Rigol DHO804 Test and Compare Thread

[TurboTom]

Did some tests on a newly arrived DHO914S. Here are a few of my findings:

The traces look nice and crisp at the first glance, but it gets quickly glear that the intensity grading is way worse than on Rigol's legacy scopes. Even the DS1000Z is performing far better than this new 12-bit series. See RigolDS18 -- this is as good as it gets! It seems like they have got only four gradations and need to dither in between...

Rise time is pretty decent for a 125MHz scope @ 1.87ns (rise time of the source < 500ps, everything's properly terminated), see RigolDS11

Colour grading doesn't seem to work properly with infinite persistance, but that may have to be expected. I tried to use it while evaluating time base jitter which apparently is quite decent with one or two channels active but gets much worse when three or all channels are enabled (RigolDS17:16:15). I used a highly stable 10MHz signal with the timebase delayed by 100ms and looked at the neares positive slope. The screenshots show approx. one minute of recording. The few traces that deviate from the "bulk" are the result of me pressing the "quick" (screenshot) button and thus applying a little pressure on the scope. This seems to slightly shift the scope's internal reference oscillator (probably a digital TCXO since during warmup, the "bulk" of the delayed traces doesn't seem to slowly walk but rather "jumps" in increments of ~1.2ns. But nothing really to complain about here.

The internal AWG won't "pull the butter off the bread" (direct translation of a local idiom... ) -- especially the reconstruction filter doesn't seem to do a good job. I'm very surprised that despite the differential design of the DAC output, we see substantially asymmetric ringing of the analog output signal (RigolDS19). The cause for this may be too high tolerances in the components of the reconstruction filter or not quite the best choice of the amplifier(s) that follow the filter. Rise- and fall times of the AWG (square wave) are in the ballpark of 10ns.

I've got the impression that there's still a considerable way to go for Rigol with this platform (both hard- and software-wise) and very likely I'll return the instrument since I'm far from satisfied, even with my initial findings. The step back in some areas vs. Rigol's legacy products really comes unexpected.

[Martin72]

@ ebastler:

I find it a little bit too dim, like the MSO5000 once.

Boot-Time : You´re right! Second Boot is 50sec, don´t know why it took so long by the first time...

[Martin72]

I find it a little bit too dim, like the MSO5000 once.

Interesting...
Ambient lights are the same, but now I didn´t find it too dim anymore..

[ebastler]

it gets quickly glear that the intensity grading is way worse than on Rigol's legacy scopes. Even the DS1000Z is performing far better than this new 12-bit series. See RigolDS18 -- this is as good as it gets! It seems like they have got only four gradations and need to dither in between...

That screenshot looks nasty indeed. I find it hard to believe that this is the best the scope can do, given that the datasheet specifically advertises "real-time 256-level intensity grading".

My DS1054Z display looks a bit like that if I switch it to "Dot" mode. But I understand that option is not even avaialble with the DHO 800, so it can't be behind that ugly dithering. Have you checked the other display and aquisition settings to see whether anything might be sub-par? I don't want to give up hope just yet that Rigol did a decent job with the intensity grading...

Colour grading doesn't seem to work properly with infinite persistance, but that may have to be expected.

Agree. I don't see how grading (whether intensity or color based) would go together with infinite persistence, which saturates all pixels which see a signal at all. Hence no concern over that one.

[Martin72]

What are the settings to get this signal ? AM modulation carrier/signal frequencies?

[Martin72]

Quick and dirty, intensity grading....looks very good.
But switching to color grade mode will deactivate the intensity adjustment.
(DS5: 25Mpts, DS7:1Mpts)

[Martin72]

So, but now is really enough for today...

-Bodnar pulser connected via 50Ohm resistor, risetime is 3.6ns which makes something around 100Mhz, not bad.
Or if you put in the formula 0.45, you get exactly 125Mhz...Aha.

When I connected my SDG2122X earlier, I thought to myself, oh run through the frequencies up to 120Mhz.
1Vrms.
At 120Mhz it was just under 700mVrms, from about 100Mhz the level became smaller.
Before that, starting at about 70Mhz the level swung up to 1.5Vrms.
A resonance point ?
What I also find interesting:
In the options menu there are three options, twice decoding, once bandwidth(100Mhz).
The first two are marked as "forever" (makes sense), the bandwidth as "limit".
So it may well be that the 70Mhz models first have 100Mhz, so I understand that anyway.
It is also interesting under the point that you can not officially buy a bandwidth upgrade for the 70Mhz model.

[iMo]

The 1.5V peak at 70MHz comes from the higher order low-pass.
It could be the 100MHz "Limit" means "limited time", like a demo version for the first 10days.. But they forgot to install the RTC, thus it means forever..

[ebastler]

It could be the 100MHz "Limit" means "limited time", like a demo version for the first 10days.. But they forgot to install the RTC, thus it means forever..

They could always enable it for a limited number of power cycles or power-on time. But it would be kind of mean to tease users with the 100 MHz performance when one can't even buy an upgrade for permanent 100 MHz bandwidth later...

Maybe the scope will replace the "limit" message with "hack me now" when time is up? 

[Martin72]

The 1.5V peak at 70MHz comes from the higher order low-pass.

Like I said, a resonance point.
If you hack the 804 to 100Mhz now, this should disappear.
If not, this is a hardware filter, but then the specification in the options menu makes no sense...
On the other hand, there is no official bandwidth option...
Oh Rigol, you're killing me.

[TurboTom]

Actually, I still like the intensity grading of the DS1000Z series better, not to speak of the higher models of this heritage...

Settings for the AM signal were 1MHz 7Vpp, AM 1kHz 100%. Timeout / duration trigger works best with the time set to 1/5 to 1/10 of the modulation frequency period.

On the DS1000Z, enabling the anti-aliasing function helps cleaning up the display. Playing with the intensity permits intensity grading configurations that are way more pronounced than on a real analog scope. The attached photo IMO shows a very nice configuration, not exaggerated but nicely visible. I won't get it better than that on the DHO914S, though. The screenshot of the scope may actually look better than the image on the scope screen itself, That's why I took a photo of a direct comparison.

Btw, the horizontal pattern visible on the DS1000Z in the photo doesn't appear on the screen to the bare eye. It may be a dynamic thing that the still camera catches but the (my... ) eye is to slow to notice...

[TurboTom]

Okay, so here's some more testing, especially regarding timebase accuracy, frequency response and aliasing...

For my first test, I connected the (warmed-up) scope to my rubidium source (square wave output) and increased the time base delay until it was possible to comfortably measure a certain deviation of the rising slope to the screen's vertical center line. This was the case at 10ms delay, and the offset measured -62.39ns (RigolDS20). So the time base of this particular scope is running approx. 6ppm too slow. This also explains why the frequency counter shows the frequency of the rubidium source 100Hz too high (one least digit). These are very acceptable figures for a scope.

The next tests were intended to evaluate the 3dB drop-off frequency for the scope at a range of signal levels. For this, I pulled an R&S SM300 RF signal generator from the basement and wired everything up, using a BNC through-terminator at the scope. I tested at levels of +10dBm, 0dBm, -10dBm, -20dBm, -30dBm and -40dBm. Initially, I checked every level at 10MHz and stored a reference to be sure the displayed AC RMS amplitude matches the expected value (shown for +10dBm in RigolDS22).

The following screenshots (RigolDS23:28) show that the -3dB frequency threshold of this DHO914S is around 180MHz with the "sweet spot" at 100mV/div with 205MHz. So it's far beyond the manufacturer's spec of 125MHz, yet one has to consider that these measurements had been performed without the supplied probes. Moreover, there's no frequency dropoff or limiter present at the lower voltage ranges (tested down to 1mV/div).

TBC...

[TurboTom]

Let's continue with the tests...

In order to check the aliasing properties, I generated five frequencies in sequence at +7dBm (since this nicely filled the screen at a 200mV/div sensitivity setting) which were 100MHz, 125MHz, 150MHz, 175MHz and 200MHz. I meassured each of the frequencies initally with only one active channel and afterwards with all four channels enabled (btw, the reason for using channel 4 for the measurement was due to the length -- or rather the lack thereof -- of the semi-rigid connection cable from the signal generator to the scope  )
The screenshots (RigolDS29:38) are attached in sequence.

As you can see, at 100MHz everything's still "nice and dandy", so the scope is definitely usable with all four channels active up to this frequency. Also the statistics of the frequency counter doesn't show any deviation. The latter is still the case at 125MHz, but the signal trace starts to show some smearing / modulation, an indication that the sin(x)/x reconstruction isn't working quite as good anymore. At 150MHz, the counter deviates heavily from the correct value and there's a lot of modulation on the signal. Here we're far beyond the useful rage of the scope with four channels active. With up to two active channels, though, the scope is performing flawlessly. Beyond 150MHz, the modulation reduces and the scope appears to show a proper waveform, except now it's aliasing completely, visible in the displayed frequency dropping with increasing input signal frequency. Since the scope's front ends work beyond the aliasing limit, there's no direct hint to this problem visible on the screen  -- the user has got to check frequencies with two or less channels active.

Another detail to observe is that an active counter or trigger source assigned to an otherwise inactive channel enables this channel internally. Even the frequency counter operates on the digitized data stream after the ADC(s)!

So to cut a long story short, consider the DHO800/900 with more than two channels active to be a 100MHz scope, and you are on the safe side. If it makes much sense to spend the premium on the DHO924(S) can definitely be questioned. I probably wouldn't even apply a hack regarding the frequency upgrade if one should become available.

[TurboTom]

Unfortunately, here's one more, rather disappointing finding: The true vertical resolution of the DHO900 (at least of the specimen that I've got that is) is rather poor:

Just let the scope auto-trigger in the 1V/div range (in my particular case, I capped the input with a 50R BNC terminator) and stopped the digitizing engine then.
After that, zoom the vertical scale to 20mV/div (that's the last step provides proper magnification). You will find a peak-to-peak noise in the ballpark of 80mV! If the range at 1V/div is 10Vpp (considering one division of margin beyond the screen limits), this means there's a resolvable range of 1:125 which equals just 7 bits of true resolution (I know, this calculation isn't completely correct, but for "home use", it should fairly accurately fit the bill). This figure is more or less independent of time base setting and initial sensitivity.

It may be interesting to check if other HDO800/900 specimen perform better, especially since the test is rather easy to perform.

[rpro]

It may be interesting to check if other HDO800/900 specimen perform better, especially since the test is rather easy to perform.

On my DHO804, using your parameters (20us/ timebase, 1.25GSa/s) the noise of the 1V/div stopped 50ohm capped signal I can measure (as std-deviation or AC-RMS) is 9.5mV at 20mv/div, 1.1mV at 2mV/div and 260uV at 500uV/div.  (I prefer to use std-dev to measure noise, rather than Vpp, since Vpp could be reflecting real tail excursions in the sampled distribution, which can make the signal Vpp noise look "better" in less capable scopes.)   

[DaneLaw]

Quick and dirty, intensity grading....looks very good.
But switching to color grade mode will deactivate the intensity adjustment.
(DS5: 25Mpts, DS7:1Mpts)

Seems weird if you lose the intensity grading in color-corrected temperature mode.
How many levels of intensity grading do the DHO800 & DHO900 series operate with? just 64 levels, like DS1054Z or even less.?
Wonder about the step-up model DHO1000 series (which I have had my eye on, because of the big screen) would be a shame if waveform-eyepron, like intensity grading, were somewhat limited on that big gorgeous screen.
The DHO800/900 series does have an inbuilt drive, If I recall alongside ext thumb drive ability.... Can you record videos, if you wanna show others how a waveform fx behaves in certain scenarios..?

Nothing pops up checking Google for gradings levels on DHO800 & 900 series:?

//
Example of another older Android-based entry scope, I believe it's 256 levels, even though I saw a seller list it as 512 levels, but with full intensity control also in color-corrected CCT mapping.

[Fungus]

-Only one "normal" USB port, which I consider a big minus.

I bought one of these a few weeks ago:


Hope it works with the Rigol...

(It seems to have a "hub" chip hidden inside it somewhere)

[Fungus]

So to cut a long story short, consider the DHO800/900 with more than two channels active to be a 100MHz scope, and you are on the safe side.

That's obvious from the sample rate.

If it makes much sense to spend the premium on the DHO924(S) can definitely be questioned. I probably wouldn't even apply a hack regarding the frequency upgrade if one should become available.

It's a 250MHz 'scope when there's only one/two channels enabled.

If you:
a) Work with 50 Ohm stuff and know to stick to two channels when doing that.
b) Need four channels for working with probes at other times.

Then ... it might be worth it.

For everybody else? Probably not.

[ebastler]

Quick and dirty, intensity grading....looks very good.
But switching to color grade mode will deactivate the intensity adjustment.
(DS5: 25Mpts, DS7:1Mpts)

Seems weird if you lose the intensity grading in color-corrected temperature mode.

Looking at screenshots from various scope brands on the web, they all seem to show color grading without putting intensity grading on top, and I think that is what you want for clarity. For me, the whole point of using color grading (rather than intensity grading) is that you want to also see the infrequent signals clearly, in "cooler" colors but not dimmed. The frequent traces will still be clearly highlighted by their "hot" colors.

Which does not rule out the possibility that some scopes may let you combine both grading modes, and their combined use is just never shwon in screenshots. Anyone with hands-on experience who can comment?

How many levels of intensity grading do the DHO800 & DHO900 series operate with? just 64 levels, like DS1054Z or even less.?
Nothing pops up checking Google for gradings levels on DHO800 & 900 series:?

The datasheet states "Digital phosphor display with real-time 256-level intensity grading". Let's hope the developers have read that too... (Google finds that nicely for me, btw.)

[Serg65536]

Rigol DHO800 Aliasing effect
Press "Default", use 12.5 MHz (or 25, 20, 10, 5 MHz), 1k memory depth, increase the timebase until aliasing appears. Firmware version 1.00.
https://youtu.be/D_kHEUk-zKo

[Martin72]

with hands-on experience who can comment?

Rigol MSO 5000 and Siglent models could do this.
The first I've shown here long ago, the second I could post it here in the later evening.

[ebastler]

Just let the scope auto-trigger in the 1V/div range (in my particular case, I capped the input with a 50R BNC terminator) and stopped the digitizing engine then.
After that, zoom the vertical scale to 20mV/div (that's the last step provides proper magnification). You will find a peak-to-peak noise in the ballpark of 80mV!

Hmm... Dave did the same measurement in his review video, starting at 1V/div and zooming in to 50 mV/div, then 20 mV/div. I would eyeball the noise at 20 mV peak-peak in his quick test, so 2 bits more than what you found. Might there be something wrong with your scope -- a power supply issue maybe?

Code: [Select]

[url]https://www.youtube.com/watch?v=S8jrpCoZyx8&t=2130s[/url]

Edit: Trying to fix that Youtube link to point to the relevant time in the video...

[TurboTom]

...
Hmm... Dave did the same measurement in his review video, starting at 1V/div and zooming in to 50 mV/div, then 20 mV/div. I would eyeball the noise at 20 mV peak-peak in his quick test, so 2 bits more than what you found. Might there be something wrong with your scope -- a power supply issue maybe?
...

The noise readings actually depend a lot on the particular configuration of the acquisition engine (sample rate, memory depth). Indeed my own findings pretty accurately match @rpro's so I don't think there's anything wrong with the scope that I'll probably return to the vendor... 

[2N3055]

It may be interesting to check if other HDO800/900 specimen perform better, especially since the test is rather easy to perform.

On my DHO804, using your parameters (20us/ timebase, 1.25GSa/s) the noise of the 1V/div stopped 50ohm capped signal I can measure (as std-deviation or AC-RMS) is 9.5mV at 20mv/div, 1.1mV at 2mV/div and 260uV at 500uV/div.  (I prefer to use std-dev to measure noise, rather than Vpp, since Vpp could be reflecting real tail excursions in the sampled distribution, which can make the signal Vpp noise look "better" in less capable scopes.)

This might point to a bad design. This scope seems to (again like DS1000Z did) derive measurements from screen decimated data. MSO5000 had capability to work from full data, much better design, despite being a bit noisy..

To explain: If you capture data at 1 V/div and do AC RMS or P-P measurement then stop the scope, changing V/div on stopped data must not change any measurement. Only thing that could change will be number representation (on different ranges it will have different display).

For instance, on different scope that I have,  If I capture at 1V/div I get 5,0484 mV AC RMS. If I view that at 10 mv/div I get 5,048367 mV, and at 10V/div I get 5,083 mV RMS. That is with 200 MHz BW. If I go to 500 µV/DIV it show over-range as it should.

If DHOX00 changes value it means that is another thing done really wrong....

@ebastler, it is not visual trace thickness that is wrong but automated measurements.
@TurboTom, time base settings are only one that in theory should change AC RMS of noise (BW changes). Rest of changes are because front end creatively combines ADC, VGA and attenuators for whole range of input sensitivities and not all combinations are best case. But S/N ratio should stay same order of magnitude except for most sensitive ranges....

P.S. did anybody notice in that video how Dave compares how blocky Siglent trace looks despite using 3 Bit ERES? Problem is that Siglent is set to 0.5 Bit ERES not 3 Bit.... If it where 3 Bits it would look much better, if not the same in reduced BW that gives...

[ebastler]

On my DHO804, using your parameters (20us/ timebase, 1.25GSa/s) the noise of the 1V/div stopped 50ohm capped signal I can measure (as std-deviation or AC-RMS) is 9.5mV at 20mv/div, 1.1mV at 2mV/div and 260uV at 500uV/div.

I understood from Dave's review video that the scope will not actually "zoom in" any further below 20 mV/div, although it claims to change the vertical scale. Are the 2mV and 500µV values really meaningful, or did I misunderstand?