Siglent SDS2000X HD, 12bit, 2GSa/s high definition oscilloscope serie.

[hhappy1]

This is the fft screen of sds2000x plus.

1mv, 50ohm.
20M Bandwith, 10bit mode, 2M Memory, 1M fft point.
center:500khz, span:1Mhz, 100MSa/s, avr:16.

Spur noise of 90khz, 280khz, 500khz and 920khz occurs on all channels.
(Adjusting LCD Brightness will change the noise.)

I wonder what sds2000x hd is like when it's the same setting.

[noisyee]

It looks like a synchronizing sampling issue to me. That probably means you have a signal frequency very close to the instrument's sampling frequency / reference clock frequency / internal processing frequency or their harmonics.
Magic may happens if they are synchronized , or some periodic behavior if not perfectly synchronized, e.g. frequency wandering of two independent source. Used to use this trick for very small signal measurement.
I tried to replicate this behavior by fine tune the signal frequency and got somewhat similar result.
Attach1: Wobble appears periodically around raising edge

Attach2: Never appear again by just changing 1Hz

Attach3&4: Same trick to break down the waveform

[miro123]

It looks like a synchronizing sampling issue to me. That probably means you have a signal frequency very close to the instrument's sampling frequency / reference clock frequency / internal processing frequency or their harmonics.
Magic may happens if they are synchronized , or some periodic behavior if not perfectly synchronized, e.g. frequency wandering of two independent source. Used to use this trick for very small signal measurement.
I tried to replicate this behavior by fine tune the signal frequency and got somewhat similar result.

It is combination of
1. Very close frequecies/ harmonics
2. High frequence component -- mostly caused by wrong signal feeding/ impedance mismatch
3. Low sample rate - In your case the signal ringing is in GHz range

On positive side - You have to be happy that your scope timebase is so close to Bodnar reference.

[Martin72]

@noisyee:

Pic 3&4 are 100Mhz ??

[noisyee]

It looks like a synchronizing sampling issue to me. That probably means you have a signal frequency very close to the instrument's sampling frequency / reference clock frequency / internal processing frequency or their harmonics.
Magic may happens if they are synchronized , or some periodic behavior if not perfectly synchronized, e.g. frequency wandering of two independent source. Used to use this trick for very small signal measurement.
I tried to replicate this behavior by fine tune the signal frequency and got somewhat similar result.

It is combination of
1. Very close frequecies/ harmonics
2. High frequence component -- mostly caused by wrong signal feeding/ impedance mismatch
3. Low sample rate - In your case the signal ringing is in GHz range

On positive side - You have to be happy that your scope timebase is so close to Bodnar reference.

That's right. I think the wobble comes from some kind of aliasing issue interacting with the synchronizing sampling. The raising edge of signal is quite fast (hundreds of ps) for a 2GS/s system.

@noisyee:

Pic 3&4 are 100Mhz ??

Yes, 100MHz, almost the same sine wave with 2Hz difference if I remember correct.
This issue can be observed at dot display mode in fast time base. Seems an issue only when the number of actual sampling points is lower than the number of displayed points.

[rf-loop]

This is the fft screen of sds2000x plus.

1mv, 50ohm.
20M Bandwith, 10bit mode, 2M Memory, 1M fft point.
center:500khz, span:1Mhz, 100MSa/s, avr:16.

Spur noise of 90khz, 280khz, 500khz and 920khz occurs on all channels.
(Adjusting LCD Brightness will change the noise.)

I wonder what sds2000x hd is like when it's the same setting.

Here SDS2kXHD

Example with this setting max level is around -51dBV (4mVpk)
Imho,  spurious (and noise*) free range is still "not bad".  *)Except when we go to lowest end of freq band where noise level rise quite lot.

Note different vertical positions, adjusted for better visibility.

 

[Martin72]

Can test it at work with lecroy WR9054(500Mhz, 8Bit) and HDO6054A (350Mhz, 12Bit).

Done, nearly on all 350...500Mhz scopes.
No scope showed this behaviour, except my HD.

[Martin72]

Auto Mode, tried several things without effect.
Waveform seems to change between dots/vectors periodically and "paused" inbetween.
I´m pretty sure the Xplus fully armed to 500Mhz didn´t show this behaviour.
Will take the bodnar pulser to work, test it on the HDO 350Mhz and the Waverunner 500Mhz.

Will check on the 6204A today too but it's currently packed in the SUV ready for a visit to a client so will pack the Leo pulser too and grab a screenshot or at least try and find if it does anything similar. 
TTYL

And did you ?

[tautech]

Auto Mode, tried several things without effect.
Waveform seems to change between dots/vectors periodically and "paused" inbetween.
I´m pretty sure the Xplus fully armed to 500Mhz didn´t show this behaviour.
Will take the bodnar pulser to work, test it on the HDO 350Mhz and the Waverunner 500Mhz.

Will check on the 6204A today too but it's currently packed in the SUV ready for a visit to a client so will pack the Leo pulser too and grab a screenshot or at least try and find if it does anything similar. 
TTYL

And did you ?

Yup, a quick look. Same effect not seen and didn't think I needed report such as SDS6000A has 2.5x sampling rate of a 2k series scope so no chance of seeing this effect. Nice clean and sharp Gibbs ears only. 

[Martin72]

Same effect not seen and didn't think I needed report such as SDS6000A has 2.5x sampling rate of a 2k series scope so no chance of seeing this effect.

On my former SDS2504X+ I didn´t got the effect the HD shows...
Hope this not something I must live with.

[tautech]

Same effect not seen and didn't think I needed report such as SDS6000A has 2.5x sampling rate of a 2k series scope so no chance of seeing this effect.

On my former SDS2504X+ I didn´t got the effect the HD shows...
Hope this not something I must live with.

You've only seen this once, right ?
As others have suggested it's a frequency/sampling artifact and now you know about it I doubt your might even strike it again. I won't be holding my breath until you see it again. 

[Martin72]

Soon as my HD arrived I took the bodnar pulser on it and wonder about the "stop and go" behaviour of the waveform displaying.
Never saw this before, thought this could be gone with a new firmware but it wasn´t so I post it.

.....

[rf-loop]

It´s not that easy..

https://youtu.be/69YJ2UcYTgM

Please explain _excactly_ every single thing what happen and have changed between time position 0:10 - 0.11s.  There start new camera shot. Why not run camera continuously without break. There can see that situation totally changes between these shots before 0:11 and after it. Before this point we can nicely see just normal things done by Sinc. After this 0.11s  point this effect looks like disappear and looks like you are in other mode.
Please do this same using (eta: also) normal DSO mode instead of (eta: only) DPO mode.
Run both modes using two ways, with Sinc on and display mode lines and  then display mode dots (Sinc is automatically off in dots mode).
And show both results with full explanation about All settings. (at least I'm not a so-called clairvoyant)

[Martin72]

at least I'm not a so-called clairvoyant)

Not ? I'm disappointed...

Its an old video from the first time, can make a new at the weekend.

[Martin72]

Videos from today, 5ns timebase, auto memory, normal acquire...

https://youtu.be/6CxySRctuMw

Dot mode

https://youtu.be/LUHSZjYWaiw

Vector mode

https://youtu.be/J1Wv7c-rCzg

Several timebases, dot mode

Everytime reproducable on every channel.
And it´s not periodically, the pauses inbetween, where the signal looks normal, are changing.

[Performa01]

Have you ever watched an old western movie, where the spoke wheels of a coach appeared to stand still or even rotate backwards at times? We get this stroboscopic effect when the "spokes per second" come close to the framerate of the recording.

It has already been hinted several times by other posters that it is the same here. The internal 10 MHz reference of your specific SDS2000X HD is within <1 Hz (<0.1 ppm) of the signal source (the pulser). In this particular case, the frequeny difference seems to be just 0.1 Hz!

Of course such a pair is rare and you won't see this with any other combination of pulser and scope within your reach. A first hint is the 10.00000 MHz trigger frequency display, which indicates a frequency difference of no more than +/-5 Hz.

What the videos actually show is the beat frequency of maybe 0.1 Hz (10 ppb!), i.e. the phase difference between signal and sampling progresses from 0 to 360° within about ten seconds. Whenever the phase difference is zero, samples are not distributed anymore and you see the individual dots, that would otherwise form a contiguous line.

In vector mode, you always get a line, but of course it looks slightly different depending on the phase and its change rate of the source data.

The fact that this phenomenon is quite constant over several days hints on the high stability of both the pulser and the SDS2000X HD clock source. Of course there will be some (also short-term) variation over time - we cannot expect a frequency difference of just 10 ppb to remain rock stable, yet it's an amazing effect.

[rf-loop]

Videos from today, 5ns timebase, auto memory, normal acquire...

https://youtu.be/6CxySRctuMw

Dot mode

https://youtu.be/LUHSZjYWaiw

Vector mode

https://youtu.be/J1Wv7c-rCzg

Several timebases, dot mode

Everytime reproducable on every channel.
And it´s not periodically, the pauses inbetween, where the signal looks normal, are changing.

ETA: Performa01 was faster    when I was editing this msg.

After carefully watching these videos, it just seems like a normal phenomenon.

If we can perfectly synchronize the clock of this generator and the oscilloscope, we can make stationary points appear.
By chance, the clocks on your Lbodnar generator and oscilloscope are very close to each other, but the variation is due to imperfections in the frequency stability of the clocks.

This phenomenon can be achieved also with many signal generators.
I just tried it with SDG1032X generator and SDS2504X HD oscilloscope. Exactly the same phenomenon, but the points move more wildly because the SDG1032X reference is not good for this purpose. If locked with an external reference, it still has a lot of frequency jitter (phase noise). Image is just gif made using some random screenshots. But this kind of things there exists. But looking from real scope screen naturally bit different.

Another thing is when the screen is in vector mode. The corner wobble of the square wave is caused by Sinc because the sample resolution is too small for that. Aren't we all familiar with this phenomenon? When you take into account that the triggering happens randomly in different places in the sample interval, even there in the corners, the position of the points varies and when we are close to sync we start to see sometimes slowly changing Sinc products just somehow same nature as Gibbs ears... This phenomenon cannot be brought out with, for example, the SDG1032X, because simply its rise time is completely insufficient - the harmonics run out in the middle to bring out this phenomenon, which can be seen nicely in one of the videos.


Just with SDG1032X  (naturally I have adjusted SDG frequency for get out as close sync with scope clock as possible with easy steps, My SDG freq just when take this image was 10.000048MHz)

[Martin72]

Thx to rf-loop, performa01, 2N3055 and all I´ve not mentioned now..
Everything´s now clearer to me and I´m glad it´s not a fault or something else.
I´ve tried the same with my sdg1062X and indeed it is as it is:

https://youtu.be/A5BSGzmJ2tE

[PartialDischarge]

New user of the scope. Bought it for the accuracy in the measurements and have been testing it with a HP 3245A and I'm very happy with it. Good build quality and overall features.

[Martin72]

Congrats, we´re growing more and more...
Hopefully I got more time in the next weeks, want to test more/making more measures.
Like decoding(with the siglent demoboard), ripple of a linear psu(which I couldn't measure with the mso5000), deskewing current/voltage with the deskew fixture, testing the new filter function and so on...
Btw filters:
I would test them in a range that are interesting me - but I got a feeling, this won´t work...
Example PWM 10.8khz, extract a 400Hz sinewave from it.
You doing this with a lowpass of about 1khz cutoff - I´m curious if I can simulate it on the scope with it´s LP filters in such manner, that the sinewave only will be displayed in the math function.
Will find it out soon.

[Martin72]

I could swear that there were more posts here after my last...
Nevertheless, today got a short time to play.
First creating a pwm with 400Hz sinewave - Not so easy on the SDG1062X as the menu is somekind of confusing.
But I got it at last and then feeding the scope with it.
Now math, filters, lowpass....And as expected you can´t do it this way.
Don´t know the timebase anymore, something in the higher µs, minimum corner frequency you can adjust is 20Mhz...
When you "allow" to control the samplerate, minimum cornerfrequency drops down to 1Mhz, I need 1Khz in this case...
Try it tomorrow again with much lower timebase in ms range but didn´t expect much from it.
Must learn more about it, but actual I think the filters are not suitable for my purposes in the lower frequencyranges.
In the deleted post (why deleted..) it was mentioned to use the track function.
Two things about it:
To use these functions you must choose advanced measurements, which is annoying because of the permanent displaying table.
Then I could only choose trend(later I know why, RTFM...), took rms for it but a straight line was the result(although the signal of the pwm changes its width)plus the time of the scale is in seconds and you can´t change this.

[MathWizard]

As much as I'd love 1 of these, I could get a car for that kind of money.

I'd like to see a comparison video of the sds2104xplus vs a similar Rigol. I like my sds1104 a lot, but I'd love a bigger screen and a few higher spec's, but it's a toss up of options.

Either way I will upgrade this winter, I can't wait. That will be a gift to myself.

[tautech]

I'd like to see a comparison video of the sds2104xplus vs a similar Rigol.

Video by Howardlong linked in the 1st post here is worth a watch:
https://www.eevblog.com/forum/testgear/siglent-sds2000x-plus-coming/

[tv84]

I updated an old list of Siglent's Product Types here.  (probably not the best place...)

But, definitely, the "NeZha" is a new machine they are working on. Anyone has any clue what is this (besides the NDA boys  )?

Looks like something with (whatever this means):

hw_adc_bits = 12
data_adc_bits = 16

[Martin72]

Aha...