Hacking the Rigol DHO800/900 Scope

[Mechatrommer]

Maybe we need to copy some ".cal" files from the 924 image to make it work, I don't know for sure. See my theory below... 
...
Theory: The extra 200uV/div range might be the reason the DHO800's cal files don't work with a DHO900's vendor.bin.

i reported earlier, DHO924 has extra files in data folder.. they are \data\cal_ext.hex and \FPGA\SPU_H12S1.bit not sure what they do, as you said, i got what i want..

...and no useless on-screen junk from the 900-series LA.

do you want 36 pages dick waving contest on this? rigol fanboy vs rigol fanboy... if i have time, i'll explore AFG and Bode plot feature, or even LA feature if i really get nasty.

The only missing thing is the extra 200uV/div range of the DHO900 but I can easily live without that.

another thing to note is when changing time scale from 2ns/div to 5ns/div, sometime on larger scale 10 to 20ns/div iirc, there is 1-2 seconds delay until the signal is updated again in DHO924 FW v 1.14, a bit annoying but i can tolerate for the sake of increased (aliased) BW... this doesnt happen in DHO804 legit FW.

[Fungus]

....i can tolerate for the sake of increased (aliased) BW... this doesnt happen in DHO804 legit FW.

Try copying all the 924's cal files to your 'scope. See if it fixes it.

[Martin72]

The only missing thing is the extra 200uV/div range of the DHO900 but I can easily live without that.

I can do without anything that is artificially zoomed in.

[ebastler]

The only missing thing is the extra 200uV/div range of the DHO900 but I can easily live without that.

I can do without anything that is artificially zoomed in.

I've been wondering what "artificially zoomed" means, actually, for a 12 bit ADC and a screen with 600 vertical pixels total, and 400 or so available for the traces.

So they only map about 1/3  of the full ADC range to the full vertical screen range. But that still leaves 3 or 4 ADC digits per vertical screen pixel. Combined with the small size of the screen, what's wrong with a zoomed-in view to show the detail?

[csuhi17]

From the DHO900 datasheet.
  [4]: 200 μV/div and 500 μV/div is a magnification of 1 mV/div setting. For vertical accuracy calculations, use full scale of 8 mV.

[Serg65536]

You can use a SD card copy from another scope of the same model on your scope with high precision.
Calibration is not secretly stored in the firmware, it's in the /rigol/data/*cal files.

Today I've flashed non original V1.14 (it's probably the oldest firmware available) to my DHO804:
1) There was initially a huge zero offset.
2) Zero offset could be eliminated with autocalibration (default mode: 1M + all ch)
3) The resulting DC precision if awful

After pushing the original cal files to the scope:
4) There was an offset initially,
5) but it's eliminated with autocalibration (default mode: 1M + all ch)
6) The resulting DC precision is perfect
7) You can update non original 1.14 to 1.01 with no issues

[Serg65536]

From the DHO900 datasheet.
  [4]: 200 μV/div and 500 μV/div is a magnification of 1 mV/div setting. For vertical accuracy calculations, use full scale of 8 mV.

So at 200 μV/div it's approx. the bit per pixel.

[Fungus]

From the DHO900 datasheet.
  [4]: 200 μV/div and 500 μV/div is a magnification of 1 mV/div setting. For vertical accuracy calculations, use full scale of 8 mV.

So... that scale doesn't have a calibration?

There goes my theory. 

[Serg65536]

To calibrate DHO924 firmware on DHO804 scope we need to place initial calibration values into the calibration files of the 924 firmware.

That's probably:
- vertical sensitivity and zero offset of the input amp for each volt per div value. That's a 15 * 2 = 30 numbers.
- vertical sensitivity and zero offset for the offset voltage source for each of the offset ranges. That's 4 * 2 = 8 numbers.

As far as I can tell, this would be much easier, than changing the vertical compensation algorithm back to DHO800 mode, to use original calibration files.

[WoodEGA]

Total newb here (Hi!), but I've been following all the discussions here since I found Dave's DHO800 videos over the weekend. The DHO804 looks like a perfect beginner scope for me. Following the threads has been a task, but let me see if I got this right:

Option 1: You can swap in the DHO914 Vendor.bin (and maybe other stuff on the SD card?) to get many options, but that frequently screws up calibration, right?

Option 2: Use hdo_tools to grab your Key.data file, generate your option strings with the BW7T10 and RLU options, then punch them into the DHO804 scope's web interface, get up to 200MHz bandwidth and some options, but calibration is not affected?

Or am I still mixed up on all this? Thanks in advance, and let me know what I might not be getting right here.

[Fungus]

Total newb here (Hi!), but I've been following all the discussions here since I found Dave's DHO800 videos over the weekend. The DHO804 looks like a perfect beginner scope for me. Following the threads has been a task, but let me see if I got this right:

Option 1: You can swap in the DHO914 Vendor.bin (and maybe other stuff on the SD card?) to get many options, but that frequently screws up calibration, right?

Yes.

nb. The calibration thing is still being worked on and may be solved in the next few days. We only got the other hacks working yesterday.

Option 2: Use hdo_tools to grab your Key.data file, generate your option strings with the BW7T10 and RLU options, then punch them into the DHO804 scope's web interface, get up to 200MHz bandwidth and some options, but calibration is not affected?

200Mhz measured bandwidth and 50MPts memory

(which is 10x more memory with all 4 channels enabled!)

I did it all manually using ADB but that's just me. I like to understand the process.

Important point: Buy with 100% confidence that it can be done.

[Serg65536]

Option 1: You can swap in the DHO914 Vendor.bin (and maybe other stuff on the SD card?) to get many options, but that frequently screws up calibration, right?
Option 2: Use hdo_tools to grab your Key.data file, generate your option strings with the BW7T10 and RLU options, then punch them into the DHO804 scope's web interface, get up to 200MHz bandwidth and some options, but calibration is not affected?

Option 2 is the only choice for now. You save the scope's accuracy and it's 100% ready for future firmware updates.
Option 1 does not allow to do calibration. It has low voltage accuracy and could be blocked in future updates.

[Mechatrommer]

7) You can update non original 1.14 to 1.01 with no issues

why people only showing the result but not how? do you mean we can run with v1.01, and then copy the DHO924S' vendor.bin and then we get DHO924 running v1.01 FW and no issue with calibration offset? how? iirc i tried that and no success...

[Serg65536]

There must be a non-volatile memory in the scope!

I flashed the memory card with a fresh copy of 924 firmware, because waveform fluffiness appeared. But vertical and horizontal settings were preserved on the first boot (even the vernier was active). And, guess what? The fluffy waveforms are there again.
The first fresh installation of exactly the same firmware copy worked fine initially.
Network settings are preserved also.
This works if you don't change the firmware type. Probably because this constant memory is used differently by different software versions.

[Fungus]

7) You can update non original 1.14 to 1.01 with no issues

why people only showing the result but not how? do you mean we can run with v1.01, and then copy the DHO924S' vendor.bin and then we get DHO924 running v1.01 FW and no issue with calibration offset? how? iirc i tried that and no success...

I think maybe you need to copy the entire 1.14 SD card image, complete with .cal files.

[Fungus]

There must be a non-volatile memory in the scope!

There's an FRAM chip to store the current settings every time you change them.

[Serg65536]

7) You can update non original 1.14 to 1.01 with no issues

why people only showing the result but not how? do you mean we can run with v1.01, and then copy the DHO924S' vendor.bin and then we get DHO924 running v1.01 FW and no issue with calibration offset? how? iirc i tried that and no success...

V1.14 is some early software, the V1.01 from official site - is the latest.
I mean you can flash your scope with DHO924S firmware from the first message of this thread, update it to the V1.01 software and "enjoy" low accuracy and high offset voltage, as you would on initial V1.14 software.

[Fungus]

Option 2 is the only choice for now. You save the scope's accuracy and it's 100% ready for future firmware updates.

It depends on what you want to achieve.

If you want maximum bandwidth with 5V signals then what's a few millivolts of offset? It's completely irrelevant.

You might want CAN/LIN bus decoding and you need a DHO900 for that.

(A few millivolts offset won't matter for that either...  )


PS: I think most people only had problems on one channel. My problem was channel 3, all my other channels were perfect.

[WoodEGA]

Thanks for the replies, Fungus and Serg65536. It's very impressive how quickly these hacks have developed. Some seriously talented folks here, for sure. Kudos to all! I'll keep reading just because it's damn fun to watch this unfold.

[mwb1100]

V1.14 is some early software, the V1.01 from official site - is the latest.

Typical Rigol firmware version control sloppiness.

[Mechatrommer]

7) You can update non original 1.14 to 1.01 with no issues

why people only showing the result but not how? do you mean we can run with v1.01, and then copy the DHO924S' vendor.bin and then we get DHO924 running v1.01 FW and no issue with calibration offset? how? iirc i tried that and no success...

V1.14 is some early software, the V1.01 from official site - is the latest.
I mean you can flash your scope with DHO924S firmware from the first message of this thread, update it to the V1.01 software and "enjoy" low accuracy and high offset voltage, as you would on initial V1.14 software.

i see now. you mean using the legit gel file v1.01 from rigol site, upgrade the "hacked to" DHO924 scope using rigol recommended method ie usb drive upgrade right? thanks.

wait a minute, but V1.01 is only upgrade a few files https://www.eevblog.com/forum/testgear/hacking-the-rigol-dho800900-scope/msg5132049/#msg5132049 this means, even if upgrade to v1.01 gel, we still using old v1.14 sparrow.apk... fyi.. i'll try later and see if sparrow.apk got updated too..

[Fungus]

I'll try later and see if sparrow.apk got updated too..

Sparrow will definitely be updated in 1.01 because some things changed in the UI.

(eg. location of probe attenuation setting)

[Martin72]

How can you actually read a bin. file ?
That would interest me.

[tautech]

How can you actually read a bin. file ?
That would interest me.

Use Siglent BIN to CSV utility.

[pickle9000]

Reading a BIN

Any old hex editor will work. Just do a search.