Rigol DHO804 Test and Compare Thread

[Performa01]

Finding: you can actually see the non-linearity at a vertical shift of 500mv (ten divisions of the 50mV/div range). That's practically a no-go.

Who needs a linear ADC, when we can have a prestigious 12 bit badge on the front cover and as long as it's cheap!

[Martin72]

Strange enough, the amplification factor appears to increase, but then there's also some kind of "integration" of the signal taking place.

That's what I meant, a kind of "filtering", plus that with the same settings the signal curve tends to be flatter with the rigol, at least that should remain the same if you don't change anything except the vertical sensitivity.

[2N3055]

Strange enough, the amplification factor appears to increase, but then there's also some kind of "integration" of the signal taking place.

That's what I meant, a kind of "filtering", plus that with the same settings the signal curve tends to be flatter with the rigol, at least that should remain the same if you don't change anything except the vertical sensitivity.

Martin,

could you please repeat measurement: 20µs/div, 10mV/div no offset (offset 0), AC coupling, NO 20MHz BW limit. Triger 0V .
Signal should got straight through center of screen with no adjustments.
Post that image. We can directly compare that with my posted images.

To make it clear, this is not prove what Thomas found. This is to see firstly how step looks like, especially top of it. To prove if there is some filtering or diff nonlinearities and nonmonotonic behavior that would show as noise band or flat line instead of clear wave on top...

Thanks.
 Siniša

[2N3055]

Heck, guys -- I don't know why I am having such a hard time to get my point across. I am not concerned about varying step sizes within one of the captures or such. I am pointing out that the images which have been compared were showing two different signals. Specifically, two different segments of the test board-generated sine waveform, with different slopes and hence step sizes (differing between the scopes). WHich made the steps look less pronounced on the Rigol.

Please have a look at the attached, which is an annotated combination of the screenshots "Rigol_50mv" and "Siglent_50mV" from Martin's post #237. Maybe that explains it better than another 1000 words?

Yes we understand. That is why we moved to AC coupling and offset and trigger level of 0.
But Thomas found that signal gets distorted at some points when offset is applied.....

[Fungus]

Yes we understand. That is why we moved to AC coupling and offset and trigger level of 0.
But Thomas found that signal gets distorted at some points when offset is applied.....

It might be to do with the AC coupling. What size capacitor is in each device?

Who needs a linear ADC, when we can have a prestigious 12 bit badge on the front cover and as long as it's cheap!

It's not the ADC.

If it was the ADC it would be on every screenshot ever. No line would ever be straight.

[Performa01]

Yes we understand. That is why we moved to AC coupling and offset and trigger level of 0.
But Thomas found that signal gets distorted at some points when offset is applied.....

There's a Déjà-vu: the ones brave enough to read my review of the SDS1104X-E back in late 2017, might have noticed the following statement in the "DC Accuracy" section:

The table below shows the results of all measurements. For each vertical gain setting, measurements have been performed for both polarities and three offsets (zero and ±3 divisions) with and without input signal respectively, resulting in a total of seven measurements per range. The reason why measurements with offset have been included is the probably widely unknown fact that not all DSOs will pass this test, so I wanted to be thorough.

Well, to be honest, I've not come across a single DSO so far, that would not pass this test - with one exception: it was ... a Rigol DS4000 (which wasn't exactly cheap back in 2015). And support never solved this problem. So Rigol and offset ... appears to be no good relationship.

I've never tested my old Rigol DS1052E though, just because i didn't think it was worth it, back when this scope still was a thing.

[ptluis]

Now something completely different, before I leave for the next hours...Just for fun.

https://youtu.be/UOFe9wmdIKk

The RIGOL seems to have a really small display area to show waveforms, signals. It is possible to hide the top and lower bars?

[Gortosch]

Hello all,

I actually ordered a DHO924S from Batronix. But now I'm thinking about whether an MSO5000 would be better for me. Since I'm currently doing more digital technology than analogue technology, the 12 bits don't appeal to me at the moment. As I said, at the moment.
The form factor is great, especially because I don't have much space, but it's not really the deciding factor. Otherwise, I don't see any more advantages with the DHO924s. Rather only disadvantages. Am I missing something important? What do you think?

Best regards
Sven

[ebastler]

I actually ordered a DHO924S from Batronix. But now I'm thinking about whether an MSO5000 would be better for me. Since I'm currently doing more digital technology than analogue technology, the 12 bits don't appeal to me at the moment. As I said, at the moment.
The form factor is great, especially because I don't have much space, but it's not really the deciding factor. Otherwise, I don't see any more advantages with the DHO924s. Rather only disadvantages. Am I missing something important? What do you think?

I would agree that, if you don't explicitly need the small form factor, potential for battery operation, and the somewhat better ADC resolution, the MSO5000 is the better choice.

The MSO5000 is relatively noisy for an 8-bit front end, but certainly good for digital electronics. It is also very hackable, all nicely documented, so you can start with an MSO5072. (which even comes with 4 probes, it seems? Nope, four inputs and front ends, but only two probes. The two included ones are 350 MHz however, according to Batronix.)

The DHO9xx is cute, but the small form factor also comes with a price to pay. The little fan is more noticeable than larger ones, and the small touch screen limits the space available for curve display. Also, it does not have individual controls for the four channels which the MSO5000 provides, of course.

[ebastler]

The RIGOL seems to have a really small display area to show waveforms, signals. It is possible to hide the top and lower bars?

No. While various overlays, results displays etc. are of course dynamic, the settings bars at the top and bottom are always there.

[TurboTom]

I just found out another detail that may save Rigol's honor (...at least a little bit) 

The observed effect of the nonlinearity at high vertical offsets appears to be only present if the total amplitude of the input signal far exceeds the screen range. Actually, we've been looking at a 3Vpp signal in the 50mv/div range, offset another 500~700mv in one vertical direction. So the maximum level of that signal outside of the "visible range" was 44 divisions -- about five vertical screen ranges.

An AC signal, riding on top of a DC signal is no problem for the scope if it is offset by 20 divisions (1V Offset in the 50mV/div range). The attached screenshot documents this, a 350mV ramp signal with -1V offset fed to channels 1 and 2, comparing the traces of channel 1 with DC coupling and vertial offset vs. channel 2, AC coupled and virtually no offset (only a little bit for trace separation).

So as long as the AC component of the input signal isn't exceeding the measurement range by far, things are okay.

[Fungus]

The table below shows the results of all measurements. For each vertical gain setting, measurements have been performed for both polarities and three offsets (zero and ±3 divisions)

Well, to be honest, I've not come across a single DSO so far, that would not pass this test - with one exception:

Finding: you can actually see the non-linearity at a vertical shift of 500mv (ten divisions of the 50mV/div range).

3 divisions != 10 divisions.

[Martin72]

Martin,
could you please repeat measurement: 20µs/div, 10mV/div no offset (offset 0), AC coupling, NO 20MHz BW limit. Triger 0V .
Signal should got straight through center of screen with no adjustments.
Post that image. We can directly compare that with my posted images.

Hi,

Of course I can...
Settings like you suggested: 20µs/div., 10mV/div, AC coupling (channel and trigger) and here we go...
Except the zero-crossing, all are looking the same, rigol dho, siglent sds1104x-e, siglent sds2504x hd..
Looks like the rigol have a DC "problem".

[Gortosch]

Hello ebastler,

thank you for your quick reply. What exactly do you mean by "relatively noisy"? Are there any tests, maybe even a video?
But basically I have already decided on the MSO5000.

[Fungus]

I actually ordered a DHO924S from Batronix. But now I'm thinking about whether an MSO5000 would be better for me. Since I'm currently doing more digital technology than analogue technology, the 12 bits don't appeal to me at the moment. As I said, at the moment.

If you're doing mostly digital stuff then you should probably get the MSO5000. It has far higher sample rate, bandwidth and memory depth.

You can easily hack it to unlock all options.

You won't get the touch screen or the cuteness, but... 

[Fungus]

I just found out another detail that may save Rigol's honor (...at least a little bit) 

The observed effect of the nonlinearity at high vertical offsets appears to be only present if the total amplitude of the input signal far exceeds the screen range.

So there could be some op-amp overload in there...

[ebastler]

What exactly do you mean by "relatively noisy"? Are there any tests, maybe even a video?

There have been long (and sometimes heated) debates on this forum. Here's an entire thread about scope noise which talks about the MSO5000 a lot. The link goes to a page mid-thread which has some comparison screenshots, but other pages might be interesting as well: https://www.eevblog.com/forum/testgear/how-much-noise-floor-and-other-things-matter-in-oscilloscope-usability/150/

Nevertheless, if your main interest is digital electronics, I'd say this should not stop you from getting an MSO5000!

By the way: The 16 digital inputs are the same on the MSO5000 and the DSO924, it seems. Rigol's probe kit is expensive (but nice -- fast comparators which work up to +- 40V, good if you should also be interested in very vintage digital electronics). But there are cheap DIY solutions if you only need typical logic levels up to 5V.

[2N3055]

Martin,
could you please repeat measurement: 20µs/div, 10mV/div no offset (offset 0), AC coupling, NO 20MHz BW limit. Triger 0V .
Signal should got straight through center of screen with no adjustments.
Post that image. We can directly compare that with my posted images.

Hi,

Of course I can...
Settings like you suggested: 20µs/div., 10mV/div, AC coupling (channel and trigger) and here we go...
Except the zero-crossing, all are looking the same, rigol dho, siglent sds1104x-e, siglent sds2504x hd..
Looks like the rigol have a DC "problem".

Thanks Martin.

That actually looks OK.

Now back to what Tom experienced.
I agree that is some type of input overload/recovery.

What I tested is following (and I suggest you test the same on Rigol):
- same demoboard sine 50 Hz non filtered.
- Input in DC mode.
- position signal that you look at top of sine. (5 ms/div).
- try at 200, 100, 50 mV/div ( I don't know the offset range of top of my head.).
Look for distortion...
Hint: once you roughly get where it starts distorting, set vertical to fine for more precise point..

[TurboTom]

Hi,

Of course I can...
Settings like you suggested: 20µs/div., 10mV/div, AC coupling (channel and trigger) and here we go...
Except the zero-crossing, all are looking the same, rigol dho, siglent sds1104x-e, siglent sds2504x hd..
Looks like the rigol have a DC "problem".

I wouldn't say that they have a DC problem, it's rather an input overdrive problem that more or less is present on all oscilloscopes. It just seems that Siglent did a better job in this area than Rigol. As long as the input amplitude stays within safe limits, there's no problem. See the attached screenshots: Once a square signal of 2Vpp total amplitude with 20% of of ramp riding on top, the other one with 1.4Vpp and otherwise the same configuration, coupled to both channels 1 and 2, yet at different sensitivities and channel 1 (50mV/div) offset by 1V /500mV to visualize the "low part" of the signal. I guess this explains the situation quite clearly. It seems, other vertical ranges are more tolerant to high AC signals. Unfortunately, the 50mV range is very commonly used with a 10:1 probe...

...
So there could be some op-amp overload in there...

Exactly! 

[ebastler]

The MSO5000 [...] is also very hackable, all nicely documented, so you can start with an MSO5072. (which even comes with 4 probes, it seems? Nope, four inputs and front ends, but only two probes. The two included ones are 350 MHz however, according to Batronix.)

Correcting my prior advice:

Having looked at the Batronix website, the MSO5074 is probably the better buy at the moment. The slightly higher price not only buys you two extra probes (which cost more than the price difference if bought separately). It also gets you a nice option bundle, including various decoders and the AWG license, so you can leave the scope unhacked at least early on during the warranty period.

And now back to our scheduled program...

[mwb1100]

I actually ordered a DHO924S from Batronix. But now I'm thinking about whether an MSO5000 would be better for me.

The MSO5000 is relatively noisy for an 8-bit front end, but certainly good for digital electronics. It is also very hackable, all nicely documented, so you can start with an MSO5072. (which even comes with 4 probes, it seems? Nope, four inputs and front ends, but only two probes. The two included ones are 350 MHz however, according to Batronix.)

The MSO5074 is being sold in the US  on the rigolna.com site for $100 less than the MSO5072 right now ($800 vs $900).  I don't think that deal has made it to the EU or if it will.  But it would be worth asking Batronix (or whoever) if the deal is coming or if they can match.  The worst that can happen is they say no.

You won't get the touch screen or the cuteness, but... 

The MSO5000 isn't as cute as the DHO900, but it does have a (larger) touchscreen.

[2N3055]

What exactly do you mean by "relatively noisy"? Are there any tests, maybe even a video?

There have been long (and sometimes heated) debates on this forum. Here's an entire thread about scope noise which talks about the MSO5000 a lot. The link goes to a page mid-thread which has some comparison screenshots, but other pages might be interesting as well: https://www.eevblog.com/forum/testgear/how-much-noise-floor-and-other-things-matter-in-oscilloscope-usability/150/

Nevertheless, if your main interest is digital electronics, I'd say this should not stop you from getting an MSO5000!

By the way: The 16 digital inputs are the same on the MSO5000 and the DSO924, it seems. Rigol's probe kit is expensive (but nice -- fast comparators which work up to +- 40V, good if you should also be interested in very vintage digital electronics). But there are cheap DIY solutions if you only need typical logic levels up to 5V.

At this point MSO5000 is actual functional product in comparison to DHOxxx series that is highly unfinshed in many regards.
If you need something that actually works and has to be Rigol, MSO5000 is actual working product. And yes it is noisy, but as other said, for digital work is much more powerful than any DHO scopes. And it has bigger screen.
Other (more expensive) option is SDS2000X +..

But I would not recommend anybody to buy DHO series today if it is going to be their only scope and they would depend on it to work flawlessly. It is in the early adopter phase...

[Gortosch]

That's exactly what I was thinking. I have just ordered the MSO5074 and cancelled the DSO924S. Thank you.

[Performa01]

The table below shows the results of all measurements. For each vertical gain setting, measurements have been performed for both polarities and three offsets (zero and ±3 divisions)

Well, to be honest, I've not come across a single DSO so far, that would not pass this test - with one exception:

Finding: you can actually see the non-linearity at a vertical shift of 500mv (ten divisions of the 50mV/div range).

3 divisions != 10 divisions.

Oh! So you really believe the DS4000 would have performed better if we'd measured it at 10 divisions offset instead of just three?

I just found out another detail that may save Rigol's honor (...at least a little bit) 

The observed effect of the nonlinearity at high vertical offsets appears to be only present if the total amplitude of the input signal far exceeds the screen range.

So there could be some op-amp overload in there...

So you still believe in "OpAmp overload", even though David Hess and me have explained numerous times what causes overload distortion in a serious scope frontend?

Yes we understand. That is why we moved to AC coupling and offset and trigger level of 0.
But Thomas found that signal gets distorted at some points when offset is applied.....

It might be to do with the AC coupling. What size capacitor is in each device?

Do you even know where that capacitor sits? And how do you suppose anyone knows its value? Do you have a DMM? have you ever measured the input resistance (DC!) of any serious scope when it is AC coupled?

[Martin72]

What I tested is following (and I suggest you test the same on Rigol):
- same demoboard sine 50 Hz non filtered.
- Input in DC mode.
- position signal that you look at top of sine. (5 ms/div).
- try at 200, 100, 50 mV/div ( I don't know the offset range of top of my head.).
Look for distortion...
Hint: once you roughly get where it starts distorting, set vertical to fine for more precise point..

I had done this with the "lower peak" of the sinus, because from 100(or was it 50)mV the rigol allows an offset of +/-1V (above this it is 8V).
When I found nothing, I shortened the time base to 500µs, see picture.