Rigol's New DHO800 Oscilloscope unbox & teardown

[Warhawk]

Am I the only one who gets heebie-jeebies because of the SD card? I assume this is same for the 800/900 family. What about the 1000 family?

No. You're not alone. Tapped the casing over to hold it in place...

But that is because of Raspberry Pi's on an SD. It is guaranteed to get corrupted at some time and then just stops working.

True it is easy to backup. But I suspect you'll have to refresh the card every now and then.

Raspberry Pi is actually what I was thinking about when sharing my thoughts. It is not the question IF it fails, but WHEN it fails. Not sure what they really did here but still. At the moment, I am looking more at Siglent but still trying to sell my ds1074z because they will become worthless, I guess   

[EEVblog]

One additional thing I noticed in the EEVblog 1563 teardown video:  One can machine their own alu heatsink better suited for a slim 120mm fan, say SilverStone FN123 (120x120x15mm, 12V, no PWM).  I suspect the existing fan is 5V – it would've been nice if you'd checked! –, so an adjustable 5V to 7V-12V boost circuit for the fan (adjustable, so one can optimize the fan RPM, as these are voltage controlled fans) would be needed, but with this mod, it could be rather quiet.  Which I like.

According to @hubertyoung's teardown photos, the fan is a 12V FirstDO FD5010H12S:
https://www.eevblog.com/forum/testgear/rigols-new-dho800-oscilloscope-unbox-teardown/msg4976278/#msg4976278

it might be possible to find a quieter fan at the same size/rpm, 2-pin or 3-pin should work (just leave the third pin unconnected as it is just for reporting the fan speed back to the PC). Certainly there is a huge range in noise levels in PC fans and buying a quality fan from Gelid or Noctua can really improve things. I'd certainly try hacking the case off a £6 Gelid Solutions Silent 5 and bodging it in before I machined a new heatsink

Some similar ones here:
https://www.pchub.com/server-frameless-gpu-fan-c1225?viewmode=grid&orderby=15&pagesize=72&pagenumber=6

[EEVblog]

This Delta at Digikey perhaps? I'll have to check tomorrow.
https://www.digikey.com.au/en/products/detail/delta-electronics/EFB0512HHAFAH/9974252

[Aldo22]

No. You're not alone. Tapped the casing over to hold it in place...

But that is because of Raspberry Pi's on an SD. It is guaranteed to get corrupted at some time and then just stops working.

True it is easy to backup. But I suspect you'll have to refresh the card every now and then.

It probably depends on how much/often is written to the SD card.
Of course, this is often the case with a complete operating system with user file system (Linux / Raspberry Pi OS).

Read-only, it can probably be used "forever".
The question is what the Rigol does.

Apart from that, my Raspberry Pi runs from SSD. 

[Warhawk]

...

....
Apart from that, my Raspberry Pi runs from SSD. 

But boots from SD card which is the biggest PI's flaw :-p

[Njk]

All that is possible BUT now there is already (at least) a Dave's copy in the wild (accessible by anyone). So, to prevent this copy there would have to be a HW change to the platform.

I'm not so sure about the changes. What if private key for secure boot is already fused in every rockchip. But I agree it's very unlikely to happen because no one wants this. The strategy is entirely different. They can't compete with the technology pioneers on performance, but they can win on affordability, which is usually associated with different performance, poor usability, poor documentation and liberal amount of immortal bugs. In most cases, you can navigate through the problems but it takes time. I don't care but those on schedule are definitely not happy with that. So the goal is not to make money but to achieve dominance by slowing the customers down. They could give the toys for free but that would be looking too suspicious. That's the conspiracy.

[Aldo22]

...

....
Apart from that, my Raspberry Pi runs from SSD. 

But boots from SD card which is the biggest PI's flaw :-p

No, Rpi 4 can be set up to work without an SD card, booting from SSD.
Here is a screenshot from my RPi (older).

[RoGeorge]

But boots from SD card which is the biggest PI's flaw :-p

Newer RPi can be made to boot from LAN or from USB without any SD card inserted.  Didn't try, but seen that advertised.
(e.g. Boot and Install Raspberry Pi Over Internet - No SD Card Needed! first search result for 'booting RPi without sdcard').

[mawyatt]

BTW, if all is inside the card, then Rigol has created a RPi style scope!!!!   

This should be the HW platform for those that are interested in assembling/compiling their own software solution with their own custom OS!

Interesting...

So, to all Rigol bashers, the path is here to stop complaining about their developers and start putting the effort.

Being like a RPi is huge IMO, and will allow custom applications and hacks to be developed by many folks with lesser software development skills, like ourselves. We've had some success in developing custom hardware and software solutions with the RPi by leveraging off the enormous RPi software knowledge base.

As we've mentioned before, these modern DSO/MSOs are data collection and display platforms in disguise as DSO/MSOs. These new Rigol HDs remove the disguise and revels the basic raw data data collection innards for those folks that want to tinker with such

This also bodes well with the necessary high quantity usage of the custom chip-set to cover the extensive NRE cost to develop such.

Might be a very clever business plan by Rigol, initially make the huge investment in the custom chip-set applicable over a broad range of instruments. Then develop various hardware implementations with these chip-sets including one DSO that is very cost friendly (large chip-set usage) and make these easily hackable, almost as Open-Source, and allow the end users to assist in the software debug and evolution with the added benefit of newly developed applications by those very end users!!

This save's considerable present software costs in debugging and firmware updates, and places the "fixes" and "updates" directly within the hands of capable end users to "assist" in such. The "Product" is released early to generate cash flow and create a core user base to provide the "assistance"!! Just the opposite of Tier One OEMs (and Siglent) for example, where a major cost factor in the price is software maintenance/updates and fixes, and considerable effort is placed in the inital software development/refinement/debugging before releasing the end Product.

Clever business model, hope this works, it certainly did with the RPi

Best,

[Fungus]

Because it is relatively small piece of not too thick aluminium (compared to full rear shell), needing only 2D/2.5D machining on each side, thus doable even for a hobbyist like myself with the help of the local Hacklab and their mill (or CNC mill).  Full rear shell?  Complicated, and much larger.

It would be much easier to just machine off a few of the fins so a larger fan can fit in the middle.

If you look at the other side it has be precisely made to exactly touch all the chips, fit around the inputs, line up the screw holes, etc.

[Fungus]

Read-only, it can probably be used "forever".
The question is what the Rigol does.

I don't imagine it has a swap file like the RPi does.

Android will also be more optimized for running from flash memory - almost every Android device depends on it.

[tv84]

Might be a very clever business plan by Rigol, initially make the huge investment in the custom chip-set applicable over a broad range of instruments. Then develop various hardware implementations with these chip-sets including one DSO that is very cost friendly (large chip-set usage) and make these easily hackable, almost as Open-Source, and allow the end users to assist in the software debug and evolution with the added benefit of newly developed applications by those very end users!!

Allowing Rigol to easily incorporate those apps in their more closed platforms. If this is the strategy, then it seems a very good idea.

OTOH, being agile in the Rigol sdcard platform will allow homebrewers to feel more confortable and migrate a full solution into those closed platforms (DHOs, 5000, 7000, 8000, etc.).

[dreamcat4]

in terms of the android, the way forwards typically would be (not always, but for certain supportable hardware)....

* make a build of lineageOS for this basic SoC, to get it to boot, with the lineageOS kernel
* try to then re-add back support any unsupported hardware (back into the build). Might require copying some pre-compiled libraries or drivers. Or compiling from "source code" that is elsewhere being published by Rigol under copyleft GPL terms

Then once that you have a (hopefully) working lineageOS build, you re-add back any scope-specific stuff too. That is unique to these Rigol over and above the general purpose IO etc...

And then for customs layer(s) ontop. You could either be installing standard android apps. And optionally something like distrobox / docker. As an extra userland layer to run (ontop) user level linux apps. Which then is in it's own linux environment. Perhaps also with some limitations due to the kernel underneath being android.

So that is generally the approach, at least when the hardware isn't so proprietary that those task(s) becomes too challenging or difficult.

[mawyatt]

Might be a very clever business plan by Rigol, initially make the huge investment in the custom chip-set applicable over a broad range of instruments. Then develop various hardware implementations with these chip-sets including one DSO that is very cost friendly (large chip-set usage) and make these easily hackable, almost as Open-Source, and allow the end users to assist in the software debug and evolution with the added benefit of newly developed applications by those very end users!!

Allowing Rigol to easily incorporate those apps in their more closed platforms. If this is the strategy, then it seems a very good idea.

OTOH, being agile in the Rigol sdcard platform will allow homebrewers to feel more confortable and migrate a full solution into those closed platforms (DHOs, 5000, 7000, 8000, etc.).

Rigol doesn't have to worry about someone copying their hardware, they are protected by the custom chip-set, and as for the software and such they would encourage this as it just increases their hardware sales!!

All in all this reminds of the early DSO introductions which were based upon an AT intel 286 motherboard with a custom PCB module that inserted into one of the AT long slots (two edge connectors). Also reminds of our first DSO back in early 80s, ET-101A which we developed and was a Digital Scope and SA based upon the Apple II, Open Source and inserted into an Apple motherboard slot.

"Back to the Future" all over again

Best

[tv84]

Also reminds of our first DSO back in early 80s, ET-101A which we developed and was a Digital Scope and SA based upon the Apple II, Open Source and inserted into an Apple motherboard slot.

"Are you telling me that you built a time machine... out of a DeLorean?"   

[gjvdheiden]

But boots from SD card which is the biggest PI's flaw :-p

Newer RPi can be made to boot from LAN or from USB without any SD card inserted.  Didn't try, but seen that advertised.
(e.g. Boot and Install Raspberry Pi Over Internet - No SD Card Needed! first search result for 'booting RPi without sdcard').

Yes, that works. My Pi4 is running with a SSD via an USB adapter, no SD card in the slot. It runs continuously, 24x7. Runs my custom KNX to HomeKit server, so big inconvenience if it's kaput. Starting from SD and then a real hard drive also works. Looks like the whole filesystem is on the SD of the Rigol scope, let's indeed hope Andriod has a clever mechanism. Directories as e.g. "lost+found" from the listings aren't read-only.

Few days ago the power adapter blew, like DevPom's YouTube channel says: "It's always a capacitor". The metal cap of the electrolytic capacitor was lifted. It still measured capacitance and no real ridiculous high ESR. The supply input of the Pi's itself never been great imho.

[Nominal Animal]

If you look at the other side it has be precisely made to exactly touch all the chips, fit around the inputs, line up the screw holes, etc.

You can use a flatbed scanner to precisely measure the heatsink, and there are only a few different Z-heights, so I don't think that is a big problem.

Flattening the fan side is not a good idea, as its surface area is already quite low with thick fins.

Could anyone measure the heatsink?  A rough measurement indicating how large a rectangular piece one needs to cover it or mill a new one from would be very nice.  Edit: Plus the distance from the heatsink to the inner edge of the plastic rear shell, i.e. the amount of room between the shell and the heatsink?

[mawyatt]

Also reminds of our first DSO back in early 80s, ET-101A which we developed and was a Digital Scope and SA based upon the Apple II, Open Source and inserted into an Apple motherboard slot.

"Are you telling me that you built a time machine... out of a DeLorean?"   

Nope it was a Porsche 911

Needed to go way faster than 77MPH

However the darn "Flux Capacitor" kept blowing up

Best,

[mawyatt]

If you look at the other side it has be precisely made to exactly touch all the chips, fit around the inputs, line up the screw holes, etc.

You can use a flatbed scanner to precisely measure the heatsink, and there are only a few different Z-heights, so I don't think that is a big problem.

Flattening the fan side is not a good idea, as its surface area is already quite low with thick fins.

Could anyone measure the heatsink?  A rough measurement indicating how large a rectangular piece one needs to cover it or mill a new one from would be very nice.

Could a hacked heatsink be 3D printed with high thermal conductivity filament if there is such ?

Best,

[dreamcat4]

Could a hacked heatsink be 3D printed with high thermal conductivity filament if there is such ?

no because those so-called conductive plastic substrate filament isn't conductive enough. in theory yes you could metal print (in a real 3d printed metal). but that option would be prohibitively expensive. typically cnc is the way to go.

[mawyatt]

An interesting experiment would be to place a thermistor on the heatsink and read the temp when the fan is running, then unplug the fan and check the temp rise. If the rise isn't too great then one might risk operating in a cool office environment without the fan enguaged, of course the risk is having one of the custom chips degrade/die due to overheating. Maybe Dave is listening

Best,

[Fungus]

I still think a large, external fan is the best option. There has to be a low profile 92mm fan out there that will fit the VESA screw holes via a 3D printed adapter.

eg. This: https://www.scytheus.com/kaze-flex-92-silm

Need to measure the available height though.

[Nominal Animal]

U-boot can be used to boot RK3399.  RK3399 itself has an eMMC (5.0 and 5.1) interface that supports HS400, HS200, DDR50, and some older "legacy" modes, and two SD/MMC/SDIO interfaces (SD3.0, MMC 4.51).  As these are built-in to the RK3399 SoC, USB, GbE, HDMI, and mass storage support is already in vanilla kernels.

[Fungus]

An interesting experiment would be to place a thermistor on the heatsink and read the temp when the fan is running, then unplug the fan and check the temp rise. If the rise isn't too great then one might risk operating in a cool office environment without the fan enguaged, of course the risk is having one of the custom chips degrade/die due to overheating. Maybe Dave is listening

The risk of hot-spots is close to NULL with that heat sink.

You could measure the temperature anywhere on it to control a fan via a small uC.

[RoGeorge]

On the spot idea, I wonder why nobody is using a tall chimney-pipe as a passive self-ventilation system.

Should work and since no fan, should be completely silent.  Except maybe for once a year, when from the chimney-cooler might come noises of  "Ho, Ho, Ho!".