Rigol's New DHO800 Oscilloscope unbox & teardown

[Nominal Animal]

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.

If you don't mind drilling new holes to your fan, you can use the VESA 100 holes to mount a 120x120x15mm fan, for example Noctua NF-A12x15 (FLX or PWM); no need to rotate the fan by 45 degrees either.  The holes partially overlap with the existing ones; the 120mm fan mounting holes are 105mm spaced, and VESA 100 are 100mm spaced (so the new hole centers are 3.5mm towards the fan hub from the centers of the existing holes).

Noctua NF-A9x14 fits within the VESA 100 mounting holes (aligned, i.e. with the mounting holes at the fan corners).

The question is, how much room is there between the top of the heatsink and the plastic shell?

A hole in the plastic shell may be necessary, to have sufficient shroud around the fan output inside the scope, to ensure good airflow along the heatsink (and not escaping out between the fan and the heatsink).  I suspect the required fan speeds to cool this thing are surprisingly low with a much larger fan.

[Fungus]

A hole in the plastic shell may be necessary, to have sufficient shroud around the fan output inside the scope, to ensure good airflow along the heatsink (and not escaping out between the fan and the heatsink).  I suspect the required fan speeds to cool this thing are surprisingly low with a much larger fan.

With a big fan I'd probably try blowing air in. Let it go out the top/bottom over the heatsink.

[Fungus]

The working environment is rated to 50 degrees C so there's a bit of margin.

The VESA mount could also be used to attach a battery pack.

[Nikki Smith]

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

@Azusa pointed a thermal camera at the scope, and the rear temperatures were 31°C-59°C (88°F-138°F), but that was after an hour of running quite a demanding game. I'd guess that works the system way harder than its normal job!
https://www.eevblog.com/forum/testgear/rigols-new-dho800-oscilloscope-unbox-teardown/msg4980634/?topicseen#msg4980634

Quite possibly it only needs the fan running in a very hot lab / server room environment (or outdoors) ?

[Fungus]

I wonder how many windows can be added...

[Fungus]

Quite possibly it only needs the fan running in a very hot lab / server room environment (or outdoors) ?

With that heatsink? Very probably.

I don't remember who it was but there was a user here who did a fan swap on his DS1054Z and noted it was now "completely silent". Six months later he found out he hadn't attached the power cable to the fan.

[mwb1100]

Code: [Select]

rk3399_rigol:/ $ getprop
getprop

...
[persist.sys.usb.config]: [mtp,adb]
...
[ro.build.type]: [userdebug]
...
[ro.debuggable]: [1]
...
[sys.usb.config]: [mtp,adb]
...


Didn't notice any comments on this getprop output.  A debuggable userdebug build with adb enabled on USB means you should be able to do nearly anything to the system pretty easily.

[mwb1100]

I don't remember who it was but there was a user here who did a fan swap on his DS1054Z and noted it was now "completely silent". Six months later he found out he hadn't attached the power cable to the fan.

Easiest silent running hack ever!

[ganevson]

It is very interesting with which oscilloscope Siglent will answer this challenge. And when.

[Grandchuck]

SDS1000 X HD and 1204X HD?  No idea when or prices.

[Dacian]

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.

Dave measured the power consumption and it is fairly high. Also there is likely almost no variation in power consumption so a fixed speed fan makes a lot of sense as there will be no advantage to have a temperature controlled fan.

[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.

Dave measured the power consumption and it is fairly high. Also there is likely almost no variation in power consumption so a fixed speed fan makes a lot of sense as there will be no advantage to have a temperature controlled fan.

Not thinking of a variable speed here, just On or Off. If the heatsink temperature is below a specific acceptable point then the fan is Off, if above then the fan is On. Of course what that "acceptable" temperature is unknown now. Most ICs are rated for maximum junction temperature of 125C, these custom chip-sets may be lower, and we don't known the chip-set junctions to heatsink thermal impedance nor power dissipation, so other unknowns. Some thermal analysis with some knowns might shed some light on this and IF one could RELIABILITY operate without the fan engaged.

Earlier was a reference to a measured heatsink temperature of ~60C in an office environment, so ~35C rise above ambient. Was also earlier stated that the maximum instrument ambient temp is 50C, so using a simple "linear" interpolation this would imply a heatsink temperature of ~85C under a 50C ambient. So if the heatsink doesn't rise above ~85C without the fan engaged, then the DSO should be OK. Of course this assumes a lot, including the "linear" relationship and benefit of the airflow to other components not serviced by the massive heatsink, and so on. Plus it's well known that higher temperatures cause component lifespans to reduce, so operating with the heatsink at a higher temperature has its' drawbacks.

We had considered this concept, or installing another lower noise fan in our Hioki IM3536, this instruments fan noise is obnoxious. However we decided against such due to the effect on instrument calibration, and this is a high precision instrument which the DSO is not, so affecting the DSO calibration wouldn't be as much of a concern for us.

Best,

[Martin72]

Six months later he found out he hadn't attached the power cable to the fan.

This doesn't mean that it doesn't matter if a fan is active or not.
He can also have unknowingly operated the device outside of the specifications, thus aging it faster.
Rarely is something done without reason, especially not in a cheap device, if a fan was not necessary, it would be omitted for cost reasons.
For example, I had replaced the fan in my SDS2504X-HD because the original one had a defect.
Now an even quieter fan from Noctua is installed and that does not really make me happy.
He is also weaker than the original.
That does not mean anything if reserves were planned in the thermal design, but I do not know that.

Most ICs are rated for maximum junction temperature of 125C

That´s mostly MIL-Spec, common ICs are rated for 85°C.
In both cases, this does not mean that they remain in the specifications(keyword derating) up to these temperatures, they just do not go defective until then.
I may be overly cautious, but in general I would not underestimate the issue.

[pmaster]

[ro.build.type]: [userdebug] => adb root should be available.
Android 7 might not have Android verified boot enabled. So remounting of system & vendor partition would be easy. Even if Android verified boot is on, on an userdebug image adb disable-verity, adb reboot, adb remount should be possible.
The console output of
su
cat /proc/cmdline

would be of interest (UART logs might be truncated on the command line entry).
Additionally
su
cat /proc/mounts
would be nice to know.

Rigol also seems to go the easy route:

androidboot.selinux=disabled
[sys.secureboot]: [false]
SecureBootEn = 0, SecureBootLock = 0
Secure Boot state: 0

[tautech]

I had replaced the fan in my SDS2504X-HD because the original one had a defect.
Now an even quieter fan from Noctua is installed and that does not really make me happy.

Exactly what ?

SDS2000X HD fans are very low noise.
Do you still have the old one to send to Siglent for a warranty claim ?

They want to know when these problems arise.

[pmaster]

Here are the partitions in the SD card image, RigolDHO800-SDcard-dump.img:

• offset=281018368,sizelimit=134217728: Ext4 Android backup/recovery partition
• offset=415236096,sizelimit=2147483648: Ext4 Android filesystem system named system (startup stuff)
• offset=2562719744,sizelimit=16777216: Ext4, empty
• offset=2584248320,sizelimit=524288000: Ext4 filesystem named rigol, contains interesting Rigol stuff
• offset=3225419776,sizelimit=28494004224: Ext4 filesystem, rest of the Android filesystem

Thanks for figuring out the offsets. Some more details on the partitions:
offset=415236096,sizelimit=2147483648:  => /system partition which contains the Android framework (="operatingsystem" without kernel)
offset=2584248320,sizelimit=524288000: => Rigol proprietary partition. DHO800_DHO900_Update.GEL is a tar.gz file which seems to contain the parts of that partition. There seems also some calibration data stored in the data folder. There is also a 148 byte Key.data file.

app/Sparrow.apk contains libscope-auklet.so which has quite some interesting strings embedded.

offset=3225419776,sizelimit=28494004224: => userdata partition. This is where on an Android system all user generated data is stored. logs/tools_log contains there some interesting logfiles.

[Fungus]

It is very interesting with which oscilloscope Siglent will answer this challenge. And when.

I don't think Siglent is working on any ASICs, and without them? Not gonna happen at this price point.

[Fungus]

Dave measured the power consumption and it is fairly high. Also there is likely almost no variation in power consumption so a fixed speed fan makes a lot of sense as there will be no advantage to have a temperature controlled fan.

Of course there is. In winter, or if I have the aircon on? Maybe no need for a fan!

[Peteg]

Yes will be interesting to see what Siglent does.

[mwb1100]

Here are the partitions in the SD card image, RigolDHO800-SDcard-dump.img:

...

If I might ask: what tools/techniques did you use to figure these offsets?[/list]

[EEVblog]

It is very interesting with which oscilloscope Siglent will answer this challenge. And when.

I don't think Siglent is working on any ASICs, and without them? Not gonna happen at this price point.

I tend to agree.

[EEVblog]

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.

There is 3.7mm from the top fins of the heatsink to the rear case.
Fins are 7.4mm deep.

[ataradov]

If I might ask: what tools/techniques did you use to figure these offsets?

binwalk tool previously mentioned/used in this thread.

[mwb1100]

Thanks.  I ran binwalk but m not familiar with the tool (and I'll admit, I was too lazy at the time to read docs or other info).  I found the output confusing, for example several volumes identified as "system" for example.

I guess it's just a matter of buckling down and reading up on it.

[ataradov]

Those extra volumes are just junk FSes with the same label. The way to find out is to mount each of them and see what is in them.