Rigol's New DHO800 Oscilloscope unbox & teardown

[2N3055]

Here's images with local lab bench lights off and on, judge for yourself.

I'm still not convinced your other 'scope is "brighter", just that it has more/bigger pixels illuminated on the trace.

(Yes, I'm aware that means there's more light coming out of it ... I'm just saying the Rigol's backlight isn't "dim" per se)

Arguing semantics is what lawyers do.

And you just explained yourself what the problem is.

Monitors are specified in nits, one candela per square meter (cd/m2).
Not per pixel.
Small screen and small pixels means less radiated optical power from single pixel. So if you make a small text, that text radiates with small amount of surface.. Makes it less luminous... It can have good contrast up close but it will be less visible from distance.

And it DOES seem to be really shiny and reflective, which does not help.

If screen is not good enough, people should complain to Rigol so they fix it.  They did it for MSO5000. If you keep quiet and trying to shut up other people complaining they'll keep suboptimal screen and nobody profits from it.

Maybe it is only some software setting.

[2N3055]

Also when booting, the large "RIGOL" shows up initially very bright and contrasty (nice!), then dims, then appears to brighten a little, then the screen appears and begins to dim again. The screen does seem to brighten some after warming up tho.

You may have a defective scope...

On my DHO804 the initial RIGOL logo starts bright, them dims and then brightens again.

With ADB you can change the brightness of the screen:

1>  adb connect scope-ip-address:55555                     [Connect via network]
2>  adb shell settings get system screen_brightness        [Returns 102.  Range is 0-255)
3>  adb shell settings put system screen_brightness 200    [This actually dims my screen, as explained below (*)]
4>  adb shell settings put system screen_brightness 255    [This brightens the screen back to Max]

(*)The shell script that starts the scope (.../start_rigol_app.sh) seems to override the android system value (set to 102) to 255 via line 149 in the script: 

echo 255 > /sys/devices/platform/backlight/backlight/brightness

I conjecture that the scope is bright at hardware/start-up, then the system dims it to 102, and then it gets reset by the Rigol start-up script to 255.

In my scope, omitting step 3> above, and setting it to 255 directly,
doesn't make the screen any brighter, consistent with my conjecture, in that:

(1) The start-up script seems to be overriding the system set value from 102 to 255.
(2) On my scope, the initial RIGOL splash screen starts bright, dims and returns to (and stays) bright, and the screen is as bright or brighter that any other instrument I have. (I also use a 22 inch hdmi monitor, which I have actually set dimmer than the scope screen.)

The value set in 3> or 4> above persists after a reboot. I have reset mine to 102, and still get the Max 255 brightness, as I have just described.

While you might be correct technically about sending ADB commands, if you read Mike's post, you will realize he didn't buy scope to play with it. It is potentially scope for a client that asked him to help evaluate if it is fit for purpose. If it doesn't work as it should coming from the factory it is not fit for purpose...Nobody is going to hack around the scope in this situation. If screen is not bright enough and there is no normal way for user to adjust brightness in GUI that means there is no solution at this moment until Rigol provides it...

[Fungus]

I just noticed a cool UI feature:

On the front of the 'scope there's two twisty knobs for moving things like cursors around.


In the setup dialogs there's two little indicators to show you that some values can be adjusted using the knobs, eg. Here I can change the Bias voltage using knob 1 and knob 2 changes the channel offset.


What I just noticed is that the indicators move around if you touch the input boxes. eg. If I touch "Scale" then "Skew" they change to this:


So now I can change Scale and Skew using the knobs. 

[2N3055]

I just noticed a cool UI feature:

On the front of the 'scope there's two twisty knobs for moving things like cursors around.


In the setup dialogs there's two little indicators to show you that some values can be adjusted using the knobs, eg. Here I can change the Bias voltage using knob 1 and knob 2 changes the channel offset.


What I just noticed is that the indicators move around if you touch the input boxes. eg. If I touch "Scale" then "Skew" they change to this:


 

Those little arrows are nice touch.
Does it have a physical Triggered LED?
You, know one that blinks when it triggers?
I miss that one on Keysight MSOX3000T.

I know it shows it on screen, but physical LED is better because you can see it with peripheral vision.
On Keysight, I put a LED on a piece of coax and connected it to trig out (aux) so I have a blinking LED.

I hate when they save on few LEDS that makes it so nicer to use...

[rpro]

While you might be correct technically about sending ADB commands, if you read Mike's post, you will realize he didn't buy scope to play with it.

If you read my post you will realize that my scope did not have a brightness problem out of the box.
"Also when booting, the large "RIGOL" shows up initially very bright and contrasty (nice!)" (as Mike puts it) is where my scope brightness ends up, by default, as it did out of the box. The adb commands and the rest of my post were used to explain why that may be the case.

[2N3055]

While you might be correct technically about sending ADB commands, if you read Mike's post, you will realize he didn't buy scope to play with it.

If you read my post you will realize that my scope did not have a brightness problem out of the box.
"Also when booting, the large "RIGOL" shows up initially very bright and contrasty (nice!)" (as Mike puts it) is where my scope brightness ends up, by default, as it did out of the box. The adb commands and the rest of my post were used to explain why that may be the case.

I honestly did not understand that from your post... But thank you for explaining.
And your explanation show you understand a great deal about how it works..
But as I said, there are many of us that won't thinker with it (lack of certain type of skills, time etc) but expect it work from the box...
This is one of the things that sure look (after your explanation) that Rigol could "fix" with correct config in software in one of the updates....

[mawyatt]

Another happy user (oscope in action - tube amp?)..

Nice video, note display here.

https://youtu.be/3OJnjXrAf5k?t=270

Best

[mawyatt]

Also when booting, the large "RIGOL" shows up initially very bright and contrasty (nice!), then dims, then appears to brighten a little, then the screen appears and begins to dim again. The screen does seem to brighten some after warming up tho.

You may have a defective scope...

On my DHO804 the initial RIGOL logo starts bright, them dims and then brightens again.

With ADB you can change the brightness of the screen:

1>  adb connect scope-ip-address:55555                     [Connect via network]
2>  adb shell settings get system screen_brightness        [Returns 102.  Range is 0-255)
3>  adb shell settings put system screen_brightness 200    [This actually dims my screen, as explained below (*)]
4>  adb shell settings put system screen_brightness 255    [This brightens the screen back to Max]

(*)The shell script that starts the scope (.../start_rigol_app.sh) seems to override the android system value (set to 102) to 255 via line 149 in the script: 

echo 255 > /sys/devices/platform/backlight/backlight/brightness

I conjecture that the scope is bright at hardware/start-up, then the system dims it to 102, and then it gets reset by the Rigol start-up script to 255.

In my scope, omitting step 3> above, and setting it to 255 directly,
doesn't make the screen any brighter, consistent with my conjecture, in that:

(1) The start-up script seems to be overriding the system set value from 102 to 255.
(2) On my scope, the initial RIGOL splash screen starts bright, dims and returns to (and stays) bright, and the screen is as bright or brighter that any other instrument I have. (I also use a 22 inch hdmi monitor, which I have actually set dimmer than the scope screen.)

The value set in 3> or 4> above persists after a reboot. I have reset mine to 102, and still get the Max 255 brightness, as I have just described.

Thanks!

Interesting the startup sequence, as we mentioned may have a defective display....not sure. Also note this is not for our use, but for a client, so altering the sequence or parameters not acceptable like 2N3055 mentioned.

Anyway, the scope is nice and we like it, but client may have issue wrt display brightness/reflections in their use environment.

Best,

[mawyatt]

Here's images with local lab bench lights off and on, judge for yourself.

I'm still not convinced your other 'scope is "brighter", just that it has more/bigger pixels illuminated on the trace.

(Yes, I'm aware that means there's more light coming out of it ... I'm just saying the Rigol's backlight isn't "dim" per se)

Arguing semantics is what lawyers do.

And you just explained yourself what the problem is.

Monitors are specified in nits, one candela per square meter (cd/m2).
Not per pixel.
Small screen and small pixels means less radiated optical power from single pixel. So if you make a small text, that text radiates with small amount of surface.. Makes it less luminous... It can have good contrast up close but it will be less visible from distance.

And it DOES seem to be really shiny and reflective, which does not help.

If screen is not good enough, people should complain to Rigol so they fix it.  They did it for MSO5000. If you keep quiet and trying to shut up other people complaining they'll keep suboptimal screen and nobody profits from it.

Maybe it is only some software setting.

Nice explanation as "nits" is related to one's visual experience (what eye "sees") and likely why it's used.

Another effect is when one takes a picture or video, the resultant is highly compressed/altered. Brights are suppressed and darks enhanced, thus bringing both closer within the limited image dynamic range. This is all part of the significant image processing that takes place in the camera because the core image CMOS sensor dynamic range isn't all that good (50~55dB). Even cameras like Nikon that claim to offer "Raw" image data, isn't actually right off the sensor, it's already been manipulated to make the result "appear" to have higher dynamic range. Having designed and worked with CCD sensors long ago, this is something we've been aware for awhile. Significant post image capture processing takes place to "make" the resultant image "appear" more natural, hiding the limited DR of the core CMOS sensor.

Here's an elementary lecture from Stanford wrt to such (see pages 8-13 and 14)

https://isl.stanford.edu/~abbas/ee392b/lect08.pdf

So in the context of the discussions going on here about display brightness, the images/videos shown will compress the variation, enhancing the dim and compressing the bright relative to what one's eye experiences.

Edit: One thing that just occurred to us while using the DHO814 relative to a larger screen DSO. This is very much like a handheld DMM vs bench DMM. Where for simple quick tasks, one "reaches" for the handheld, and more complex, precision and intensive tasks one "reaches" for the bench top DMM. In that respect one should have both types DMMs (suspect most already have such), and maybe both types DSOs 

Best,

[Fungus]

I honestly did not understand that from your post... But thank you for explaining.
And your explanation show you understand a great deal about how it works..
But as I said, there are many of us that won't thinker with it (lack of certain type of skills, time etc) but expect it work from the box...
This is one of the things that sure look (after your explanation) that Rigol could "fix" with correct config in software in one of the updates....

The display goes bright-dim-bright on bootup.

All the tinkering does is avoid the "dim" phase of that during bootup. It doesn't change the brightness of anything else.

(or whatever "brightness 255" is)

[Fungus]

Nice video, note display here.

https://youtu.be/3OJnjXrAf5k?t=270

It's something that goes with "touchscreen" territory. Micsigs are similar out of the box but Micsig includes a matte screen protector with every 'scope. It really helps.

I'm waiting on a couple of screen protectors to try with my Rigol.

[DaneLaw]

Don't you have a light/nits-meter, then you could get some tangible numbers that other people could replicate as I would guess many do have light-meters, as they are quite cheap and even the budget ones, perform quite decent.

Should be possible to dim the room, and create a lid scenario on the scope in the circumstances where you sense that its "dim", that other owners can replicate, and compare values -
To conclude if the unit's display is underperforming, as brightness/dim is relative, varies from one to the next..

Was unsure if it's a TFT or IPS in DHO8/900 series but from the looks of videos on the DHO800-900 series and the glow from certain angles, I would guess it's an LCD TFT display, which is quite common on scopes.

Checked the max brightness on Micsig-STO-scopes which peak at 414 on battery with just a white slide as light-sample, haven't tested fx a wide waveform to see what it musters in the scope app, but it's bright enough that it's fully useable outdoors, & indoors needs to be dimmed down quite a bit for my taste..
the scopes topglass screen also seems to be frosted like you see on fx mat or anti.glare screens. even though it's a glossy touchscreen.

- An example of the frosted native glassscreen & the LCD TFT RGB grid   - https://tinyurl.com/nw958xat - three first pic is the same and vice versa for the two last, only the focal depth varies, - that glass-frosting also makes it difficult to get the pixels into focus...

Tried briefly a few years back to put a screen-protector on the Micsigs screen, to see what amount it would counter FX glare, but it made it hazy from the TFT light glow underneath clashing, as I doubt it is laminated, so that was a clear no-go for my taste and peeled of again.. - it undermined some of the dark contrast that I personally prioritize a lot, as I prefer the waveform-background to be as close to inky black as possible.

Likely also why many scope vendors pursue TFT vs IPS, as IPS has a tendency to undermine the dark in the infamous IPS-hazy-glow (light bleed) that partly tones black into grey'ish...though, it is getting better for higher-end IPS panels.
An OS brightness slider hopefully will see daylight in future FW.

[DaneLaw]

Here it is next to my Micsig. The Micsig is at full brightness.
---

Are you using a screen protector on your Micsig.? was not a good experience on my unit, as mentioned above..

// it does look like you have some kind of screen protector on it, as your screen also has air bubbles. (we've all been there, they are a high-maintenance b to squeeze out)

Example with two Saelig Mcsig models that I guess are fitted with the incl. screen protector as theirs are also plotted in bubbles..https://tinyurl.com/2s3p3djy
 it creates this hazy greyish look that will counter glare'  but to me also undermines the contrast a lot, and the more I blasted the brightness the worse it got, did not like it
Was left with the impression that the angled-screen light on the non-laminated display to an extent conflicted with that added plastic layer..

[mawyatt]

I'm waiting on a couple of screen protectors to try with my Rigol.

Will that screen protector affect the touchscreen performance?

Best,

[Fungus]

it does look like you have some kind of screen protector on it, as your screen also has air bubbles. (we've all been there, they are a high-maintenance b to squeeze out)

I put that on when I was a screen protector noob. 

There's a couple of small glitches on the black bezel on the left but none in the screen area.  Micsig gave me two protectors so I've got another one somewhere. I've been meaning to put it on for years but never got around to it.

These days I can put on any size screen protector perfectly in about five minutes. 

I ordered some as soon as I measured the screen of the DHO. Aliexpress says they're in transit on a guaranteed 10-day delivery. Watch this space...

[Fungus]

FWIW: If you tell your DHO to go to the Android home screen (via ADB) all you get is the same black background with "Rigol" that you see during power on.

There's no icons or menus or anything like that.

It doesn't respond to any inputs, you have to power it off and restart.

[Martin72]

I wanted to get the data sheet for the display earlier, I need to know what it is called.
So I opened the scope and...no chance.
The display is glued to the front, like a cell phone.

[mwb1100]

FWIW: If you tell your DHO to go to the Android home screen (via ADB) all you get is the same black background with "Rigol" that you see during power on.

There's no icons or menus or anything like that.

It doesn't respond to any inputs, you have to power it off and restart.

I wonder if you can `adb install` some launcher apk to get a decent home screen?

[Fungus]

FWIW: If you tell your DHO to go to the Android home screen (via ADB) all you get is the same black background with "Rigol" that you see during power on.

There's no icons or menus or anything like that.

It doesn't respond to any inputs, you have to power it off and restart.

I wonder if you can `adb install` some launcher apk to get a decent home screen?

No idea... I'm not an Android guy, I was just fiddling around.

I pulled the main .apk out to have a poke around inside it. Can anybody recommend an apk decompiler.

Maybe this should be in another thread...

[Howardlong]

A DHO924S turned up here yesterday.

The UI is reasonably snappy, but I'm not keen on the implementation: you have no choice but to resort to using the touch screen.

I had to resort to RTFM to figure out how to set the probe ratio. Thankfully, the later firmware 1.01 is more obvious in this regard.

I haven't figured out how to set a one button user default rather than factory default, or if you can even do it, other than saving you favoured settings to storage and restoring them as necessary. As the probes are in 10X 99% of the time, and pressing default resets them to 1X. It also resets the LA channel threshold to 0V.

I headed straight over to doing serial decodes which is pretty much a staple in my line of work, but made the fatal error of trying to do it with the LA rather than the analogue channels.

Firstly, I note it doesn't support 10 bit I2C, which was a bummer because the demo board I was using uses that extensively (Tek MDO Demo 1 board).

Secondly, SPI decode doesn't seem to function at all on the LA.

The USB C port supports PD as a sink, but I can confirm that it isn't a USB host. Be aware that if you decide to use a multi port USB C adapter, many of these reset the ports when you plug or unplug another device, thus shutting down your scope.

It looks like the bandwidth is reduced in line with Nyquist when you start adding analogue channels beyond 2, determined by measuring the associated decrease in rise time.

When measured with an RF signal generator with a 50 ohm through attenuator, no attempt is made to filter the signal in line with Nyquist, so aliasing is a problem, including measurement.

Something also of note is that when you add the LA, the sample rate is halved again, so if you run 3 or four analogue channels and any number of LA channels. the maximum sample rate drops to 156.25MSa/s. Whether the trigger is on an analogue or digital channel didn't affect the maximum sample rate.

Code: [Select]

Analogue / Digital / Max Sample rate / Max mem depth/channel

0        / 1 to 16 / 625MSa/s        / 25Mpts
1        / 0       / 1.25GSa/s       / 50Mpts
1        / 1 to 16 / 625MSa/s        / 25Mpts
2        / 0       / 625MSa/s        / 25Mpts
2        / 1 to 16 / 312.5MSa/s      / 10Mpts
3 or 4   / 0       / 312.5MSa/s      / 10Mpts
3 or 4   / 1 to 16 / 156.25MSa/s     / 1Mpts



Had another go on it tonight for an hour or so, which confirmed what I'd surmised yesterday: the firmware in its current state is far to buggy for this to be relied on for my purposes, so it'll remain off the bench until the firmware matures. My use case for this was as an alternative to the Analogue Discovery for field work, but it's just too unfinished.

[ebastler]

Something also of note is that when you add the LA, the sample rate is halved again, so if you run 3 or four analogue channels and any number of LA channels. the maximum sample rate drops to 156.25MSa/s. Whether the trigger is on an analogue or digital channel didn't affect the maximum sample rate.

That's a bummer -- both, the sample rate restriction and falling back to the measly 1 Mpt memory. I can't find these limitations mentioned at all in the DHO 900 datasheet.

In a mixed-signal scope, I expect to be able to observe a "mix" of digital and analog inputs without incurring an extra penalty. Disappointing, and probably the finding so far which detracts me most from a DHO900. Is that a compromise made in other mixed-signal scopes too?

As a side note, we can probably give up on the hope that someone will come up with a hack to upgrade the DHO 800/900 to the 2 GSa/s sample rate of its larger sibling. Looks like the FPGA is already running at capacity and is limiting the total data rate?

[Warhawk]

Something also of note is that when you add the LA, the sample rate is halved again, so if you run 3 or four analogue channels and any number of LA channels. the maximum sample rate drops to 156.25MSa/s. Whether the trigger is on an analogue or digital channel didn't affect the maximum sample rate.

That's a bummer -- both, the sample rate restriction.....

I agree. Honestly, the scope is great value but after reading first user reports here I pass and go for a different option.

[Howardlong]

Something also of note is that when you add the LA, the sample rate is halved again, so if you run 3 or four analogue channels and any number of LA channels. the maximum sample rate drops to 156.25MSa/s. Whether the trigger is on an analogue or digital channel didn't affect the maximum sample rate.

That's a bummer -- both, the sample rate restriction and falling back to the measly 1 Mpt memory. I can't find these limitations mentioned at all in the DHO 900 datasheet.

In a mixed-signal scope, I expect to be able to observe a "mix" of digital and analog inputs without incurring an extra penalty. Disappointing, and probably the finding so far which detracts me most from a DHO900. Is that a compromise made in other mixed-signal scopes too?

As a comparison, the Rigol DS-1000Z LA goes up to 1GSa/s if you only had on 8 LA channels and no analogue channels.

On the other side of the coin, on the DS-1000Z there is no option to have 3+16 or 4+8 channels at all. IME. in practice, I find that this isn't much of a limitation as the use cases are rare.

I'm not sure if the 1Mpt limitation is an architectural or technical limitation, or just a front end feature.

The LA integration as a whole is so unfinished, I'd be willing to state the various metrics I have presented are subject to change in forthcoming firmware releases. For example, that triggering on an unviewed channel doesn't seem to affect sampling rate suggests to me that something's not quite right. At one point, I fruitlessly tried adjusting the position of the D0 LA channel on the screen, and suddenly it stopped displaying any signal... it turned out to have reappointed the signal on D2, and nothing short of a restart resolved it.

It's not the end of the world, but having a sample period on an LA that's not an integer multiple of 1ns might get a bit of getting used to. It reminds me of the egregious deviation from 1/2/5 timebase increments on the Tek MDO3000/4000 series in favour of 1/2/4, and not even offering a horizontal vernier.

[iMo]

..As a side note, we can probably give up on the hope that someone will come up with a hack to upgrade the DHO 800/900 to the 2 GSa/s sample rate of its larger sibling. Looks like the FPGA is already running at capacity and is limiting the total data rate?

It is a big question how they ported their DSO architecture from their previous models (like the 1000Z) into the new Zynq FPGA (2xarm9 mcus + Artix7 fpga in the XC7Z015). It could be they migrated that 1:1 without any special optimization or trickery. The Zynq model used is the lowest speed grade I would assume (none speed grade indicated on the package, that is interesting), which is 667MHz (the highest is 866MHz).

And perhaps ( ?? ) the Artix7 could be more performant as the Spartan6 used in the DS1000Z models..

[iMo]

Something also of note is that when you add the LA, the sample rate is halved again, so if you run 3 or four analogue channels and any number of LA channels. the maximum sample rate drops to 156.25MSa/s. Whether the trigger is on an analogue or digital channel didn't affect the maximum sample rate.

That's a bummer -- both, the sample rate restriction.....

I agree. Honestly, the scope is great value but after reading first user reports here I pass and go for a different option.

It cannot be a "great value" when you pass, though.. 

The prospective buyers (like you and me) should really wait a little bit till the first surge of "happiness" after the purchase of a new cute lovely gadget with VESA mount cools down (a buyer's psychology phenomenon) and people start to evaluate the stuff much more critically. I've seen only 1 statement in this forum the buyer indicated he will return the scope..

As I wrote before, this scope would be "a great value" when produced in 802/812/822 configuration only, at say $199/$249/$299, targeting explicitly the entry level segment (thus setting the expectation properly), moreover shipped without those most visible bugs (and I would perhaps tolerate the bugs knowing they will fix them soon). With adding another 2 channels and the LA and the AWG they simply overloaded the low cost tiny hw infrastructure of this scope, imho.

Even though it is a good value for the beginner, the people migrating from say their basic 2ch CRT scope to something more digital, and people requiring a 2nd, portable scope for occasional work battery powered, but - all the users here reviewing the scope are the advanced users - and it could easily be after a month or two many of them would consider this specific series a flop..