How much noise floor and other things matter in oscilloscope usability

[2N3055]

But It has also a lot more noise.... I have an idea to show you how much the signal degrade due to this problem.

All 'scopes degrade the signal, none are perfect.

The real problem with this Rigol vs. Siglent comparison is that you're comparing two devices side by side where one of them happens to work better at 1mV then the other one does.

ie. When you turn the Siglent's vertical control to "max" then it makes the Rigol look bad at that level. If you were comparing 1V signals or 5V signals in your screenshots then you wouldn't see the same difference between them.

Guess what? Electronics doesn't magically stop at the exact place where the Siglent's vertical control does. There's plenty of signals below 1mV. You'll need an amplifier to see them and the exact same same amplifier would work with a Rigol, too.

Plus: You originally said you do digital stuff, so... 

Your thinking:

Siglent that has 5-10X times better noise and performance and sensitivity but cannot measure properly something at 200uV levels,
Rigol MSO5000 cannot measure something properly even at 10mV
That makes them equal. Because they both have something they cannot measure.

By the analogy:
You are same as a gorilla. You both have eyes.. Let's ignore other, irrelevant, details..

Get a grip..

[2N3055]

Also you keep repeating about some magical amplifiers. Amplifiers that have DC-100 Mhz bandwith and less noise as even a little Micsig or Siglent SDS1104X-E cost as much as a good scope from Keysight.

Yes, but amplifiers from DC to 1MHz are incredibly cheap (ie. a $2 OP-amp) and would be perfectly adequate for power supplies and audio work.

Amplifiers from 1MHz to 2GHz are also incredibly cheap.

Switching power supplies are measured by convention in DC-20MHz range. And that was for old switchers switching at up to some 100ths of kilohertz. Today switchers are in MHz range, and you better be looking at them up to 300-400 MHz range because of EMI problems...

Also those RF amplifiers are not replacement for a proper scope front end.

[Fiorenzo]

I 'm reading some of these posts and 

It is obvious MSO5000 has much more noise, and that it is a problem, unless you only look at digital signals and just want to look at general shape and some timing information.
It is shame, really, because new Rigol scopes are much more powerful processing wise than the old ones, and generally held great promise but analog front end/ADC noise performance is not very good.

Scope with low noise is always going to be better instrument than the one with high noise. Why is that even a discussion?
Is this some audiofool discussion how this huge noise is pleasant to look at because it's pretty? 

I don't use bandwidth limiting, averaging or any "signal cleanup" features when I'm looking into a signal I want to understand. You would want to look at this switcher signal with a full 1GHz bandwidth and with as low noise scope you can.
To really see what is there... Switching ripple is most of the time least interesting part of switching PSU. We expect it to be there, and most of the time it will be roughly what we calculated. Other, higher frequency stuff (those little hairs on top) is much more problematic and most of the time those will give you headaches.. Nanovolts of those will already be seen on any EMI test...

You filter, limit and "cleanup" signal in circumstances where you understand your signal and you want to ignore noise and other parts of signal on purpose. If your signal is buried inside the noise, you average.
But is that noise part of signal you're measuring or your scope is not irrelevant. If it comes from DUT I want to know that. I want to see it..
Only way to do that is to have low noise scope.

Of course, like OP correctly asked, there is a point of diminishing returns..
Is scope with 5 uV of RMS noise so much better than one with 50uV RMS noise for measuring this switcher signal from this example? Probably not.
It would be definitely better but probably not usefully so in this case. But one with 50uV of RMS noise is definitely better than one which has trace that is whooping 20 mV wide... On a signal that is 60mV P-P...
On this test I would call MSO5000 from Rigol useless for this measurement. And averaging this not autocorrelated signal ( it doesn't repeat cleanly and doesn't retrace it's waveform exactly but varies slightly all the time) will not extract more detail but will hide even more information about signal..

OTOH Siglent shows pretty much perfect representation of the signal, big peaks, ripple AND little hairs. That is your switcher output. That is useful information..

Little Micsig STO1104C/E, or Siglent SDS1104X-E could do equally good job here.

Sad part is that little Rigol DS1054Z would be much better for this signal than MSO5000.. DS2000A had excellent low noise front end .. But new series of Rigol scopes is very powerful in processing power but analog performance is worse than older series. Shame really, otherwise they are very nice scopes.

Thank you for the great explenations, I also agree with you

[Vestom]

The lower sampling frequency of the SDS2000X+ is much less of a problem than it sounds due to the excellent ETS-like trigger implementation. There are plenty of examples showing the Leo Bodnar pulse with very good fidelity on the SDS2k+.

However frame-by-frame averaging is really a kludge one should be careful about using, since it hides transients, glitches etc. Also with modern low-amplitude digital signals measured using x10 probe, low noise performance has become more important for digital signals than you might think...

Of course you can add external amplifiers - but why not buy a proper scope instead?

[Fiorenzo]

With channels 1 and 3 of each scope activated a very different picture may emerge at some timebase settings.

Yes, because if you enable to adjacent channels the Siglent only has 1GS/sec to look at 350Mhz signals, ie. it's getting uncomfortably close to Nyquist.

Turning a channel off can bump the sample rate to 2GS/sec and give a different picture.

No. Actually i did many experiment about It. If you set the Rigol so it has a lower sampling speed the noise doesn't become lower.

I checked It in many ways.

For example if you turn on all the channels together the noise stay the same.

If you lower the memory buffer the sampling rate becomes lower but the noise stay again similar until you force It to work with a very very little Memory like 20k or so.

I have tried all i could think to get a better sampling from the Rigol and It was not possible.

In regard of me now I am doing digital stuff but i am going to work a lot with analoge signals and circuits so i preferred to switch to the Siglent.

As i said i think the Rigol is a very great scope but noisy

[Fungus]

I 'm reading some of these posts and 

It is obvious MSO5000 has much more noise, and that it is a problem, unless you only look at digital signals and just want to look at general shape and some timing information.

Fiorenzo originally said his main work was digital.

Scope with low noise is always going to be better instrument than the one with high noise. Why is that even a discussion?

Because: "Price"

(ie. This isn't a 100% technical discussion, if it was we'd all be driving 10-bit R&S 'scopes)

On this test I would call MSO5000 from Rigol useless for this measurement.

The Rigol displayed "14.003mV" on screen and the Siglent displayed "14.0482mV"

That's 0.3% difference between them.

Little Micsig STO1104C/E, or Siglent SDS1104X-E could do equally good job here.

Yep. My original recommendation was to save 1000 Euros and get the SDS1104X-E.

[Fungus]

No. Actually i did many experiment about It. If you set the Rigol so it has a lower sampling speed the noise doesn't become lower.

That's because it won't actually change the ADC clock speed, it just discards samples.

What's being discussed is the sample rate to bandwidth ratio. When it approaches a ratio of 2 you'll start to see artifacts in the display.

The Siglent has 350Mhz bandwidth and can drop to 1Ghz sample rate if you enable all channels. This can produce visibly different results at maximum zoom. You'll see it most on digital signals.

The Rigol doesn't drop below 2GHz sample rate so it should never have a problem.

[nctnico]

No. Actually i did many experiment about It. If you set the Rigol so it has a lower sampling speed the noise doesn't become lower.

That's because it won't actually change the ADC clock speed, it just discards samples.

What's being discussed is the sample rate to bandwidth ratio. When it approaches a ratio of 2 you'll start to see artifacts in the display.

The Siglent has 350Mhz bandwidth and can drop to 1Ghz sample rate if you enable all channels. This can produce visibly different results at maximum zoom. You'll see it most on digital signals.

The Rigol doesn't drop below 2GHz sample rate so it should never have a problem.

In the end neither is suitable for looking at 350MHz signals using 4 channels. The Siglent SDS2k due to low samplerate, the Rigol MSO5000 due to excessive noise. Also note what David Hess wrote: Rigol typically performs math on decimated data which can give the wrong results when doing measurements on noise.

All in all, if you venture into the HF arena, you'll need to look at more expensive scopes. For a general purpose daily driver scope, low noise is king all day long.

[2N3055]

No. Actually i did many experiment about It. If you set the Rigol so it has a lower sampling speed the noise doesn't become lower.

That's because it won't actually change the ADC clock speed, it just discards samples.

What's being discussed is the sample rate to bandwidth ratio. When it approaches a ratio of 2 you'll start to see artifacts in the display.

The Siglent has 350Mhz bandwidth and can drop to 1Ghz sample rate if you enable all channels. This can produce visibly different results at maximum zoom. You'll see it most on digital signals.

The Rigol doesn't drop below 2GHz sample rate so it should never have a problem.

In the end neither is suitable for looking at 350MHz signals using 4 channels. The Siglent SDS2k due to low samplerate, the Rigol MSO5000 due to excessive noise. Also note what David Hess wrote: Rigol typically performs math on decimated data which can give the wrong results when doing measurements on noise.

All in all, if you venture into the HF arena, you'll need to look at more expensive scopes. For a general purpose daily driver scope, low noise is king all day long.

MSO5000 actually can use all data.it has propper implementation  of math but it is let down by noise.. Shame.

[Fiorenzo]

Yes fungus I think so.
In regard of your suggestion to buy a cheaper scope my thinking was in line with you but i needed some function that are only in the Siglent sds2000.

About this topic my need was to understand how more important is sample rate against front end noise.

Because of my bad english maybe i could not explain well my doubts.

In the end if you or other have suggestions i can do some other test with the limited equipment i have until I have both the scopes at home..... because i am going to send back the Rigol as soon as possible.

[Fungus]

In the end neither is suitable for looking at 350MHz signals using 4 channels. The Siglent SDS2k due to low samplerate, the Rigol MSO5000 due to excessive noise. Also note what David Hess wrote: Rigol typically performs math on decimated data which can give the wrong results when doing measurements on noise.

All in all, if you venture into the HF arena, you'll need to look at more expensive scopes.

Probing a 350MHz signal with passive probes is also a minefield, the artifacts from the probe will usually be bigger than the signal.

I wouldn't buy either of these for the bandwidth, I'd buy them for the big touch screens, large memory, etc.

[Fungus]

Yes fungus I think so.
In regard of your suggestion to buy a cheaper scope my thinking was in line with you but i needed some function that are only in the Siglent sds2000.

What function?

Siglent has added more functions to the SDS1104X since launch. Maybe they added it but you're looking at an old manual.

In the end if you or other have suggestions i can do some other test with the limited equipment i have until I have both the scopes at home..... because i am going to send back the Rigol as soon as possible.

The only other interesting test is the effect of intensity on the color-graded display.

For that a photo of the screen isn't really good enough though. You'll need to put in a USB stick and press the 'print' button.

[Martin72]

I´ve changed from MSO5354 ( ) to the sds2k+ - After I did the very first ripple-measurement with the mso.
The noiselevel is "horrible", useless for doing low noise measurements - By the way, Dave mentioned it also in his first rigol mso 5000 video.
I got the sds2k+ over a year now, before the rigol over a year, there is nearly nothing the rigol can do better.
Display is worser, much worser ( same resolution btw), UI is mickey mouse style and sometimes really confusing, the response itself is slower...
It´s a shame, because the hardware itself is powerful, except the noisy frontend.
And as I´ve asked the rigol support, if there is a chance to get it better by buying the 7000, their answer to me kills everything.
Frontend are the same and that was it, I´ve changed to siglent without any regrets so far, except the missing of 4 math-traces at the same time..

[mawyatt]

We had to sift thru all this a year and half ago. After a few months decided to "listen" to those that had actual experience with the scopes and for comparisons actually had used various brands. Dave's videos are a great resource, and should be reviewed many times.

Couple folks here (think 2N3055 & Martin72) have actually used and worked with both the Rigol and Siglent scopes, so value these experiences and responses.

BTW we decided on the Siglent SDS2000X+, the deciding factor was the Rigol noise. Now have 2 Siglent scopes and maybe getting a 3rd, so that alone says enough about our experiences.

Best, and Happy Scope Hunting

[mawyatt]

Not to discredit Rigols' attempt and deployment of integrating the entire front end and ADC, actually we should applaud them

Coming from a career that often involved low noise, wide bandwidth and high dynamic range requirements, designing such at the discrete level is extremely difficult, and speaking from experience, at the Integrated IC level, this becomes exceedingly difficult and quite expensive. Things that work at the discrete level may not work so well on an IC, and an entire different design approach is often required. Modern IC processes that feature amazing small devices, and very fast, are dictated by the digital requirements. The analog use of such is "you get what you get" and you must figure out how to achieve the analog performance goals, which often means a complete departure from a conventional discrete approach.

Anyway, hopefully Rigol (Siglent, maybe some other mid-level players) will continue with an integrated scope front-end and ADC solution because we'll all benefit from this integration.

Best,

[Performa01]

Your thinking:

Siglent that has 5-10X times better noise and performance and sensitivity but cannot measure properly something at 200uV levels,

Are you sure? See reply #130 in the following thread

https://www.eevblog.com/forum/testgear/suggestions-for-a-dmm/msg2766948/#msg2766948

[Fungus]

Rigol MSO5000 cannot measure something properly even at 10mV

This, in a thread full of screenshots showing the Rigol measuring things to sub-millivolt resolution with the same accuracy as a Siglent.

[Performa01]

Rigol MSO5000 cannot measure something properly even at 10mV

This, in a thread full of screenshots showing the Rigol measuring things to sub-millivolt resolution with the same accuracy as a Siglent.

It appears hard to understand for some (well, actually I know only one), but even though the amplitude measurement for the ripple is correct - which doesn't come as a surprise since noise averages out to zero - engineers still want to see the details apart from the very predictable ripple, which includes switching noise and other potential RFI components, as well as any other glitches that might hint on hidden problems with the circuit.

EDIT: And the result is only equal for the Vrms measurement. Has anyone noticed the huge difference in the Vpp measurment, where the noise does not average out to zero?

[2N3055]

Now let's see if you can understand this:

Imagine I have a sub mV signal for you to look at ... what are you Siglent owners going to do?

Now let's see if you can understand this:
Rigol cannot see anything less than 10mV.

Siglent users will see this, with stable triggering:


That's almost two orders of magnitude better.

I rest my case. You're trolling now, nothing else... Stop wasting our time. Please.

[nctnico]

R&S owners are equally blessed with low noise 



And the option to use a filter on the signal.

[gf]

R&S owners are equally blessed with low noise 

Compared to the previous Siglent screenshots, the sine wave looks a bit distorted, though. Or is the signal generator to blame here?

[nctnico]

R&S owners are equally blessed with low noise 

Compared to the previous Siglent screenshots, the sine wave looks a bit distorted, though. Or is the signal generator to blame here?

There is some noise riding on top and keep in mind the bottom trace is not averaged but filtered.

[David Hess]

What are you going to do when the ripple is even smaller?  For example, here is a ~150uV_rms 10MHz signal being clearly triggered and displayed.  This signal was displayable--triggerable and above the noise threshold on a Tek 2465B (shown), a Tek 2221A (digital and analog) and the Siglent 1104X-E (although just barely and not as reliably).  I was able to do the same thing with a 1MHz and 100Hz signal of approximately the same amplitude.  What would it look like on the Rigol? 

You can't average your way out of this when you are looking for noise--possibly non-periodic--in the first place.  On the Tek 2221A and the Sig 1104X-E, averaging made the signal look nicer but I'm not convinced that means better.

I have run across that problem a couple times, where only a lower noise oscilloscope would do when looking for subtle anomalies.  The usual solution is to use averaging or high resolution mode and trigger off of a different channel from a signal which is synchronous to the signal of interest, but this is not always possible.  See below about differential probes.

My 10 bit digitizing but not a DSO Tektronix 400 MHz 7854 can reveal signals that my 40 uV RMS over 100 MHz 2232 or 2230 cannot even see in digital storage mode, and do it with an RF sampling front end if needed, but an equivalent instrument today would be 10s of thousands of dollars and we are not discussing those.

Which initiates another question; what would the MSO5000 sampling drop to if one more channel was activated ?

Lower cost DSOs do that because they either have only one bank of acquisition memory with limited bandwidth shared between 4 digitizers, or because the digitizer is interleaved between the 4 channels.  The later is practically universal in lower cost DSOs.

Either can be seen as a cost saving measure or a way to maximize performance when only a single channel is used, and more expensive instruments may use completely separate digitizing and storage for each channel so sample rate, and usually record length, does not depend on the number of channels.  Tektronix used to refer to a DSO as being "real time" if its maximum sample rate, and record length, did not change with the number of channels and for many years they maintained a separate line of DSOs which worked this way because some applications demand it.

Also you keep repeating about some magical amplifiers. Amplifiers that have DC-100 Mhz bandwith and less noise as even a little Micsig or Siglent SDS1104X-E cost as much as a good scope from Keysight.

Yes, but amplifiers from DC to 1MHz are incredibly cheap (ie. a $2 OP-amp plus power supply) and would be perfectly adequate for audio work and looking at power supply ripple.

Amplifiers from 10kHz to 2GHz are also incredibly cheap.

What is not cheap or easy are getting good flatness and settling time over the bandwidth of interest.  However the high noise is a problem of design and not cost.  Up to 350 MHz, there is no excuse for such a noisy implementation and obviously Siglent is doing something right that Rigol is not.  The noise on the Siglent is high compared to what it could be, but that makes the noise on the Rigol much worse than high.

In the end neither is suitable for looking at 350MHz signals using 4 channels. The Siglent SDS2k due to low samplerate, the Rigol MSO5000 due to excessive noise. Also note what David Hess wrote: Rigol typically performs math on decimated data which can give the wrong results when doing measurements on noise.

It is not the decimated data which is the problem but performing math on the display record which has already been processed for the display.  More expensive instruments maintain a separate full resolution record for processing what is essentially the "raw" data, that is separate from the processed display record.

Probing a 350MHz signal with passive probes is also a minefield, the artifacts from the probe will usually be bigger than the signal.

Passive high impedance probes are more difficult to use at higher frequencies than active probes, but if this is taken into account, the usual problem is noise from the ground loop with a singled ended probe, which even an active probe does not solve.  A differential probe solves this but at the expense of greater noise, and this tradeoff is almost always worth it if their higher cost can be accepted.  Several times I have probed signals approaching the limit of the input noise of my oscilloscope where differential probing was a solution because it removed common mode noise.

I wouldn't buy either of these for the bandwidth, I'd buy them for the big touch screens, large memory, etc.

I was thinking that about the Rigol MSO5000 series before I saw how much better the Siglent is for noise.  They are both noisier than the general state of the art of more than 2 decades ago, but for most applications that is good enough.  By preference I do not even use the lowest noise oscilloscope that I have available.  Fiorenzo did the right thing by inquiring.

[Fungus]

Fiorenzo did the right thing by inquiring.

And has provided some valuable data.

[masterx81]

How did you get so low noise waveforms? Did all you live in a faraday cage? 
I can't go so low with my 1104x-e as it pickup always random noises also with all things turned off in the room except the scope. It get also noises from the cooking gas sparks for starting the flame two rooms away from me (kitchen). I get the 34khz from the fluorescent lamp on the desk when powered on. The scope itself send a decent ammount of noise from the lcd panel (i think the lcd backlight buck is causing it), and you need to be very careful to the routing of the wire of the probe. Also if i connect the output of a battery powered signal generator, i can't reach those so low level of noise (still better than the fg fy6900, correctly grounded and with all-linear psu). In any case before the siglent i had an owon 7102... and hell, that thing was horrible in noise.