Rigol 5074 vs Siglent SDS2104X Plus vs Keysight DSOX1204G

[tautech]

Did you watch the comparison video by member Howardlong ?

[Fungus]

Did you watch the comparison video by member Howardlong ?

Yes, but depending on what you work on that might have no effect at all, eg. If you're working on 5V logic all day long.

[rf-loop]

Did you watch the comparison video by member Howardlong ?

It is fun but I have not seen so much comments about things what happen when starting from video there is continuously running this 1kHz sinewaves.
Look carefully this Rigol screen... first I thing this is just camera reason etc but when compare with eyes Siglent and Rigol live screen carefully... why this Rigol have continuous horizontal jitter. It is horrible. Same time Siglent is rock solid, just as still image.
Also even with this signal level noise in Rigol is really visible.


But then other things..

If you have ever worked on repairs you would realize your hands get dirty and touch screens and dirt don’t go well together.

If your hands are really dirty you do not need even touch scope.
Just use mouse. (SDS2000X Plus)

Do not go fooled with this samplerate. In this case it is least partially marketing trick and nothing more in practice. In camera world there is pixels and pixels... in scope world there is samples and samples.
In scope, look also Effective bits:.

Here @Howardlong test what is quite well made afaik.

Image from https://www.eevblog.com/forum/testgear/benefits-of-going-with-all-siglent-setup/msg3281580/#msg3281580

8GSa/s scope... barely 5 bits scope.

Also max memory when option activated is 200M  it is divided between all channels. One channel get max 200M, two 100M and four 50
But it need also think that many times we are interested about how long capture we can get when we think time, not amount of samples.
If our thinking is based to time, 200M with 8GSa/s equals 50M with 2GSa/s

Siglent have 2x 200M and 2x 2GSa/s
Rigol have 1x 200M and 1x 8Gsa

Think worst case
4 channels on.
Rigol 2GSa/s  and 50M / channel
Siglent 1Gsa/s and 100M / channel. You get 4 times more max capture length in time.

Then Rigol claim they have up to 12 bit resolution in HiRes mode... how about Effective bits...
After then... also. What do with these 12 bits. Just nothing but  bit less noise in image. How you get these 12 bit out from scope, do you have 12 bit data. No, you have not.

How about Siglent 10 bit mode. It beats this Rigol 12 bit "HigRes" mode hands down AND it also give 10Bit data to you. Of course it also eat more memory, in 10bit mode every data point takes two bytes from memory, naturally.

Rigol have 1M FFT Siglent 2M.  But what you do also need is dynamic range. Where is Rigol noise floor, where is Siglent noise floor.

Then just for fun because it is so fun... this Rigol BodePlot joke what is nothing but one featuire more in salesmens brochure. It is toy. It is nothing with this real tool what is in SDS2000Xplus. No need here explain details... up to 300 pints / decade when siglent have 1+3 channels FRA with up to 275204396/decade (yes if dirty salesman tell it, but still he do not lie because Siglent can use 500points also minimum span 500Hz and this can do up to 120MHz. From 119.9995MHz to 120.0000MHz sweep with 500 data pints is just this told amount per Decade.)
Naturally not all peoples need FRA or FFT.

Then, you are looking Rigol screen and you note something weird just happen. You stop scope... but of course you are bit late.
With Siglent, depending settings, you stop scope and look history.

Naturally, all depends just one individual user needs, there is not totally common better, but perhaps there can find totally common shits but these I do not name at all. Rigol is better to one and Siglent is better to one other. Keysight if for one and Tek is for other, also LeCroy, also R&S also ZLG. All these make good scopes and if go more up in needs (and pocket) still there is LeCroy, Tek, KS and R&S, every brand and model is just best for one. There is not universal best, unlike long time ago when there was only three best names, Tektronix, Tektronix, Tektronix. I have started touching and using oscilloscopes in these  decades. Today I know there is many good scopes. Everyone can find personally best scope for his needs and his budget.

[TurboTom]

...and here we go again with the table that's been faulted by the oh-so-frequent DC offset mistake when measuring Vrms, as has been discussed starting from here.

Simply speaking: Yes, the MSO5000 is considerably more noisy than the SDS2000X. But it's not as bad as shown in the table. As it's been told before, the more noisy front-end will be a showstopper for one while it's a minor nuisance for another, depending on the application.

For my own part, if I was in the market for a new scope in this class right now, the price would make the difference: If the "R" was within budget while I would have to stretch to get the "S", "R" would be it. If "S" is well within reach, I'ld go for this one!

[Fungus]

Simply speaking: Yes, the MSO5000 is considerably more noisy than the SDS2000X. But it's not as bad as shown in the table. As it's been told before, the more noisy front-end will be a showstopper for one while it's a minor nuisance for another, depending on the application.

And if you really need to see low-level signals get yourself a signal amplifier and see them properly on any 'scope. A low noise front end will only take you so far.

[nctnico]

Simply speaking: Yes, the MSO5000 is considerably more noisy than the SDS2000X. But it's not as bad as shown in the table. As it's been told before, the more noisy front-end will be a showstopper for one while it's a minor nuisance for another, depending on the application.

And if you really need to see low-level signals get yourself a signal amplifier and see them properly on any 'scope. A low noise front end will only take you so far.

Low noise is not just good for low level signals but for all kinds of signals. A big fat trace on the screen is hard to put a cursor on.

[normi]

I do admit that each scope has its strong points and week points, some more than others. My point is cost is often times the biggest factor when choosing a scope, if it were not we would be talking about Tektronix and Keysight not Siglent and Rigol. If the cheaper scope does not suit your needs, then money becomes the lesser factor and you spend more to get more.

Based on the early reviews by Dave before firmware updates the Rigol does have numerous flaws, and Howardlong's review also points outs some remaining issues. Most of Daves issues appear to have been fixed and had the scope been reviewed after firmware update, it would have looked good. Even in the video which I had watched multiple times before my original post, Howardlong said it was a hard decision to make and he had both scopes side by side. If the difference was so huge he would have quickly made a choice. Dave also said if the MSO5000 was hacked, they couldn't keep it on the shelf. So if we were to conclude that the Siglent has less issues, would customers feel the difference is worth spending much more. If we were discussing the 100Mhz versions then cost would not be an issue since they would be similarly priced.

rf-loop you have a lot of knowledge about the Siglent, I am not finding many persons with that similar knowledge and support for the Rigol, and this makes checking claims less easy. 

[2N3055]

Simply speaking: Yes, the MSO5000 is considerably more noisy than the SDS2000X. But it's not as bad as shown in the table. As it's been told before, the more noisy front-end will be a showstopper for one while it's a minor nuisance for another, depending on the application.

And if you really need to see low-level signals get yourself a signal amplifier and see them properly on any 'scope. A low noise front end will only take you so far.

You have a source for 350Mhz bandwidth 10x low noise amplifiers that cost less than price difference?
Cause that would be awesome...

[IAmBack]

Low noise front-end is important also because almost all measurements are done with 10x probes and requires extra front-end gain. My 2 cents. Of course, if we have active probes...

[tautech]

Low noise front-end is important also because almost all measurements are done with 10x probes and requires extra front-end gain.

So simple and so often overlooked.

[Fungus]

And if you really need to see low-level signals get yourself a signal amplifier and see them properly on any 'scope. A low noise front end will only take you so far.

You have a source for 350Mhz bandwidth 10x low noise amplifiers that cost less than price difference?
Cause that would be awesome...

Sure, just search for "RF signal amplifier" on Aliexpress. Or build your own (see video below).

When did you become price sensitive anyway? You're happy recommitting the Siglent as if it costs exactly the same as the Rigol.

I do admit that each scope has its strong points and week points, some more than others. My point is cost is often times the biggest factor when choosing a scope, if it were not we would be talking about Tektronix and Keysight not Siglent and Rigol.

Yep. It's turtles all the way up. Why stop at the Siglent? eg. There's some tasty R&S 'scopes with genuine 10-bit DACs out there that are a real pleasure to use. They're only a few hundred dollars more!

Wait... I know why it stops at Siglent - because there's only Siglent salesmen hanging around these forums. There's no R&S, Tek or Keysight salesmen here.

Bottom line: Buy the 'scope that fits your needs. The Rigol offers plenty of bang per buck.

[2N3055]

DC-350 MHz, 10x low noise amplifier with less than 1% gain error at DC and better than 1 dB amplitude flatness up to -3dB at 350MHz.
With switchable 50 Ohm/1MOhm/15pf oscilloscope like input.
For pennies on Ali...
As they say: " I am confusion, Ameriga explain!"

And I didn't say an iota about Siglent.
And I remember fondly my DS1074Z. Best spent money ever, those few years ago.
Sweet thing, game changer back then, changed the market...
But world marches on, new stuff comes.

I disagree that low noise, 500uV/div is irrelevant and that any scope with higher noise and lower sensitivity is equal.
It's not, it is worse, and for many uses it is important difference.
If you are doing only digital stuff, then you might not need it.
And no, you cannot buy some cheap preampf from Ali (or make one with 10kHz bandwith) and pretend it is same as real scope front end. Not by a mile. Designing scope front ends is pretty much the holly grail of industry, and is hard.

Sad thing is that few years ago, Rigol made DS2000A series that had excellent, high sensitivity, low noise, front end. 5000 would have been waay better scope if they used that front end with new A/D instead of integrated front end chip whose design goal was 4 GHz bandwidth...
Paired with A/D that was optimized for 4 GHz bandwidth. Quite frankly, it is not even that noisy, when you think about it.

That front end chip paired with their A/D makes sense for the DS8000 series. For a 4GHz scope that one has respectable specs.
Maybe they will make second generation that will have improved specs, and bandwidth filtering done in a way that pulls the noise down proportionally with bandwidth reduction. I don't know.

But, while MSO5000 noise and sensitivity specs are not catastrophic, they are very not good either.  So for low level work, there are better choices. And to clarify, for that kind of work I wouldn't recommend Keysight 3000A/T series either, or any of the Infiniivision scopes.
Up to 100 MHz, I actually use my Micsig STO1104E when I need to see something "in a grass".

[Fungus]

I disagree that low noise, 500uV/div is irrelevant and that any scope with higher noise and lower sensitivity is equal.

OK, not irrelevant.

Neither is perfect, a low noise front end paired with an amplifier would be even better. 

Up to 100 MHz, I actually use my Micsig STO1104E when I need to see something "in a grass".

Me too. 

[Electro Fan]

I disagree that low noise, 500uV/div is irrelevant and that any scope with higher noise and lower sensitivity is equal.
It's not, it is worse, and for many uses it is important difference.
If you are doing only digital stuff, then you might not need it.
And no, you cannot buy some cheap preampf from Ali (or make one with 10kHz bandwith) and pretend it is same as real scope front end. Not by a mile. Designing scope front ends is pretty much the holly grail of industry, and is hard.

Sad thing is that few years ago, Rigol made DS2000A series that had excellent, high sensitivity, low noise, front end.

2N3055, a little off topic but some questions for you, plz/Thx.

If you were looking at power supply ripple in the range of roughly 1-2mVRMS or 3-4mV Pk-Pk, would you expect to see any difference in the measuring or display rendering capability of the the Rigol 2k vs the Rigol 1k vs the Rigol 5k vs the Siglent 2k Plus?  If not in this range, at what level of ripple or other noise would you expect to see the front end capabilities of the Rigol DSO2000A stand out, or in what use case and at what measurement levels would you expect to see differences in these models' ability to accurately discern, measure, and display noise?  Thx

[2N3055]

I disagree that low noise, 500uV/div is irrelevant and that any scope with higher noise and lower sensitivity is equal.
It's not, it is worse, and for many uses it is important difference.
If you are doing only digital stuff, then you might not need it.
And no, you cannot buy some cheap preampf from Ali (or make one with 10kHz bandwith) and pretend it is same as real scope front end. Not by a mile. Designing scope front ends is pretty much the holly grail of industry, and is hard.

Sad thing is that few years ago, Rigol made DS2000A series that had excellent, high sensitivity, low noise, front end.

2N3055, a little off topic but some questions for you, plz/Thx.

If you were looking at power supply ripple in the range of roughly 1-2mVRMS or 3-4mV Pk-Pk, would you expect to see any difference in the measuring or display rendering capability of the the Rigol 2k vs the Rigol 1k vs the Rigol 5k vs the Siglent 2k Plus?  If not in this range, at what level of ripple or other noise would you expect to see the front end capabilities of the Rigol DSO2000A stand out, or in what use case and at what measurement levels would you expect to see differences in these models' ability to accurately discern, measure, and display noise?  Thx

Off topic but quickly.. Yes, that would be roughly the point where you would see the difference.. Make sure to use 20Mhz bandwidth limiter. Also make note scope noise is actually noise(thermal and quantize noise), while what we call "noise" in PSU is a combination of real broadband noise, ripple from power grid (50 Hz and 100Hz, and harmonics), switcher ripple, and various resonances, including instabilities in control loop.

Noise is added as RMS, so to have good results rule of the thumb is that your instrument has to have 3x less noise than DUT to have cca 5% error in noise floor measurements.
This is place where FFT is very useful, especially long FFT, that can achieve good frequency resolution..If you enable averaging, you can extract more data...
That's quickly.

For specialized measurements like measuring low frequency noise on LTZ1000, you will need preamp. For common switchers, you can do good measurements with low noise scope with 500uV/div. For verifying good linear PSU (cca 200-300uV rms) with a scope  that has 80 uV RMS noise at 500 uV/div, you will be able to verify that it roughly passes specs. You won't be able to measure that accurately, but for repair and verification, you would be able to spot something is wrong or not.

One more note and then back to topic.  I know it's been said milion times, but measuring at 500uV/div level on scope is HARD...
You need to make sure you're not measuring everything else in your environment except DUT... good probing is paramount.

[Electro Fan]

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?)

[Fungus]

Up to 100 MHz, I actually use my Micsig STO1104E when I need to see something "in a grass".

Me too. 

And note that you could almost buy a Rigol and a Micsig for the price of a Siglent.

[tautech]

Up to 100 MHz, I actually use my Micsig STO1104E when I need to see something "in a grass".

Me too. 

And note that you could almost buy a Rigol and a Micsig for the price of a Siglent.

And a SDS2kX Plus would be a better instrument than both of them put together !
Get one and you'll see.

[2N3055]

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.

[2N3055]

Up to 100 MHz, I actually use my Micsig STO1104E when I need to see something "in a grass".

Me too. 

And note that you could almost buy a Rigol and a Micsig for the price of a Siglent.

For people that need nothing advanced and are OK with Micsig being very simple but portable, and MSO5000 not very good for low level analog, that might be an option.
SDS2000X+ is next level up, analog performance wise, and capability wise..
Of course not everybody needs that, most people will be happy with DS1054Z or Micsig STO1104 because they only use some capabilites and not doing anything that needs more advanced stuff.

Really all of those scopes are good choice in their price range. But they do go in different price range for a reason.

[jemangedeslolos]

Up to 100 MHz, I actually use my Micsig STO1104E when I need to see something "in a grass".

Me too. 

And note that you could almost buy a Rigol and a Micsig for the price of a Siglent.

And a SDS2kX Plus would be a better instrument than both of them put together !
Get one and you'll see.

I wonder on what argument you claim that ?

[rf-loop]

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.

10kHz sweep and measurement dynamic inside single sweep roughly 100dB. Just ONE small tiny and easy example.
This is just as easy bisquit what do not need even any limits knocking.

[Fungus]

Do this with these Rigol + Migsig together.

Rigol has bode plot all by itself:

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

[2N3055]

By the look of it, it is a crystal measurement...
Quite usable one too..

Rigols BodeWave is quite a simple implementation, and wouldn't do well on this task...
Not that it isn't possible to do better, just Rigol never bothered to make it better..

[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 very basic operation. The Anritsu network analyser I bought for $400 OTOH can also tell me the equivalent circuit parameters to actually model the crystal. The $100-ish NanoVNA can also be used to measure crystals.