Rigol DHO800/900 Oscilloscope Bug Reports + Firmware

[ptluis]

That's normal behavior for all oscilloscopes. It's due to the input amp overload. You need to remove input circuitry overload, to see the undistorted signal.

They should create a message limit to avoid this situation, in my old GW-instek this situation simply blanks the screen, on the cheap zeeweii dso2512g it shows a limit message and you can't change the vertical anymore. This kind of small details make the experience of using the scope much better.

[2N3055]

Ax+B is industry standard name for that function. With uppercase A and B and lower case x

Please, check other operators on the DHO800 (A+B, A-B, A||B, etc.): all other operators show A and B as actual captured waveforms, but in the operator "AX+B" letters A and B represent the numeric constants.
So, to prevent confusion, either the mentioned operator should be renamed, or all the other.

PS: I would prefer mathematical conventions: "x y" - for the waveforms, and "a b c" - for the constants.

I understood you the first time. I understand your logic and it is not wrong, if observed in isolation...

My point is, that in order to avoid confusion, name should be kept same as on other 100 scopes from different manufacturers.

Those are names of functions and A and B are operands, whatever they might be.. That is why it is named as it is now...

And I agree mathematical way would be prefered. In which case all the functions should be properly named etc..
But sometimes there is established practice and if others have been naming something for 20 years in certain (suboptimal) way it sometimes becomes de facto standard..

[Fungus]

I just found another bug: Only channel 1 can be set to the full 1.25Ghz sample rate. The other three channels only go to 625MHz sample rate even when they're the only channel enabled.

That's three bugs on my list:
1) The serial decoder threshold voltage doesn't factor in the probe attenuation, it's only correct when probe is set to 1x mode.
2) The setting for fine/zoom (the action when you push the horizontal timebase knob) isn't remembered when you power off.
3) Channels 2,3,4 can't be set to 1.25Ghz sample rate.

[Fungus]

3) Channels 2,3,4 can't be set to 1.25Ghz sample rate.

Did you change the trigger to the active channel as well?

Aha! Good question. It can still trigger off invisible channels...

Yep, that was it! 

That's only two bugs on my list then:
1) The serial decoder threshold voltage doesn't factor in the probe attenuation, it's only correct when probe is set to 1x mode.
2) The setting for fine/zoom (the action when you push the horizontal timebase knob) isn't remembered when you power off.

[TurboTom]

Moreover, you need to deselect all measurements, especially the "hardware" frequecy counter that may (still) be assigned to another channel. It appears that in the DHO800/900 everything is being done on the "digital side" of the signal, i.e. after the ADC. So if anything's still assigned to another channel than the displayed one and is active, it will cause the ADC to work with its sampling rate split over the affected channels.

[Fungus]

Moreover, you need to deselect all measurements, especially the "hardware" frequecy counter that may (still) be assigned to another channel.

I just tried it and ordinary measurements show asterisks on hidden channels and don't affect the bandwidth:


Frequency counter shows asterisks but still reduces the bandwidth.

[2N3055]

Data point: on some scopes (Keysight Megazoom) they have analog comparators in parallel with ADC. That same comparator is used for hardware decoding. In which case they can have counter running from that, and not drop sampling rate.
Most modern scopes with full digital trigger work from ADC data and use similar processing block (or reuse one) to edge digital trigger.
So they will use channel like if it is enabled, just for trigger and counter.
That is just how it is.

[Serg65536]

Frequency counter shows asterisks but still reduces the bandwidth.

No, it doesn't reduce the sample rate (according to your picture also).

DVM and counter (from the main menu) and trigger activates the channel in the background. Measurements from the "measure" menu use screen data, but do not occupy the channel.

[Fungus]

Frequency counter shows asterisks but still reduces the bandwidth.

No, it doesn't reduce the sample rate (according to your picture also).

Oooops, I meant "sample rate", not bandwidth! 

625MSa/s is the max I can get on C2 when the counter is enabled on C1:

[Mechatrommer]

625MSa/s is the max I can get on C2 when the counter is enabled on C1:

just use frequency measurement if you dont want 625MSa/s limit on 1 channel.

[ebastler]

just use frequency measurement if you dont want 625MSa/s limit on 1 channel.

But how would a frequency measurement done on no data (from a non-enabled channel) do you any good?

If you want to view and measure the same, single channel, I would expect that you can use the dedicated counter without sampling rate penalty. If you want to view one channel and measure the frequency in another, I don't see a way around cutting the sampling rate in half.

[Serg65536]

To summarise: 
The DVM and the counter (from the main menu) or the trigger can activate the channel in the background (if it's disabled).
Measurements from the "measure" menu use screen data, but do not occupy the channel.

[ebastler]

To summarise: 
The DVM and the counter (from the main menu) or the trigger can activate the channel in the background (if it's disabled).
Measurements from the "measure" menu use screen data, but do not occupy the channel.

To repeat:
How would measurements from screen data work on a non-enabled channel?

[Serg65536]

How would measurements from screen data work on a non-enabled channel?

I don't argue with you, as far as I can see, you don't contradict my statements.   

[ebastler]

Thanks. Seeing your repetition of your earlier statement, I thought indeed you were assuming a contradiction.

I am not contradicting you either, but don't see how your second statement would be of practical use. Yes, calculating measurements from screen data does not occupy the channel. But acquiring the screen data does. So in which scenario do you get an extra measurement without dedicating sampling rate to the respective channel?

[lollokara]

I’ve seems to have found some bugs with the LA (with the probes) on the 924S (V01.01)  .

• It is impossible to move logic tracks, once enabled you get the screen full of tracks, later you can disable some, but the space between them will stay constant, you can increase or decrease size but if like me tend to use 2/3 tracks you are still stuck at the bottom of the screen, TLDR you can’t move (apart from switching them) LA tracks on the screen
• Labels do not update / get stuck on screen if you edit the size of the LA tracks. Example, enable LA choose the first 8 channels, increase to size big, edit label, the edit will not update the screen, even disabling the labels still will not update the screen. Restarting the LA by disabling it and enabling it again fixes the issue. TLDR Can’t edit LA track size if you want labels they won’t follow you.

If needed i can try to get you a video

[Howardlong]

I’ve seems to have found some bugs with the LA (with the probes) on the 924S (V01.01)  .

• It is impossible to move logic tracks, once enabled you get the screen full of tracks, later you can disable some, but the space between them will stay constant, you can increase or decrease size but if like me tend to use 2/3 tracks you are still stuck at the bottom of the screen, TLDR you can’t move (apart from switching them) LA tracks on the screen

Yes, I concur. As it stands, the LA is far from finished. Won't decode SPI either.

[Serg65536]

Small bug: auto screenshots during the pass|fail test should not contain the file path popup.

[Mechatrommer]

FFT RBW display is buggy. when increasing timescale, RBW usually go lower, but near 10-50us/div, when you go back and forth, the RBW number will change inconsistently.. setup: viewing single channel at 1.25GSa/s.. FFT span 625MHz (center 312.5MHz) FFT window is arranged at bottom of dsos signal windows for better viewing, while increasing timescale up until 50us, take note on FFT RBW for each timescale, then reduce the timescale, RBW will show different value at around 10-20uS/div...

[Fungus]

I want to add a request:

Could the on-screen measurement indicators only be shown when the measurement side panel is open?

[dzungpham0703]

Hi, I am new here. Anyone can unlock 914s to 924s? I am ordering it.

[Martin72]

Hi and welcome,
But this is the wrong thread for it, here the right one:

https://www.eevblog.com/forum/testgear/hacking-the-rigol-dho800900-scope/msg5074423/#msg5074423

[dzungpham0703]

Hi and welcome,
But this is the wrong thread for it, here the right one:

https://www.eevblog.com/forum/testgear/hacking-the-rigol-dho800900-scope/msg5074423/#msg5074423

Thank you, I will try that post.

[ebastler]

I have a "feature" request: Could Rigol please be more space-conscious in the UI, and waste less space (in particular vertical space) on the small display?

With reference to the attached screenshot from the manual -- showing the DHO900 here, since it needs to accommodate the most features:

7 - Is this "menu" really still useful if it only has two visible icons? It made sense on the 10" screens, but here? Maybe move the status indicator (10) here instead?
15 - Why do we need this separate button? Couldn't the RIGOL logo in the upper left serve this purpose?
10 - Why is this field so large? It used to display date and time, but those are not available on the small scopes.
11 .. 14 - If we can make these fields a bit wider, after omitting/reducing/moving 10 and 15, can they be less tall please? Two lines of text, same total height as the bar on top of the screen.

All sub-Windows on the screen: Why does every window need its own title bar? I can see that they show Waveforms, Math and Measurements respectively; no need to spell that out for me. If additional per-window controls are necessary on the screen, put them all into one burger-style menu (per window), and let that icon float over the data area; don't waste vertical space for a dedicated title bar.

In general, I think Rigol only made half-hearted adjustments to the UI when adapting it from the 10" screen of the DHO1000 and 4000 to the 800/900 series. I do like their graphical design choices (which some might find too playful and not professional-looking). But they should not consume quite that much space on the display.

[Fungus]

Here's another bug...

Every now and again the Window transparency seems to go to 100% all by itself. I have to go and adjust it back.


I prefer it like this:


None of the other values changes, only transparency.

I haven't figured out what triggers it.

Anybody else seen this?

As a test I just set my wave intensity to 75%. Let's see if my window transparency jumps to 75% next time it happens.