it is used on non 50 Ohm systems.
NanoVNA is amazing little device that anybody that can afford it should have. Absolutely agree. But serves slightly different purpose than FRA.
Same as FFT on scope serves different purpose than real SA...

[Michael YYZ]

What is this issue about zooming out all about? My SDS2104X Plus is on its way, so I’m not familiar with it, but the way I understand things is one could capture a waveform over a longer time domain - i.e. zoomed out - and then zoom in to see the details. In other words, the scope captures only what’s on the screen, and not beyond. What’s the problem with this approach?

Are you serious?
Again with this... You have problems.. Really...
You cannot go more than 2 weeks without "I need zoom out, please let me have zoom out... Seriously, man just one zoom out please.. I need it.."
 
And no, it is marginally useful thing very few use on daily basis.

And whats up with " Well, peasants should buy VNA if they want to measure something"...

There is need for good frequency resolution in FRA apart from crystals. Crystal is just used to to show good resolution and dynamic range...

[nctnico]

What is this issue about zooming out all about? My SDS2104X Plus is on its way, so I’m not familiar with it, but the way I understand things is one could capture a waveform over a longer time domain - i.e. zoomed out - and then zoom in to see the details. In other words, the scope captures only what’s on the screen, and not beyond. What’s the problem with this approach?

It hinders some workflows. Here is the long thread about it (with several people not understanding the problem at all and causing lots of noise nevertheless):
https://www.eevblog.com/forum/testgear/oscilloscope-zoom-out-quirk/

Most scopes can zoom out so there probably is a good reason to do so. Dave tested many oscilloscopes and found out Siglent is the one of the two brands (the other is Lecroy which is more geared towards signal analysis anyway) which can't:


Needless to say an oscilloscope which can't zoom out is a hard pass for me. It takes too much fiddling of knobs and restrictions (=slow and inefficient workflow) to make it use the amount of memory I want it to use. I need to be able to set the memory length in order to do certain measurements quickly & efficiently. Been there, done that.

However, it if perfectly fine if it doesn't bother you. Everyone has different priorities where it comes to oscilloscope features.

[2N3055]

What is this issue about zooming out all about? My SDS2104X Plus is on its way, so I’m not familiar with it, but the way I understand things is one could capture a waveform over a longer time domain - i.e. zoomed out - and then zoom in to see the details. In other words, the scope captures only what’s on the screen, and not beyond. What’s the problem with this approach?

Are you serious?
Again with this... You have problems.. Really...
You cannot go more than 2 weeks without "I need zoom out, please let me have zoom out... Seriously, man just one zoom out please.. I need it.."
 
And no, it is marginally useful thing very few use on daily basis.

And whats up with " Well, peasants should buy VNA if they want to measure something"...

There is need for good frequency resolution in FRA apart from crystals. Crystal is just used to to show good resolution and dynamic range...

Scope captures time interval that is same as screen time interval. So 10 ms/DIV gives 100 ms worth of data. Screen is WISIWIG "map" to your FULL acquired buffer.
If you set a scope to 100MS, you will get 100E6 points in that interval. Screen will do it's best make a visible representation of that FULL time scale. On such an acquired buffer, you can now use zoom to show magnified portion of that buffer, with secondary timebase, or you can just change timebase and horizontal position and look around.

So it is something that is very important to Nico (he refuses to use zoom mode because he has his reasons), but not really a problem to most of other people... I simply use zoom mode and consider it better, easier and more intuitive to work. And he keeps repeating it is easier and less button fiddling, despite being proven that is not true. At one point we literally counted steps and clicks and proved his way is more work and setup. But he keeps broadcasting his alternate reality..

Anyways, from what you wrote, I don't think you will have any problem using scope without "Nico's special".

[nctnico]

At one point we literally counted steps and clicks and proved his way is more work and setup.

No you didn't! It is clear you work on completely different measurement tasks compared to me. That is totally fine. Again, respect other people's priorities! Bashing or even crusading against (yes, you got to that point!) other people's opinions just reflects badly onto yourself. You can't even let a tongue in cheek comment go -my response to rf-loop bringing up FRA for the umpteenth time while it is not even the subject-. Just let it rest.

[2N3055]

It hinders some workflows. Here is the long thread about it (with several people not understanding the problem at all and causing lots of noise nevertheless):

Most scopes can zoom out so there probably is a good reason to do so. Dave tested many oscilloscopes and found out Siglent is the one of the two brands (the other is Lecroy which is more geared towards signal analysis anyway) which can't:

I hinders your some of your workflows... Thanks for sharing..  For 3432nd time..

And other statement and Dave's video are incorrect. He keeps repeating Keysight captures outside screen, when it is proven thing, documented by Keysight, it ain't so.
If anything, Keysight Megazoom scopes are so praised because of brutal auto control and no user settings whatsoever, because it allows for such speeds.
Fact that Keysight uses many tricks ( a separate Single acquisition in full length after you press Stop to fake full buffer capture) should not confuse anybody.
Other parts of that video are not Dave's best work, he wanted to capitalize on controversy and inflammatory note on forum discussion. He lives by making videos.
It makes sense for him to exploit such themes. His research was, well, non existent, using Keysight as example. He simply partially took sides, and moved on.

Siglent scopes (Siglent is OEM for simpler LeCroy lines) are inspired by LeCroy concept. They took same approach.
LeCroy are pure digital scopes. They dont pretend to be CRT emulation, like Keysight. And you might like it or not. Or maybe understand it or not? I don't know.
I personally like, respect and use both concepts.. Tool for the job.

And as much I support your freedom to use instruments you bought with your money any way you want, stating that some instrument is fundamentally flawed because you can not use some arcane workflow that you like because you don't want to use zoom for pretty much highly personal reasons, is simply a buffoonery. It is grandiose overblowing out of proportion of something that is a difference and artificially engineering arcane scenarios that make it look like a defect or serious problem. It isn't. Not in any universe. Except your's.

If you were to point out to every beginner, the fact that there a different types of scopes, and that some scopes can do this and some can do this, I would never replied to you on any of these postings.
But you literally demonize Lecroy, Siglent, and any other manufacturer (like Picoscope) that uses that same architecture, proclaiming it is  stupid, useless, failed, unusable, something you would never buy, how bad it is. And in meantime, millions of happy users worldwide, create most fantastic products with those, send people to the space etc. People working on stuff beyond your dreams are not only capable of doing great things with those, but are actually full of praise.

So excuse me if I find your parroting of hate to something that is different from what you think it should be, quite tiresome.

[2N3055]

At one point we literally counted steps and clicks and proved his way is more work and setup.

No you didn't! It is clear you work on completely different measurement tasks compared to me. That is totally fine. Again, respect other people's priorities! Bashing or even crusading against (yes, you got to that point!) other people's opinions just reflects badly onto yourself. You can't even let a tongue in cheek comment go -my response to rf-loop bringing up FRA for the umpteenth time while it is not even the subject-. Just let it rest.

Yes we did, Someone and myself did some time ago. But you obviously didn't catch that for some reason.

Go back and count and let me know to how many people I respond like this. I 'll give you a hint, i had quite a few words with Rf-loop at beginning. And more heated than with you.
Since then I learned he has trouble expressing himself in English, so sometimes he comes off harsh even without that intention. Also, very experienced old school engineer..Like you and me, he likes repeating stuff he thinks is important. Also he is quite an expert in FRA related matters, so yeah, he likes talking about that..  I know.. But it is always to the point.
He is also very critical of Siglent when he think they deserve it and has pointed out many things he found suboptimal, publicly and without pulling punches.
So I learned to respect the hell out of him.
And also go back a count how many times I called you an excellent engineer, smart and how many times I agreed with you and stated I learned something from you...

So no, it's not black and white.. We disagree.. That's life, to be honest, I would rather argue with you than speak with people without any will and stance of their own.
This way I always learn something, be it technology, human nature or whatever..

On that note, stay safe and healthy, so we can argue until old age...

Sinisa

[Michael YYZ]

I get the point about the zoom out feature, or lack thereof for Siglent, but I don’t understand a never-ending debate on a topic which appears relatively minor.

I am an electronics hobbyist. I am also an engineer and have a PhD degree, but in a different field. I’ve done an enormous amount of research on current oscilloscope models and eventually I closed in on the Siglent for offering one of the best performance vs. price ratio available.

Like for any other devices, the designers of a scope have to make decisions, compromises and accept trade offs for the design decisions that they make. The resulting features or lack of them may work perfectly fine for some and not at all for others.

I had watched Dave’s video and had read tens and tens of pages of forums and reviews before I decided to get the Siglent. I believe I understand the no-zoom-out characteristic of this scope, which I would not call it “a flaw”. It should, I reckon, work perfectly fine for me.

The good thing is we live in a world where we’ve got choice. Lots of it! That’s why comparison threads like this exist. They help us decide, and learning from others’ expertise is a tremendous support. But one should simply point out one’s likes and dislikes, and then move on. We should not become a mob of fanbois and hategals. It’s perfectly understandable that some characteristics of Siglent scopes may not work for some. It is good and very appreciated to be told why, so we’ve got complete information. But that’s it, there’s absolutely no reason to keep pounding sand on this topic relentlessly. If one’s happy with Siglent, go and have fun with it. If you don’t like it, stop bickering and go and use another type. In the meantime, let’s help each other!

I surely am looking forward to enjoying mine!...

[Fungus]

Zooming out is much more useful feature. But the Siglent can't do that very basic operation.

Yep. If I count the number of Bode plots I've done in my life vs. the number of times I've zoomed out on captured data, the zoom wins.

Siglents don't do that most basic thing.

[2N3055]

Zooming out is much more useful feature. But the Siglent can't do that very basic operation.

Yep. If I count the number of Bode plots I've done in my life vs. the number of times I've zoomed out on captured data, the zoom wins.

Siglents don't do that most basic thing.

Well, as I said, that is more of a testament of your (wrong?) scoping technique and the fact that you don't do anything that would need FRA.

That excludes all kinds of audio filters, tone controls, audio crossovers, active and passive, control loops that include anything that works on negative feedback principle (be it PID controllers, power supplies) etc etc... Also any circuit that uses opamps or any kind of amplifier that you would want to check gain/phase for any reason (stability or simply to verify specs..). You can also check IF parts of the radios (455kHz and 10,7MHz), check filters etc... Test small signal and audio transformers characteristics...
It can also be used to test CMRR and PSRR (with adequate accessories). List is huge..

There is no scenario where those are less important than ability to set capture length by side effect of setting buffer sizes manually because there are people that like that instead of what says in a manual..

[Fungus]

Well, as I said, that is more of a testament of your (wrong?) scoping technique

I'm holding it wrong?

[2N3055]

Well, as I said, that is more of a testament of your (wrong?) scoping technique

I'm holding it wrong?

  

[Performa01]

Do this with these Rigol + Migsig together.

Rigol has bode plot all by itself:

https://int.rigol.com/products/oscillosopes/mso5000.html

Impressive! 

The vertical grid is evenly spaced, isn't it? How comes that we get 5 dB/div in the bottom row, then 4 dB /div above and finally 3 dB/div around zero? Generally, numbering on both amplitude and phase axes is all over the place.

Doesn't look very professional, does it?

RF-Loop showed an example with narrow frequency span and a dynamic range of 100 dB. Do you really think that a screenshot with severely flawed axis numbering, showing a dynamic range of just about 25 dB will excite anyone but the Rigol fanbois?

[Fungus]

The vertical grid is evenly spaced, isn't it? How comes that we get 5 dB/div in the bottom row, then 4 dB /div above

If I had to guess I'd say the numbers have fractions and they're using a round() function to display them.

I don't own one but maybe somebody who does can confirm if the cursor measurements show a better value.

[Performa01]

Thanks for the helpful info above.

Just to try to tie it off, roughly what difference (quantitatively in terms of mV, uV, or % and or/qualitatively) would you expect when measuring ~1mVRMS ripple with a 20MHz bandwidth filter assuming good probing technique (and really measuring the PS's performance rather than noise in the environment) when comparing a Rigol 2000 vs a Rigol 1000, and a Rigol 2000 vs a Siglent 2k Plus?  (Is the Rigol 2k front end going to have some clear advantage over the other two for such a measurement, or are the results going to be barely noticeable?)

 1mV P-P, 138uV AC RMS signal fed into scope (just an AWG signal that resembles superficially something you might see on the switcher output):

Keysight MSOX3104T with 20MHz bandlimit (50Ohm, so lower noise than with 1Meg input) is a good approximation of MSO5000. It has a bit lower noise but has same software magnified 5mV/div for 1mv/div. So MSO5000 would be like this, and slightly worse..

Note errors in measurements in Keysight. Which otherwise has excellent measurement implementation.
And Keysight couldn't get stable trigger. Snapshots were single, manual force triggered.

And a Micsig STO1104E ... Real 1mV/div, less than 65 uV AC RMS noise (20 MHz bandlimited). Triggering and all...

And on a scope with real 500uV/div it would be even better than on Micsig. Trust me on that. I tried. It is amazing how clean it looked.

Since the TO considered an SDS2000X+, here is the corresponding noise demonstration, see attached screenshots.

ARB Signal 1 mVpp, 120 µV AC-RMS, same timebase as yours: 20 µs/div. We don't even need 500 µV/div...

Signal at 1 mV/div and 20 MHz BW limit can be triggered and easily stands out of the noise.
Measurements aren't too far off (especially considering the signal is only about 1 division high) : 1.25 mVpp, 131 µV AC-RMS.
SDS2354X+ Sig BW20M.png

Base noise at full bandwidth (~580 MHz) is nearly as low as on your Micsig with 20 M bandwidth limit: 77 µV AC-RMS.
SDS2354X+ Noise BW500M

With 20 MHz bandwidth limit, the noise drops down to <27 µV AC-RMS.
SDS2354X+ Noise BW20M

[nctnico]

It hinders some workflows. Here is the long thread about it (with several people not understanding the problem at all and causing lots of noise nevertheless):

Most scopes can zoom out so there probably is a good reason to do so. Dave tested many oscilloscopes and found out Siglent is the one of the two brands (the other is Lecroy which is more geared towards signal analysis anyway) which can't:

I hinders your some of your workflows... Thanks for sharing..  For 3432nd time..

So excuse me if I find your parroting of hate to something that is different from what you think it should be, quite tiresome.

Geez...  Let's not count the number of times you brought up the R&S scopes not being able to search on digital decoding. It is likely you brought that up more often than I brought up zooming out. To me search is not a very important feature (I use the trigger setup to filter unwanted messages which then also makes a search futile because the search options usually aren't more elaborate than the trigger settings) but I'm not going to crusade or ridicule you about it.

[nctnico]

Zooming out is much more useful feature. But the Siglent can't do that very basic operation.

Yep. If I count the number of Bode plots I've done in my life vs. the number of times I've zoomed out on captured data, the zoom wins.

Siglents don't do that most basic thing.

Well, as I said, that is more of a testament of your (wrong?) scoping technique and the fact that you don't do anything that would need FRA.

That excludes all kinds of audio filters, tone controls, audio crossovers, active and passive, control loops that include anything that works on negative feedback principle (be it PID controllers, power supplies) etc etc... Also any circuit that uses opamps or any kind of amplifier that you would want to check gain/phase for any reason (stability or simply to verify specs..). You can also check IF parts of the radios (455kHz and 10,7MHz), check filters etc... Test small signal and audio transformers characteristics...
It can also be used to test CMRR and PSRR (with adequate accessories). List is huge..

People for which such measurements are important usually have a VNA and LCR meter. FRA gets you a crude amplitude / phase graph which you can also get from sweeping a function generator. Add some math and you can use the cursors to read magnitude in dB(m/V/uV) and phase if you like. FRA is nice but it runs out of steam quickly. For example if you want to know impedances, Q factor, equivalent component values (RL  / RC), etc then a VNA or LCR meter becomes a better choice quickly.

[2N3055]

It hinders some workflows. Here is the long thread about it (with several people not understanding the problem at all and causing lots of noise nevertheless):

Most scopes can zoom out so there probably is a good reason to do so. Dave tested many oscilloscopes and found out Siglent is the one of the two brands (the other is Lecroy which is more geared towards signal analysis anyway) which can't:

I hinders your some of your workflows... Thanks for sharing..  For 3432nd time..

So excuse me if I find your parroting of hate to something that is different from what you think it should be, quite tiresome.

Geez...  Let's not count the number of times you brought up the R&S scopes not being able to search on digital decoding. It is likely you brought that up more often than I brought up zooming out. To me search is not a very important feature (I use the trigger setup to filter unwanted messages which then also makes a search futile because the search options usually aren't more elaborate than the trigger settings) but I'm not going to crusade or ridicule you about it.

Again, to you, it is not, because you download stuff to PC and search there.
I'm sorry but, in practice, search is more important than manual buffer control ..
And people are usually genuinely surprised 8000 € scope doesn't have it.
Other, higher end, R&S scopes have it. Keysight 3000/4000/60000 have it, despite having two orders of magnitude less memory.
And I use it all the time. Because I use combination of triggers and search to travel quickly through captures..
It is like second set of trigger like filter on top of basic trigger. And if you have more than 10-20 packets you cannot really find much by visually looking at one by one. You need some search support.
So it is either:
1. You capture, put in USB stick, save decode to stick, eject stick, put it in PC, open file in your favorite coma delimited file editor (Excel?), then search there. Here you either find what you need or have to repeat all of it to capture again with different settings.
2. You capture something, quickly search through it with a scope and reacquire differently if needed. When you're satisfied you have data you want(need), than you can save to stick and open it on PC for further analysis. Or not if you already verified what you needed.

I prefer 2nd option. Many others do too. For those knowing that scope cannot do this is probably important.

[nctnico]

The one time (ONE TIME) I have exported decoded packets to Excel it was about thousands of packets spanning over one hour in time where one packet went wrong for an unknown reason (and I didn't had a way to add an error detection/reporting to the firmware). So where do you even search for? At some point getting data into a PC is just more convenient for various reasons. You can delete the uninteresting data from the set, add annotations, filter (grep), plot a graph, etc. Beyond that I never ever had the need to search for a packet. The main reason is that I use different, more efficient ways/tools (like a CAN analyser, specific events on other signals or integrated protocol error detection in firmware for example) for tasks that otherwise might need searching for particular packets. I do highly value the ability of a DSO to capture a lot of packets just in case anything else fails (last resort).

BTW it is interesting to see your reasoning for finding search important is exactly the same for me finding zooming-out is important.

[2N3055]

Zooming out is much more useful feature. But the Siglent can't do that very basic operation.

Yep. If I count the number of Bode plots I've done in my life vs. the number of times I've zoomed out on captured data, the zoom wins.

Siglents don't do that most basic thing.

Well, as I said, that is more of a testament of your (wrong?) scoping technique and the fact that you don't do anything that would need FRA.

That excludes all kinds of audio filters, tone controls, audio crossovers, active and passive, control loops that include anything that works on negative feedback principle (be it PID controllers, power supplies) etc etc... Also any circuit that uses opamps or any kind of amplifier that you would want to check gain/phase for any reason (stability or simply to verify specs..). You can also check IF parts of the radios (455kHz and 10,7MHz), check filters etc... Test small signal and audio transformers characteristics...
It can also be used to test CMRR and PSRR (with adequate accessories). List is huge..

People for which such measurements are important usually have a VNA and LCR meter. FRA gets you a crude amplitude / phase graph which you can also get from sweeping a function generator. Add some math and you can use the cursors to read magnitude in dB(m/V/uV) and phase if you like. FRA is nice but it runs out of steam quickly. For example if you want to know impedances, Q factor, equivalent component values (RL  / RC), etc then a VNA or LCR meter becomes a better choice quickly.

Well made FRA gives you nice graphs with lots of control. And also nice data tables to calculate and/or plot something later.  And much more control than using sweep generator, that will not show you phase, or scale or frequencies. And FRA is easier to setup, has cursor support, vertical and horizontal scaling and type of plot and units. No need to reinvent the wheel, math is already there.
I agree that well equipped lab should have VNA and LCR meter. Ideally real impedance analyser..
But FRA is not impedance analyzer, nor LCR meter or VNA. It is gain/phase response analyser, and one that fills gap from almost DC to vhere VNA starts to operate. And for input/output impedances different from 50Ohm...

[nctnico]

Well, my Anritsu network analyser works from 10Hz and has switcheable 1MOhm / 50 Ohm inputs so it isn't limited to 50 Ohm and 10Hz is very near DC.

[2N3055]

Well, my Anritsu network analyser works from 10Hz and has switcheable 1MOhm / 50 Ohm inputs so it isn't limited to 50 Ohm and 10Hz is very near DC.

Yes that is very good. What model exactly is that?

[nctnico]

Well, my Anritsu network analyser works from 10Hz and has switcheable 1MOhm / 50 Ohm inputs so it isn't limited to 50 Ohm and 10Hz is very near DC.

Yes that is very good. What model exactly is that?

MS4630B