Rigol DHO804 Test and Compare Thread

[nctnico]

i see tea'ism comes to the party now, after siglenism... find Dave's video in the early age... DSO is not for Vdc accuracy... why this has to happen so many times? why?

The selling point of 12bit DSOs is that they're more accurate than 8 bit DSOs.

No, resolution isn't accuracy.

[Fungus]

No, resolution isn't accuracy.

It is when it's doing self-cal. 

[Mechatrommer]

i see tea'ism comes to the party now, after siglenism... find Dave's video in the early age... DSO is not for Vdc accuracy... why this has to happen so many times? why?

The selling point of 12bit DSOs is that they're more accurate than 8 bit DSOs.

No, resolution isn't accuracy.

true, but probably he's not talking about 12bit is being accurate. he probably talk about self-cal algorithm and voltage/frequency reference used.

i see tea'ism comes to the party now, after siglenism... find Dave's video in the early age... DSO is not for Vdc accuracy... why this has to happen so many times? why?

The selling point of 12bit DSOs is that they're more accurate than 8 bit DSOs.

if its that important, then need to plot dso's frequency respond as well, it may accurate at DC but wobbly respond in frequency domain. if you get that kind of scope, i great you congratulation!

[Martin72]

Needless to say we were impressed then as we are now with the Rigol DHO824 measurement capabilities, altho the FFT needs some polish, however our client isn't interested in the FFT (nor Bode Function) so for us a moot point in that respect.

If you disregard FFT and the missing Bode, this is also a really nice entry-level scope.
Firmware-wise, FFT can be improved and Bode could at least be implemented with an external source to help, but I don't believe that rigol will change much in this regard.
But as you wrote, if there are people who do not need that, that can be the purchase decision in this price range.

[mawyatt]

Needless to say we were impressed then as we are now with the Rigol DHO824 measurement capabilities, altho the FFT needs some polish, however our client isn't interested in the FFT (nor Bode Function) so for us a moot point in that respect.

If you disregard FFT and the missing Bode, this is also a really nice entry-level scope.
Firmware-wise, FFT can be improved and Bode could at least be implemented with an external source to help, but I don't believe that rigol will change much in this regard.
But as you wrote, if there are people who do not need that, that can be the purchase decision in this price range.

Agree the 800 series is a good entry level DSO!! Since the 800 series has no AWG, nor Bode capability, won't discount it for that. Rigol could implement an external AWG control and enable Bode Feature in the 800 series, but from what's been shown with the 900 series wrt the Bode Capability, doubt that's going to happen.

The FFT isn't that bad, and with some work could become a nice usable feature IMO. Again that's up to Rigol if they want to invest any resources ($) into improving the FFT capability.

Or clients intended 824 use involves viewing various complex waveforms for comparison purposes, and the measurement capability could yield an easier Go/NoGo decision for end users wrt to waveform interpretations.

Best,

[Fungus]

Agree the 800 series is a good entry level DSO!!

Yep, I'll say it's a bargain for that price. Very capable and much nicer to use than a twisty-knob 'scope.

The screen is a bit small but it goes with the overall size which I really like. I don't have any problem using it and the text is really legible - more legible than my Micsig with its bigger screen

Mine hasn't crashed on me even once and there's only one bug that really affects me and it has an easy workaround (the serial decoder reference voltage bug, set probes to 1x mode to fix it).

The 12 bits and low noise really shows in actual use.

It's also been fun to hack around with and explore via ADB. 

[Martin72]

Mine hasn't crashed on me even once

I had a few crashes in the beginning, but after the firmware update not a single one - Except for the incompatibility of plugging in the usb-hub during operation.

[Fungus]

Features everybody needs to try:

I'm really liking the "Indicator" feature on the measurements, eg. here I'm indicating the preshoot on a pulse:


It moves around when you touch the measurement boxes in the panel. After using it for a while I've left it permanently turned on. 

[Martin72]

Ah, you mean the dash-dot line, right ?

[Fungus]

Ah, you mean the dash-dot line, right ?

Yes. That's not a cursor, it moves around with the signal.

Edit: It would be better if they disappeared when the measurement panel is off. I'll add that to the wish list thread...

[Martin72]

Edit: It would be better if they disappeared when the measurement panel is off. I'll add that to the wish list thread...

I know these "auxiliary lines" from the lecroy scopes with MAUI.
There they are displayed only as long as one is directly at the parameter in the measuring menu, when leaving then only the corresponding measured value is displayed.

[Fungus]

Mine hasn't crashed on me even once

I had a few crashes in the beginning, but after the firmware update not a single one - Except for the incompatibility of plugging in the usb-hub during operation.

If it does crash there's a daemon that starts it right back up again in the state it was at.

You can test it by killing the man task with ADB.   

adb root
adb shell kill -9 $(pidof com.rigol.scope)

[asmi]

If it does crash there's a daemon that starts it right back up again in the state it was at.

That's an Android thing. All apps should be ready to be suspended/killed at any time, and then restore themselves to the way they were once re-launched. Because that's how OS works on cell phones when it starts running low on memory - it kills oldest used apps, and when you switch back to them, they are supposed to restore itself to whatever state they were when they were last interacted with.

[Martin72]

Two Tone tests...
I wanted to try the other method of combining 2 signals instead of using the "wave combine" function of my generator.
There the rigol came as called...
As with the dho4204 there are a few oddities regarding FFT, but if you ignore that you get good "pictures".
The little jewel is doing quite well, at 100Mhz I got nothing more, but that's not really dramatic.
Oh well, I just don't like the color of the math channel, a selection palette would be a great thing.
So, here are some pictures of various frequencies.

[Fungus]

I like this one best:

[ebastler]

Two Tone tests [...]
So, here are some pictures of various frequencies.

Nice to see that the peak search can produce meaningful results! 

In your early exploration of the DHO800, I seem to recall that you got a very messy list of apperently random peaks which the DHO had selected. It did not seem to prioritize them by peak height at all. Have you explored that any further? Is there a (higher than expected) minimum peak/background contrast required for reliable peak identification, or some other constraint?

[Fungus]

Oh well, I just don't like the color of the math channel, a selection palette would be a great thing.

You have four to choose from:

[Martin72]

In your early exploration of the DHO800, I seem to recall that you got a very messy list of apperently random peaks which the DHO had selected. It did not seem to prioritize them by peak height at all. Have you explored that any further? Is there a (higher than expected) minimum peak/background contrast required for reliable peak identification, or some other constraint?

No, I hadn't paid attention to that, I might try that again.
Here again the former post:
https://www.eevblog.com/forum/testgear/rigol-dho804-test-and-compare-thread/msg5122722/#msg5122722
But that was also a bit "unfair", as noisy as the fundamental wave was and the rigol no FFT averaging function got.
The siglent had set most of the peaks there, but did not "forget" the other two harmonics.
I'll test this again.

[Fungus]

In your early exploration of the DHO800, I seem to recall that you got a very messy list of apperently random peaks which the DHO had selected. It did not seem to prioritize them by peak height at all. Have you explored that any further? Is there a (higher than expected) minimum peak/background contrast required for reliable peak identification, or some other constraint?

Yes, you can order peaks by amplitude/frequency, set maximum number of peaks, set threshold and excursion values.


Edit: It's not very obvious how to get there. The trick is to either read the manual or to press the up-arrow at the bottom of the FFT parameter page. 

[Martin72]

You have four to choose from:

No, I want my palette ! (*footstomping*) 

[ebastler]

Here again the former post:
https://www.eevblog.com/forum/testgear/rigol-dho804-test-and-compare-thread/msg5122722/#msg5122722
But that was also a bit "unfair", as noisy as the fundamental wave was and the rigol no FFT averaging function got.
The siglent had set most of the peaks there, but did not "forget" the other two harmonics.
I'll test this again.

Yes, that's the post I was thinking of, thanks! Looking at the first picture there -- even with the noise shown, I would have expected that the scope can pick out the fundamental and the harmonics and list those first.

Agree that Rigol would do themselves a big favor by introducing an option to average over the FFT spectra. From the Siglent comparisons I saw somewhere, that has so much more impact than FFT'ing over averaged traces.

Or is there an "Average" function in the Math menu, which one could set up in a second Math channel and apply to the FFT results? Couldn't find anything in the manual.

Yes, you can order peaks by amplitude/frequency, set maximum number of peaks, set threshold and excursion values.

In Martin's early measurement (link above), it looks like they are sorted by frequency, but only in a very narrow range around the fundamental. Why would the scope do that?

Edit: Typos...

[Fungus]

Can your Siglents do four simultaneous FFTs each with a different window function and peak search?



Ok, that's a bit silly but you could easily show different areas of an FFT, expand one part, etc., by having multiple windows open.

[ebastler]

That is quite neat (also the fact that it does not slow down the scope, as you mentioned in the other thread)! It does highlight the limitations of the small display though.

Hopefully, in a future software release Rigol gives some thought to proper rounding in tables, and to the choice of meaningful increments and decimal places for the axis labels. Especially with the limited space on the display, it hurts to see all those "4.9..99 kHz" table entries and the awkward values on the vertical axes. 

[Fungus]

Yes, you can order peaks by amplitude/frequency, set maximum number of peaks, set threshold and excursion values.

In Martin's early measurement (link above), it looks like they are sorted by frequency, but only in a very narrow range around the fundamental. Why would the scope do that?

I don't know but the default number of peaks is five so he obviously fiddled with the peak detection settings.

Maybe he set a very low value for "excursion" (the amount a peak has to stick out for it to be added to the table).

[Martin72]

Hi,

Tried again, the menu Fungus showed I already knew, it makes no differece between frequency or amplitude order.
But there´s the excursion button...Excursion, what´s this ?
Manual do not describe it in a manner I could understand it.
But google do:

Press Excursion to specify the amplitude above the FFT waveform's noise floor necessary to be recognized as a peak.

Aha...
Now you don't know where the noise carpet starts for the excursion, at the bottom, in the middle, rather at the top....
It is also very "thick", so the only thing that helps is to try it out.
With 35dB I "caught" at least 3 peaks, better than nothing.
But an averaging function for the FFT would be really desirable.