Hacking the Rigol DHO800/900 Scope

[helmy]

... are there any other hacks available?...

You can improve the UI of your DHO924

In this video, how did you access the launcher with that swipe-down gesture from the top edge of the screen? Also, which launcher are you using?

[mrisco]

In this video, how did you access the launcher with that swipe-down gesture from the top edge of the screen? Also, which launcher are you using?

I used Zone Launcher

[Fungus]

are there any other hacks available?

It's also an Android device so you can install a launcher and any other apps you want to.

[norbert.kiszka]

Android device

In my case, it's a Debian Linux device. This can run anything, since it's a regular ARM processor (RK3399).

[shapirus]

Android device

In my case, it's a Debian Linux device. This can run anything, since it's a regular ARM processor (RK3399).

Have you been able to run the actual scope app under debian yet?

[norbert.kiszka]

Android device

In my case, it's a Debian Linux device. This can run anything, since it's a regular ARM processor (RK3399).

Have you been able to run the actual scope app under debian yet?

Yes, but it was in a temporary way, which made it incredibile slow. In January I was close to make working environment for it, but I had to do my job and I left it as it was.

Anyway, the goal is to run (modified) existing open source app or make a new one. Rigol app working under Debian will make it much easier to do reverse engineering.

[helmy]

https://github.com/Andy-Big/Rigol-DHO800-900-Sparrow_mod

https://github.com/mriscoc/RIGOL_DHO800_DHO900_GUI/tree/master

Can anyone explain the differences between them?

[norbert.kiszka]

https://github.com/Andy-Big/Rigol-DHO800-900-Sparrow_mod

https://github.com/mriscoc/RIGOL_DHO800_DHO900_GUI/tree/master

Can anyone explain the differences between them?

Price? Everything other is visible and only one is open source (kinda... because both are based on decompiled code).

The choice is Yours.

[mrisco]

Can anyone explain the differences between them?

The extended UI version on Patreon is intended for advanced users, you can use the DHO Actions panel to run SCPI scripts, save usage profiles and easily recall them from the same panel, etc. Here is a video showing some of the features:

[mrisco]

And this is an example of the DHO Actions panel:

[helmy]

Is there a way to save the recording frames to a file? It feels like an obvious feature to have, but I haven’t been able to figure out how to do it. The only related option I found was under “Storage,” which seems to deal with saving waveforms — but nothing that explains how to save a full “recording” of, say, 200 frames all at once to disk.

If this isn’t currently possible, would it be feasible to add this as a feature in the extended UI?

[dzwer]

Did you try to configure in the Quick menu for Recording and use the Quick button on the front panel?

[norbert.kiszka]

Finally today I added more capacitors parallel with others - after switching regulators and after LDO (is it LDO?). Just in ~11 places, which is less than half of my plans. Result is 2-3 times lower noise on modified channels (3 and 4 - increased bandwidth). Unmodified channels has 1-2 % less noise...

Now it's way more usable with normal acquisition instead of average acquisition (last screenshot).

[helmy]

Finally today I added more capacitors parallel with others - after switching regulators and after LDO (is it LDO?). Just in ~11 places, which is less than half of my plans. Result is 2-3 times lower noise on modified channels (3 and 4 - increased bandwidth). Unmodified channels has 1-2 % less noise...

Now it's way more usable with normal acquisition instead of average acquisition (last screenshot).

Where can I find more information about adding more capacitors? and modifying the channels for more bandwidth.

[helmy]

Did you try to configure in the Quick menu for Recording and use the Quick button on the front panel?

Yes, I tried that, but it only saves the current waveform regardless of whether the record window was open or not. What I'm looking for is how to save all the recorded frames to either one or multiple files

[helmy]

I think there might be an issue related to sleep mode and potential screen burn-in on the oscilloscope. If this has already been discussed, I’d appreciate a link to the relevant thread.

I left the oscilloscope in sleep mode for several hours. When I pressed the power button to wake it, I noticed the RIGOL boot logo appeared with a noticeably brighter area around it. That lighter “shadow” remained visible even after the system fully loaded.

Running a screen test confirmed it — the RIGOL logo from the boot screen was clearly visible during solid color tests. The ghost image faded gradually after repeating the screen test a few times.

Interestingly, while the oscilloscope is in sleep mode, I can still access Web Control via a browser. The remote screen shows the Rigol logo displayed on a black background with a spinning animation. So the screen isn’t truly off — it’s just dimmed to 0 brightness while still displaying the same static image for hours.

This seems to explain the temporary burn-in effect.

[norbert.kiszka]

Where can I find more information about adding more capacitors? and modifying the channels for more bandwidth.

I have 470 nF from Samsung (CL21B474KBFNNNG). I just added to missing places and on top of another. Off course only bypass capacitors. First 12 capacitors gave me result as in my previous post. Currently Im adding more as an trial and error nothing. Still I didn't added anything close to AFE or ADC - that theoretically should give best results - maybe I will do this today.

On this forum, there was a link to a Russian forum, where somebody changed LC filters (just below ADC) from 250 MHz to 400 MHz. Increasing BW to 400 MHz gives little more noise. After removing first or all two stages, noise is quite big, very big. Which makes it almost unusable, unless You are using averaging acquisition or you will decrease noise in another way (like I did). Also, You have to remember, that there is a capacitor before those two stages, added for line termination at higher frequencies (coils impedance). So You have to change it value accordingly - I remember it was something like 10 or 15 pF.

From my own reverse engineering, looks like most noise comes from AFE power supply - as I wrote here before.

I think there might be an issue related to sleep mode and potential screen burn-in on the oscilloscope.

Script quick_boot_test.sh (/rigol/shell/quick_boot_test.sh) and Linux kernel modified by Rockchip, gives some answers. Since GPU, VOP, MIPI, panel-simple, etc drivers are compiled as built-in, LCD screen most likely will not be switched off. If mipi driver will be compiled as module, unloading it (rmmod command) should switch off the panel completely (I made a port of RK3399 MIPI driver from one kernel to another, so I learned code in this module a bit).

Rigol didn't gave source code of kernel, beside of GPL licence that requires them to do it. But I did not one, but two kernels with recreated drivers (based on decompiled code and some tests), so theoretically You should be able to make Your own kernel. Also there was a post of compiling kernel modules from (maybe) original kernel, so maybe it's also an option (with or without drivers that I did).

[iMo]

In case of switchers there it would be better to add an LC or CLC filter (would need to cut the trace, however)..

[norbert.kiszka]

Low noise LDO for AFE should give better results than filters.

[norbert.kiszka]

Now I disabled AFE (same as before - commented insmod), measured noise across channels, added capacitors (bypass off course) around ADC (looks like both analog power lines) and I did same measurements. With those capacitors, noise is around 20% less than before.

[ptluis]

The oscilloscope's normal consumption is around 35W. Are there possible ways to reduce consumption to make it more efficient when using a power bank? What is the lowest consumption you can get with this oscilloscope? Has anyone try it?

[norbert.kiszka]

You can turn off PLL and AFG. It will drop to around 10-15 W.

Edit: put backlight to 0. Somewhere in /sys

[ptluis]

You can turn off PLL and AFG. It will drop to around 10-15 W.

Edit: put backlight to 0. Somewhere in /sys

The DHO804 doesn't have AFG, my mistake, didn't specify the model I'm refering to. Even lowering the backlight the best I can is 31W.

[norbert.kiszka]

Ok.. so I did some experiments with adding capacitors parallel to existing ones (decoupling only off course) and even I put wires to decrease ground impedance and... I did false positive - as on my previous screenshots from scope, when noise was 2-3 times lower (on modified channels only).

So, at some point I noticed noise came back to original value  After some hard thinking, I noticed, that previously I omitted two screws from heatsink.

Finally I developed couple possibilities to make lower noise (practically only on channels with increased bandwidth) - I can't say which one is the best and which one will work at less modified scopes than mine:

1. Remove screws 3 and 4.
2. Remove screw 1.
3. Place electrical tape on shield (between PCB and front panel) under screws 2, 3 and 4.
4. Point 2 and 3 - that's what I have right now.

I also did experiments with other screws, but finally I spent too much time on it. Maybe somebody else will develop something better.

[Thripster]

I thought he stated at the start that the power bank is 20,000mAh?