Rigol DHO804 Test and Compare Thread

[Serg65536]

P.S. Even those 70µV you see are generated by that being non-shielded measurement, all kinds of unwanted thermocouples and multimeter zero offset

If on DHO800 you set bias to 8 V on 100 or 200 mV/div vertical, you get 60 mV output voltage, not uV!

[ebastler]

If on DHO800 you set bias to 8 V on 100 or 200 mV/div vertical, you get 60 mV output voltage, not uV!

I have repeatedly been impressed by your findings on the DHO 800 -- great work! But may I suggest that you tone down the choice of fonts in your posts. The bold+red create a dramatic "red alert!" vibe which seems unhelpful for a productive discussion.

Using bold typeface without the red color should be enough to direct attention to very important messages, and italics or underlining work well for less pronounced emphasis.

[Fungus]

You don't seem to get it.
Vertical zoom is a SW special feature and little to do with vertical offset limits.

You're saying a software feature can cancel out 10 screen divisions of offset?

(or whatever it is they were doing earlier in the thread)

Edit: Well, yes, it can, but then you'd complain about it not being 12 bits and more and besides, Mr Red Ink was specifically talking about the thing below:

[2N3055]

the scope can generate output voltage at its input!!

Ummm... yes. It wouldn't be able to calibrate itself without that feature.

What are you talking about?

Calm down and count to 10 before posting.

 I'm referring to the "offset" part of the "offset+gain amplifiers" that ALL oscilloscopes have.

(Hopefully Mr. Red Ink is referring to the same thing)

It is a software thing. No BOM price attached.

I'm fairly sure offset amplifiers are hardware.

Calm down and count to 10 check your information before posting.

Serg65536 reported that when he adjusts vertical offset (yes, that hardware one you are talking about) that results in output voltage appearing on input BNC.  There is absolutely nothing right about that, that is defect if true. That is either that he measured wrong (in which case it is false alarm) or his scope is damaged or all of the scopes have that defect. Thing to do here is for other users try to replicate same thing.

Zoom he is talking about is real proper vertical/horizontal zoom mode. I keep repeating that all the idiotic insistence of use of  word "zoom"  for changing the timebase or vertical attenuation/offset in channels is confusing as hell but nooo... I'm an word nazi...

To explain (I will use Siglent as example because that is what I have but other scopes have that too):  When I enter zoom mode, i can choose both vertical and horizontal magnification from captured data.
See image zoom1 where you can see a small box in upper overview. That part is magnified in bottom zoom window.
In image zoom2 you can see additional horizontal zoom (small box is smaller this time).
These two views are from the same physical capture, and works with stopped and live data...
There is no change in acquired time-base or physical vertical attenuation or voltage offset on channel.
It is post processing.
That is zoom.

So we don't have any more misunderstanding.

[ebastler]

[...] So we don't have any more misunderstanding.

Thanks for the clarifying post. I was about to write something along those lines, but could not have provided the helpful illustrations.

I think the wording "Rigol DHO800 and DHO900 don't have a PROPER ZOOM function" from Serg65536's post above may have pushed this discussion in the wrong direction. The software functionality (user interface) is there, looking very similar to the Siglent screenshot you shared. The actual limitation is that Rigol do not allow large offsets in Y, right? Which Serg found and reported as a bug a while ago, https://www.eevblog.com/forum/testgear/rigol-dho800900-oscilloscope-bug-reports-firmware/msg5093817/#msg5093817.

EDIT: Not sure why I can't get that link to point to the relevant post. It's reply #31 in that thread.

[Fungus]

Serg65536 reported that when he adjusts vertical offset (yes, that hardware one you are talking about) that results in output voltage appearing on input BNC.  There is absolutely nothing right about that, that is defect if true. That is either that he measured wrong (in which case it is false alarm) or his scope is damaged or all of the scopes have that defect. Thing to do here is for other users try to replicate same thing.

Have you tried it yourself?

I just set mine to output 1kV and tried the tongue test. I didn't feel any tingle, so.... 

[csuhi17]

Can you set it so that you measure 0mV on the channel?
Did you try based on the description in the user manual? 
Has there been a change?

[ebastler]

Serg65536 reported that when he adjusts vertical offset (yes, that hardware one you are talking about) that results in output voltage appearing on input BNC. 

Actually, Serg65536 reported a problem when he sets a large bias, not offset. Those seem to be different settings, and I just realize that I am not clear what the "bias" is and does. Maybe someone can confirm or correct my understanding:

The offset (position) knob on the front panel serves two different purposes, depending on the operating mode of the scope. During live acquisition is controls the physical offset input to the front end amplifier; during Stop mode it lets the user zoom in or out of the captured data via pure software functionality. So far, so good.

I always thought that "bias adjustment" uses a combination of hardware and software to compensate for small offset errors in the amplifier: Physically apply a small offset voltage, but call that "zero" in software, such that 0V signal input are actually reported as zero.

If my concept of "bias" is correct, that raises two questions for me: (a) I thought that this is adjusted automatically by the scope during self-calibration. Is it typical to have that as a user control? (b) Why would the "bias" mechanism cause an output voltage on the inputs, while adjusting the offset does not? (Or does it?)

Thanks for your thoughts or corrections of the above!

[Fungus]

Actually, Serg65536 reported a problem when he sets a large bias, not offset. Those seem to be different settings, and I just realize that I am not clear what the "bias" is and does. Maybe someone can confirm or correct my understanding:

I think it's this:

If my concept of "bias" is correct, that raises two questions for me: (a) I thought that this is adjusted automatically by the scope during self-calibration.

Yes.

These Rigols also seem to have an extra manual adjustment for some reason.

(b) Why would the "bias" mechanism cause an output voltage on the inputs

It wouldn't.

(If it did then calibrating the 'scope would also produce a voltage there)

[ebastler]

(b) Why would the "bias" mechanism cause an output voltage on the inputs

It wouldn't.
(If it did then calibrating the 'scope would also produce a voltage there)

Well, Serg65536 has apparently measured an output voltage -- for bias voltages which far exceed what would be required to compensate amplifier offset errors. Could you try and repeat that experiment on your scope?

BTW using bias at 100 and 200 mV/div gives OUTPUT VOLTAGE from -40 mV (for -8 V bias) to 60 mV (for 8 V bias). Yes, the scope can generate output voltage at its input!!
For the 500 mV, 1, 2 V/div ranges output voltage range is -1,7...3.5 mV.
So don't try to move you waveform with the "bias" field!!

I would not be too concerned, since I don't see any reason to input large "bias" settings. And Serg' has come to the same conclusion; just don't "abuse" the bias setting to enter large offsets. Still I am wondering why Rigol is exposing the "bias" control at all?

[Fungus]

(b) Why would the "bias" mechanism cause an output voltage on the inputs

It wouldn't.

Well, Serg65536 has apparently measured an output voltage

His 'scope must be broken. I can't see anything on mine.

I would not be too concerned, since I don't see any reason to input large "bias" settings.

I just set mine to 1.8 Megavolts and there was no arc flash. I think we're safe.


(Sorry to disappoint those who wanted to use their 'scope as a bench power supply)

[2N3055]

Serg65536 reported that when he adjusts vertical offset (yes, that hardware one you are talking about) that results in output voltage appearing on input BNC.  There is absolutely nothing right about that, that is defect if true. That is either that he measured wrong (in which case it is false alarm) or his scope is damaged or all of the scopes have that defect. Thing to do here is for other users try to replicate same thing.

Have you tried it yourself?

I just set mine to output 1kV and tried the tongue test. I didn't feel any tingle, so.... 

Are you deliberately trying to provoke me, are you rushing to answer before thinking, or have trouble comprehending..??

He set "Scale" to 100 or 200mv/div and Bias to 8V.

Reproducing problem by definition means reproducing it.
You cannot just choose random settings and say "Ahaa!! See, no problem!"
You need to use exactly same settings to see if you can reproduce result.
Changing any variable is not valid.

If preamp injects voltage/current into input, if input range is set to such setting that preamp is not used will remove it by definition.
Also if input attenuator is used it will load/divide that said voltage too.

Any voltage seen on input will be maximized when front end is in a config where influence from input attenuation is minimal and preamp is used.

[ebastler]

I just set mine to 1.8 Megavolts and there was no arc flash. I think we're safe.

Please note that Serg65536 found the largest voltages on the input (10s of mV) in the lower amplifier ranges (200 mV/div). A multimeter would be really helpful to repeat his test, rather than checking for tongue tingles or arc flashes. 

[trp806mo]

May someone check the speed of the (statistical) measurement by fitting the scope with a 1Mhz waveforme and measuring the time to get 10K samples of width at 5us and 5ms of timescale ?

[ebastler]

PS: Yes, I twisted the knobs all over the place. Didn't see a single microvolt.

Could you please confirm that your twisting of knobs also included a setting of 200 mV/div and +8V bias voltage? What voltage did your multimeter show on the scope input jack in that state?

[TurboTom]

Yes it's there. I measure approx. +60mV at 100 or 200mV/div (either of these two ranges "works"), 8V bias and approx. -45mV at -8V bias. The effect is highly non-linear, a bias of +4V results in ca. +12mV and -4V in just between one and two millivolts negative. But the ugly thing is: The effect is just the same if you use offset instead of bias! So they are basically just the same function and add up internally, bias probably meant to compensate for a few millivolts of thermal drift so offset could be kept at zero for a centered trace.

[Fungus]

Could you please confirm that your twisting of knobs also included a setting of 200 mV/div and +8V bias voltage? What voltage did your multimeter show on the scope input jack in that state?

I'll do it after 2N3055 explains to us how measuring a voltage across something with a 1MOhm impedance has any significance.

FWIW I measured the current across the connector with those settings and saw 80nA.

Does that answer the question?

[Fungus]

I just measured the output of my Micsig and it can supply up to 6.5 millivolts with the right settings.

[ptluis]

See image zoom1 where you can see a small box in upper overview. That part is magnified in bottom zoom window.
In image zoom2 you can see additional horizontal zoom (small box is smaller this time).
These two views are from the same physical capture, and works with stopped and live data...
There is no change in acquired time-base or physical vertical attenuation or voltage offset on channel.
It is post processing.
That is zoom.

So we don't have any more misunderstanding.

What's your Siglent scope model? That zoom feature looks really nice 👍

[ebastler]

What's your Siglent scope model? That zoom feature looks really nice 👍

The DHO800 zoom function looks quite similar to me? See the attached screenshot by Martin72, https://www.eevblog.com/forum/testgear/rigol-dho804-test-and-compare-thread/msg5115042/#msg5115042.

[2N3055]

Could you please confirm that your twisting of knobs also included a setting of 200 mV/div and +8V bias voltage? What voltage did your multimeter show on the scope input jack in that state?

I'll do it after 2N3055 explains to us how measuring a voltage across something with a 1MOhm impedance has any significance.

FWIW I measured the current across the connector with those settings and saw 80nA.

Does that answer the question?

No it does not.

And 80nA will fully open MOSFET or FET if enough voltage is available and create all kinds of problems.  Fact that you don't work on high impedance circuits does not make it non problem for everybody...

Tom already confirmed that DHO900 behave the same ...

And to answer your other post, yes I am eager to find the truth..
As a service to my fellow Rigol users.. I don't care if DHO800 has defects because of me.
I'm not in a market for one, I would be more interested in DHO4000 series as a type of product.

Inexpensive does not mean cheap nonfunctional shit. DHO800 is limited enough in capabilities that it's price is very good only if it works flawlessly.
Who's side are you on? On your side and all other users and demand best product for your money, or you work for Rigol hiding stuff?
I actually think Rigol is decent company and all problems are there because of rushing to the market. Given the chance they will do you right and fix the stuff. But nobody ever fixes problems they are not aware of...

But hey, it might just one more of those imperfections that are simply showing it is entry level instrument and what is its price range.
Maybe I expect too much.

But this shows that sometimes you need to pay more to get more...
On SDS2000X HD I can have 80V offset on 200mV/div range. And it shows full 140µV on input in that case... yes 140 microvolts. That is worst case scenario. On average it is 20-30µV.

[2N3055]

See image zoom1 where you can see a small box in upper overview. That part is magnified in bottom zoom window.
In image zoom2 you can see additional horizontal zoom (small box is smaller this time).
These two views are from the same physical capture, and works with stopped and live data...
There is no change in acquired time-base or physical vertical attenuation or voltage offset on channel.
It is post processing.
That is zoom.

So we don't have any more misunderstanding.

What's your Siglent scope model? That zoom feature looks really nice 👍

That is SDS2000X HD. But it is same on 6000A too. And incoming SDS1000X HD will be exactly the same.

[2N3055]

What's your Siglent scope model? That zoom feature looks really nice 👍

The DHO800 zoom function looks quite similar to me? See the attached screenshot by Martin72, https://www.eevblog.com/forum/testgear/rigol-dho804-test-and-compare-thread/msg5115042/#msg5115042.

If you look more closely, that is only horizontal zoom.
On my images you can see I selected a small square, both horizontally and vertically..

[ebastler]

If you look more closely, that is only horizontal zoom.
On my images you can see I selected a small square, both horizontally and vertically..

Ah -- I thought that you can also zoom in vertically in the lower part of the DHO screen, e.g. by a pinch/zoom finger gesture. Not sure whether that is actually supported, though; I don't see it mentioned in the manual.

[Fungus]

And 80nA will fully open MOSFET or FET if enough voltage is available

Would 60mV be enough?

The Fluke 87V outputs over 7V in continuity mode. You'd imagine it would destroy half the circuits in its path...right?