Measuring Distortions with the Scope:What you see is not what you really have..

[Martin72]

For the SDS2000X+ I can say, they did.
Currently the manual status is issue "D" from may 2022.

[tautech]

Btw, it is good question why it is so many times seen that some publications do not have any date and/or version with history (changelog).

Agreed.

Which invites a comment on Siglent's own documentation -- I look forward to the day when  Siglent steps up to the next level of documentation standards for their products, not only including change logs but also having an index at the back, displaying chapter numbers in the footer, updating the manuals to keep them consistent with firmware changes, etc. 

Maybe 2023 will be the year for that. I can't wait.

Much is already available about document release dates on the HQ website:
See the dates for each doc in this list:
https://int.siglent.com/products-document/sds2000x-e/#navs

[rf-loop]

Btw, it is good question why it is so many times seen that some publications do not have any date and/or version with history (changelog).

Agreed.

Which invites a comment on Siglent's own documentation -- I look forward to the day when  Siglent steps up to the next level of documentation standards for their products, not only including change logs but also having an index at the back, displaying chapter numbers in the footer, updating the manuals to keep them consistent with firmware changes, etc. 

Maybe 2023 will be the year for that. I can't wait.

Much is already available about document release dates on the HQ website:
See the dates for each doc in this list:
https://int.siglent.com/products-document/sds2000x-e/#navs

Yes of course some times Sherloc Holmes work and guessing is ok but...

But if we talk good documenting practices... Only right place for datecode is in document itself. Period. And it need be in every single published document. Period.
Siglent need understand this and do it. Period. It is mandatory.
And it need be started over 10 years ago but still better later than never. It need do asap and including whole product portfolio. There need be "house standard" for documentation and not so that different design groups do different way. Some better and some not.

We can see there has already been a slight improvement in this matter over the years. That's good, but not at all enough yet.
I don't think I'm the only one who has tried to move a mountain in this matter.
Lot of development is still needed.
In terms of documentation, Siglent is still not at the appropriate level. The hardware can be done quite well, but there are major flaws in creating documents.

And it's also pointless to claim (if some people want to do so) that someone else doesn't make good documents either and that's why they aren't needed - maybe no one even reads them (as is often seen unfortunately).
Such thinking would be a mistake. Let's hope it doesn't happen.
If that happens, you have to ask who you are trying to ski after. Usually the skier behind is not the winner.

 

[Hexley]

I don't think I'm the only one who has tried to move a mountain in this matter.
Lot of development is still needed.
In terms of documentation, Siglent is still not at the appropriate level. The hardware can be done quite well, but there are major flaws in creating documents.
 

To further illustrate the point about the standards of Siglent documentation:

Consider this post from May, 2020 https://www.eevblog.com/forum/testgear/siglent-sdg1000x-waveform-generators/msg3051348/#msg3051348 that describes issues in the SDG Series Programming Manual. That post gives details on how the description of one of the essential commands for the unit, the modulated wave command, MDWV, is simply incorrect.

Though that was posted more than 2 1/2 years ago, with full details, the latest version of the programming manual (version PG02-E05A) contains the same errors. Verbatim.     I found that disappointing.

Of course, the Programming Manual does have a version label printed on its title page, so there's that.

[Martin72]

Back to topic, comparison...

FFT of SDG1062X and 2122X, sinewave 1khz, appx 0dBm @50Ohm:

SDG2122X:



SDG1062X:

[2N3055]

Martin,

Try using Markers, On harmonics, and enable relative scale... Start FFT on horizontal from 500Hz to get rid of DC peak..
When using 0dBm or 0 dBV you have direct readout, but this way you can use it easily on any level..
Try varying AWG output level a bit. You will see there are sweet spots where AWG uses full DAC scale and output amplifier drive has low distortion. Play with input sensitivity of scope too... Like most SA, best distortion is not when driven full scale... There is a fine balance between a noise floor and amplifier spurs..
SDG2000 will generally have lower distortion, especially in sweet spots.

Best,

[Martin72]

Hi my friend,

Can I test tomorrow(vacation), although it looks like it's not too far from what the spec sheet states.



(OK, the 2nd harmonic is a little "high" (appx 0.09%))

[markone]

Hi my friend,

Can I test tomorrow(vacation), although it looks like it's not too far from what the spec sheet states.



(OK, the 2nd harmonic is a little "high" (appx 0.09%))

I would take with grain of salt averaged FFT spectrums, especially for THD and similar math, the wonderful dynamic range is only apparent, my trust ends at best at "SNR = 6.02N + 1.76dB" 

Anyhow, here is my SDG2042X @ 1KHz - 0dBm as seen by HDO1000 :

[rf-loop]

I would take with grain of salt averaged FFT spectrums, especially for THD and similar math, the wonderful dynamic range is only apparent, my trust ends at best at "SNR = 6.02N + 1.76dB" 

Anyhow, here is my SDG2042X @ 1KHz - 0dBm as seen by HDO1000 :

You can take what ever salt or sugar...
Btw, this averaging doesn't affect the signal levels, but it does affect the noise floor, so we can peek "underneath" the noise level...(without true need here)  but there we can ONLY see these signals that are in sync(!) what is sometimes very important (depending what we are doing). As long as this is true etc...... you know. Of course it works and this way we can communicate in a very simple way with radio signals that are much below the noise level, which cannot be detected at all by a normal "modern/high-end" radio.  And no bystander knows that communication is taking place there (until they find how to sync and how to open messages, but how they do it if they do not know what to try find and where.

https://www.analog.com/media/en/training-seminars/tutorials/MT-001.pdf


Btw, in Rigol image there is RBW: 90.90m
Where is specified what it really mean.
In your image can read that FFT samplerate is 1MSa/s  and you have there 20kHz span.
With 1MSa/s normal FFT full span is 500kHz and you have then selected 0-20kHz span. If there is 1Mpts and full span is 500kHz then Delta f is 0.5Hz. But what is this RBW "90.90m"?

What is FFT window used in this image.
DHO1000 user manual or data sheet do not tell anything but "resolution".
Also in DHO1000 (HDO1000) manual: "FFT resolution is the quotient of the sample rate and the number of FFT points. If the number of FFT points is a fixed value (65,535 at most),......." 

[markone]

Btw, in Rigol image there is RBW: 90.90m
Where is specified what it really mean.
In your image can read that FFT samplerate is 1MSa/s  and you have there 20kHz span.
With 1MSa/s normal FFT full span is 500kHz and you have then selected 0-20kHz span. If there is 1Mpts and full span is 500kHz then Delta f is 0.5Hz. But what is this RBW "90.90m"?

What is FFT window used in this image.
DHO1000 user manual or data sheet do not tell anything but "resolution".
Also in DHO1000 (HDO1000) manual: "FFT resolution is the quotient of the sample rate and the number of FFT points. If the number of FFT points is a fixed value (65,535 at most),......." 

I can confirm, among many other, there is a bug regarding acquisition parameters display during FFT operation.

Right yesterday, when i was doing the meas that i attached in my post, i found out that rotating horizontal scale FFT acquisition sample rate, sample interval and samples number  were not changing in accordance 

Anyhow, the produced spectrum graph is "aligned" with an equivalent meas made with my SSA3021X in sweep mode, so i'm confident about that image, RBW apart for obvious reasons.

[Performa01]

(OK, the 2nd harmonic is a little "high" (appx 0.09%))

So the third harmonic is about the same on both generators, but the 2nd is about 10 dB stronger on the SDG2000X, which is quite unexpected.

Do you happen to have some decent voltmeter that can measure down to 100 µV with sensible accuracy?

If so, please check the offset voltage on both generators and report the results here...

[Martin72]

So the third harmonic is about the same on both generators

It is 6dB lower on SDG1062X, the table (peaks) is a little bit irritating.
"Best" multimeter here is a calibrated brymen 869s...
Could try with it.
Martin

[Martin72]

Checked DC-offset, on both generators the same (appx 400µVdc (Meter shows 0.41mVdc), output 640mVpp@50Ohm terminator).

[Performa01]

So the third harmonic is about the same on both generators

It is 6dB lower on SDG1062X, the table (peaks) is a little bit irritating.

Oh yes ... maybe it would be more intuitive if you sort by frequency instead of amplitude?

[Performa01]

Checked DC-offset, on both generators the same (appx 400µVdc (Meter shows 0.41mVdc), output 640mVpp@50Ohm terminator).

I'm too lazy to do a detailed calculation, but a quick estimate tells me that in this case the 2nd harmonic would be worse than -60 dBc. In other words: this DC-offset measurement is just a streetnumber (no wonder: check the specifications and calculate the error margins from that). What do you get if you measure a short? Does it show precisely 0 µV then?

Anyway, my theory was that even harmonics, especially the second, would occor especially with unsymmetric waveforms as we could get them with non-zero DC offsets. You could still try to tweak the DC-offset and see how it affects the even harmonics.

[Martin72]

Hi,

What do you get if you measure a short? Does it show precisely 0 µV then?

It shows 000.00mV when shorted or the 50 Ohm termination is connected - Open the last digit change between -/+ 5.
I agree that could be only streetnumber.
If you set a dedicated offset voltage on the generator, it will also be measured accurately by the meter.
I could test how "low" I can get with this...
With the next test I´ll sort the table.

[mawyatt]

All this and the other somewhat related thread got me thinking about what's going on, so just a quick setup to check things out. Here's the related thread.

https://www.eevblog.com/forum/testgear/siglent-sdg1032x-sine-distortion-at-1-khz/25/

So, we got our Pico Scope 4262 (16 bit ADC) out and hooked up to a laptop. Then used the builtin AWG (only goes to 20KHz tho) and setup a 10KHz sine wave signal.

We setup the scope for a 0-100KHz span and captured the spectrum shown below as Pico_AWG.

Then connected the Pico AWG output to a SA as shown in PNG7.

Then connected to our SDS2000X+ as shown in PNG_147.

No attempt was made to "normalize" things and get all the displays setup up for a detailed comparison, just a quick look and see.

It appears that the SDS2000X+ is introducing the harmonic distortion from within, as the dramatic difference in the spectrums show.

Later we'll try and look at the SDG2000 & 6000 AWGs we have.

Best,

[mawyatt]

See above, here's the added files for the SDG2000X and the SDG6000X set at 10KHz.

Pico_AWG 2 is SDG2000X

Pico_AWG 3 is SDG6000X

PNG 8 is SA SDG2000X

PNG9 is SA SDG6000X

PNG 148 is DSO SDS2000X+ for SDG2000X

PNG 149 is DSO SDS2000X+ for SDG6000X

The AWGs seem much better than the SDS2000X+ can display regarding harmonic content.

Best,

[Martin72]

Hi....
Following one of Sinisas hints I´ve played with the vertical resolution between 100mV and 200mV/Div., same FFT settings as before.
Two pics, one with 100mV/div. (as the measure before) and one with 130mV/div.
Judge by yourself..

[electr_peter]

Likely explanation of SDS FFT issue described in other topic.

[Martin72]

One of the findings one could derive from this would be of a fatalistic nature, namely that one can forget the FFT function with oscilloscopes.
Is this really the case, or do you have to try a little harder to get a more credible result?
And if so, in which direction one would have to go for it....

[rf-loop]

Hi....
Following one of Sinisas hints I´ve played with the vertical resolution between 100mV and 200mV/Div., same FFT settings as before.
Two pics, one with 100mV/div. (as the measure before) and one with 130mV/div.
Judge by yourself..

You can also do same but take all averaging off and watch screen more time than just one eyes blink because we do not have good min-max feature in this FFT application, and not much else either. After enough watching...  take a piece of paper or nice cloth and make curtains to hide these very low level peaks and "problem" is "solved".     
:- // 
(= if the measurement instrument  can't show enough truth then instrument must not display it)


Btw, voltage band change between 100│102mV/div. (and 1.00│1.02V/div)
Using this have minimal other effects than just band bottom and band top, so minimal change in signal level in ADC but different front end noise figure.

[Performa01]

Btw, voltage band change between 100│102mV/div. (and 1.00│1.02V/div)
Using this have minimal other effects than just band bottom and band top, so minimal change in signal level in ADC but different front end noise figure.

... because of vastly different PGA gain, which in turn could also lead to a different distortion level of the PGA (which probably is the main source of nonlinearity in a scope frontend).

[rf-loop]

Btw, voltage band change between 100│102mV/div. (and 1.00│1.02V/div)
Using this have minimal other effects than just band bottom and band top, so minimal change in signal level in ADC but different front end noise figure.

... because of vastly different PGA gain, which in turn could also lead to a different distortion level of the PGA (which probably is the main source of nonlinearity in a scope frontend).

Yes, this can apparently suspect.

More research is needed.

[switchabl]

One of the findings one could derive from this would be of a fatalistic nature, namely that one can forget the FFT function with oscilloscopes.
Is this really the case, or do you have to try a little harder to get a more credible result?
And if so, in which direction one would have to go for it....

Not at all, in the same way that the frequency measurement function is not useless just because it does not have the same resolution and accuracy as a standalone frequency counter. An oscillscope is not a replacement for an audio analyzer (neither is a spectrum analyzer btw).

Every instrument creates some distortion of its own (not only in the ADC but more importantly the front-end as well). It will depend on the frequency and amplitude of the input signal and the vertical setting as well. And unfortunately, it is usually poorly characterised in oscilloscopes (the SDS2000X HD datasheet only says SFDR >= 45dBc which is not very helpful). Once you get close to those distortion limits, the error bars increase. The distortion of the generator and the receiver do not necessarily add in a straightforward way either. In some cases they may cancel partially and a noisier input signal may show less distortion because of dithering effects.

The SDG2000X has pretty clean output in the audio range (significantly better than the datasheet specs at most settings). Almost all oscilloscopes will struggle there, with some exceptions like the remarkable but specialized PicoScope 4262. Note that you are trying to characterize a 16 bit generator with a 12 bit oscillscope. Now that doesn't tell you anything definite about linearity at all but IMHO it should give you pause and make you proceed with careful consideration and healthy scepticism.

What can you do to improve your measurements? First, know your instrument and its limitations. You can try to measure an oscillator that is known to have distortion levels significantly better than your oscilloscope as a reference. That also allows you to find the settings that give you the best results. A quick sanity check is to add some in-line attenuation (on a spectrum analyzer, change the input attenuator setting). If the relative levels of the harmonics change, you will know that the distortion is from the receiver.

You can also add an external notch filter to attenuate the fundamental. This will enable you to measure down to very low distortion levels.