Siglent SDS800X HD 12 bit DSO's

[tautech]

Now more of you have this model series I have unlocked voting on the requirement of WiFi support for a further 4 weeks.
Thanks for your support.

[Muttley Snickers]

I'm still wondering how accurate the time keeping would be on these units once the clock is set?   

For example, say I needed to monitor for an intermittent communications fault (ie: data bus) at an industrial plant over a weekend and needed event timestamps but could not run a network cable for the NTP sync. I wonder if the time was manually set on the oscilloscope how well it could maintain its accuracy after say a 24 or 48 hour period in a stand alone application. 

   

[pdenisowski]

Your videos are excellent, I'm a fan of your work. The point I was getting at though, is that (probably) nobody is going to take a mortgage out to buy a scope if they don't know how to use it. 😉😉

Thanks!  And I agree, but the vast majority of our sales are to companies, not individuals. And most companies have employees at different skill / knowledge levels.

On a side note, it probably helps your video production quality that none of your gear competes in price range with the scopes in this thread (which are excellent in their own right).

Confession:  I’m basically a one-man production team. I do the researching, storyboarding, slides, text, illustrations, recording and editing all myself.  I do have a tremendous team of people who help me in many ways (and to whom I am perpetually grateful) but the actual composition and production are all done by one person - it’s honestly not that hard

And yeah, I’m really here on EEVblog as a hobbyist more than as an R&S employee.  There are so many people here who share their unique (and very valuable) knowledge from their own areas of expertise, and my small contribution is having some inside insight into the test and measurement industry.

[BillyO]

I'm still wondering how accurate the time keeping would be on these units once the clock is set?   

For example, say I needed to monitor for an intermittent communications fault (ie: data bus) at an industrial plant over a weekend and needed event timestamps but could not run a network cable for the NTP sync. I wonder if the time was manually set on the oscilloscope how well it could maintain its accuracy after say a 24 or 48 hour period in a stand alone application. 

 

I'd guess there are two possibilities.

1) The clock is run from the internal timebase which is accurate to 25ppm.  In that case given that there are 172,800 seconds in 48 hours it would be accurate to less than +/- 4.4 seconds.

2) The clock is run from a standard 32,768 Hz watch/clock crystal.  They've been making these for eons and they are at the heart of every quartz watch made today.  So possibly even better than the internal timebase and certainly not significantly worse.

What sort of accuracy do you need over 2 days?

[Electro Fan]

I'm still wondering how accurate the time keeping would be on these units once the clock is set?   

For example, say I needed to monitor for an intermittent communications fault (ie: data bus) at an industrial plant over a weekend and needed event timestamps but could not run a network cable for the NTP sync. I wonder if the time was manually set on the oscilloscope how well it could maintain its accuracy after say a 24 or 48 hour period in a stand alone application. 

 

I'd guess there are two possibilities.

1) The clock is run from the internal timebase which is accurate to 25ppm.  In that case given that there are 172,800 seconds in 48 hours it would be accurate to less than +/- 4.4 seconds.

2) The clock is run from a standard 32,768 Hz watch/clock crystal.  They've been making these for eons and they are at the heart of every quartz watch made today.  So possibly even better than the internal timebase and certainly not significantly worse.

What sort of accuracy do you need over 2 days?

Another opportunity to ask why modern oscilloscopes don’t come with or offer an option for a 10 MHz reference input….?

[ebastler]

Another opportunity to ask why modern oscilloscopes don’t come with or offer an option for a 10 MHz reference input….?

Rigol has it in the DHO1000 and 4000 (but not the small 800/900 series). I don't think Siglent has provided it in any of the X HD series scopes?

[KungFuJosh]

I've seen some scopes (don't recall the brands) that offer multiple references outputs, but never in less expensive models.

[hfleming]

I've seen some scopes (don't recall the brands) that offer multiple references outputs, but never in less expensive models.

With the capabilities of these new scopes it would really be nice if they added a ref. input. Until now it didn’t really make sense, but with the new measuring capabilities it becomes more needed.
BTW, I just love the fact that you can control the scope with VNC or Remmina, even though the screen-update is slighly more laggy than observing it on the scope itself, quite noticeable in XY-mode. Still another way to control the scope remotely.

[KungFuJosh]

I just love the fact that you can control the scope with VNC or Remmina, even though the screen-update is slighly more laggy than observing it on the scope itself, quite noticeable in XY-mode.

How does it compare on this model to the webserver? I found VNC more laggy than the webserver in XY mode on my SDS2kX+.

[2N3055]

I've seen some scopes (don't recall the brands) that offer multiple references outputs, but never in less expensive models.

With the capabilities of these new scopes it would really be nice if they added a ref. input. Until now it didn’t really make sense, but with the new measuring capabilities it becomes more needed.
BTW, I just love the fact that you can control the scope with VNC or Remmina, even though the screen-update is slighly more laggy than observing it on the scope itself, quite noticeable in XY-mode. Still another way to control the scope remotely.

Saying something would be nice without explaining why is not something we can learn from.

I can also make a point that I estimate that out of millions of scopes sold and used around the world very few in percentage are connected to external references even if they have the input.
For majority of use only parameter that is important is that scope has low internal clock jitter, and funny enough, not so much because scope timing measurements but because of performance of ADC.

What is your scenario where you think you would need scope disciplined to laboratory frequency standard, could you please elaborate?

[BillyO]

If you had the skills and the internal reference is 10MHz, putting in an external reference input would not be too horrendous.

But  yeah, why?  I can see that for a 10 digit frequency counter or a high precision signal generator, but a scope?  I have never felt I needed it.

[2N3055]

C'mon Paul,....

You are heavyweight player here, and we are glad for your presence, and thankful for shared knowledge.
And in addition to that, it also shows how we live in interesting times..   


Best
Siniša

[hfleming]

I've seen some scopes (don't recall the brands) that offer multiple references outputs, but never in less expensive models.

With the capabilities of these new scopes it would really be nice if they added a ref. input. Until now it didn’t really make sense, but with the new measuring capabilities it becomes more needed.
BTW, I just love the fact that you can control the scope with VNC or Remmina, even though the screen-update is slighly more laggy than observing it on the scope itself, quite noticeable in XY-mode. Still another way to control the scope remotely.

Saying something would be nice without explaining why is not something we can learn from.

I can also make a point that I estimate that out of millions of scopes sold and used around the world very few in percentage are connected to external references even if they have the input.
For majority of use only parameter that is important is that scope has low internal clock jitter, and funny enough, not so much because scope timing measurements but because of performance of ADC.

What is your scenario where you think you would need scope disciplined to laboratory frequency standard, could you please elaborate?

FFT with an accurate reference, for one.

[2N3055]

I've seen some scopes (don't recall the brands) that offer multiple references outputs, but never in less expensive models.

With the capabilities of these new scopes it would really be nice if they added a ref. input. Until now it didn’t really make sense, but with the new measuring capabilities it becomes more needed.
BTW, I just love the fact that you can control the scope with VNC or Remmina, even though the screen-update is slighly more laggy than observing it on the scope itself, quite noticeable in XY-mode. Still another way to control the scope remotely.

Saying something would be nice without explaining why is not something we can learn from.

I can also make a point that I estimate that out of millions of scopes sold and used around the world very few in percentage are connected to external references even if they have the input.
For majority of use only parameter that is important is that scope has low internal clock jitter, and funny enough, not so much because scope timing measurements but because of performance of ADC.

What is your scenario where you think you would need scope disciplined to laboratory frequency standard, could you please elaborate?

FFT with an accurate reference, for one.

Ok thanks for answer, I appreciate it!

FFT will have limited frequency resolution. Simplest SDS800xHD has max timebase err. of ±25 ppm.
Bins frequency distribution won't be clean round decimal number but sample rate divided by an power of 2 number of bins.

Meaning that you can't get an frequency of FFT bin exactly centered on 10MHz for instance.
With 2097152 bins and 2GS/s your bin spacing will be 953,67431640625 Hz.
10Mhz peak will actually have to be sorted into 999450,68359375 Hz or 1000404,35791015625 Hz centered bin.
Bin spacing will be ±40,4 ppm

Scope will not perform some kind of parabolic or Gaussian frequency bin interpolation to increase frequency resolution beyond that. Those are advanced analytic methods. It simply does FFT.

On SDS6000A that supports FFT up to 8 Mpts, timebase accuracy is worst case ±2ppm (±3ppm over 20 years) and on SDS7000A that supports 32 Mpts FFT you have both 0,1ppm OCXO option and 10MHz REF IN.

Did you have something else in mind that would make ppb level timebase accuracy important in regards to FFT ?

Best

[rf-loop]

Despite everything, my wish is that we could use a better reference. An oscilloscope is specifically a time axis measuring device. The time axis should be as accurate as reasonably possible. Today we have better than initial +/- 25ppm references available for the price of a shirt button. Why leave time axis in the measuring device to the poor accuracy just to save a penny or some jiao?

On the other hand, how often do we need precision on the time axis.
On the other hand, this inaccuracy appears extremely easily also in this case in the frequency axis FFT measurement. Here's a very modest picture of it that still shows the matter graciously.

In this picture, the signal source is my "house standard" Datum Rb .
Its error is absolutely sure below 1ppb. (I "believe" well better, but my own calibration is a bit aged, and well done cal takes least one week! (and Rb need first run least one week before start fine adjust what then take least one week (because need take very long averages, partially also due to GPS "wow and flutter" )

As can see this individual scope reference is roughly 13ppm off.

But then, very importantly. It is quite stable. In this image there is infinite persistence on and picture is after 40min. (oscilloscope have been on 1 hour before this test because main reason was to do selfcal after just new beta FW.)

As can see there is not so much frequency jitter in oscilloscope internal 25MHz  reference.
(also based to other measurements over whole time I have this unit, this reference do not have notable ageing drift. )

And this quite low jitter is very much more important that nearly constant offset. Still I will ask, why save one penny here. 25ppm to 2ppm TCXO is not big jump in price. it is very different if we start talking about ppb class..

But still independent of this offset there is still jitter. If we use oscilloscope for characterize DUT time jitter...  we need be careful to know if we are measuring oscilloscope or DUT (but about this jitter, I will leave it to other time. )

So it need also be careful when think FFT and freq accuracy. As can see...




In this simple FFT ... freq axis one div can think as 10ppm (100Hz @10MHz)
And as said, input signal is 10.000 000 00x MHz
40min run (persistence set for infinite)

[2N3055]

Despite everything, my wish is that we could use a better reference. An oscilloscope is specifically a time axis measuring device. The time axis should be as accurate as reasonably possible. Today we have better than initial +/- 25ppm references available for the price of a shirt button. Why leave time axis in the measuring device to the poor accuracy just to save a penny or some jiao?
...................

My friend, I don't disagree with you, as you know it 
Better is better.
But in reality, not many users of this very inexpensive scope would even notice.
That is my point.

[tautech]

I'm still wondering how accurate the time keeping would be on these units once the clock is set?   

Preliminary results from a 4hr test between SDS814X HD running internal Linux time and SDS1104X-E on interweb NTP time.
SDS814X HD ~-3s

I'll run this overnight to get ~14hrs from start of test which should be accurate enough to indicate 24 and 48 hr accuracy.

[rf-loop]

Despite everything, my wish is that we could use a better reference. An oscilloscope is specifically a time axis measuring device. The time axis should be as accurate as reasonably possible. Today we have better than initial +/- 25ppm references available for the price of a shirt button. Why leave time axis in the measuring device to the poor accuracy just to save a penny or some jiao?
...................

My friend, I don't disagree with you, as you know it 
Better is better.
But in reality, not many users of this very inexpensive scope would even notice.
That is my point.

I understand better than just well your point.   What is also my opinion if we look "average" users...
But also, we still need think that even when this is inexpensive and small, it is (and can be)  much more than just inexpensive entry level scope.


And even when time base have offset... once after owner know his unit offset he can take it to count.

Much more bad thing is IF it have high jitter. But this do not have - least when we look this instrument in its class



Here, it is certainly good for the device owners and those considering the purchase to know that jitter is something completely different than someone who can too quickly be scared by the +/-25ppm rating in the data sheet.
Short-term jitter is completely different.

The following picture gives a hint of it.
For information, the setting of the vertical scale of the Trend screen in the oscilloscope menu is the smallest possible, i.e. 5ns. (however, the scale only display a delay time of 100.000ms (total lack of enough decimals in trend vertical scale to display scale values with 5ns resolution)

Signal is still same Datum Rb



(And needless to note, I have noticed the sign of the time numerical values on the time axis in the  Z  window  = The Bug  )
In this image we can detect that data sheet 25ppm is like elephant in porcellain store and nothing to do with jitter (what is much more important).
25ppm would correspond to 250div in this zoom window.

Delay from Trigger position is 100ms. 
Also there in image persistence is 30 sec.
100ns corresponds 1ppm with 100ms
Now there in  Z1 window scale is 10ns/div  ... 0.1ppm 
(over 2 hour I have seen now 0.3ppm max drift.  And then this visible faster jitter up and down inside 30s period (Persistence time)

Not bad at all in this kind of instrument.
But if really want do some serious time jitter measurements then user need know  his instrument limits for avoid situation where we measure oscilloscope instead of DUT.

[ebastler]

(first 30sec: Normal, next 30sec: Eres 3.0, same scenario)

Now there's an example where it becomes limiting to have ERES only as a math function: Unless I am overlooking something, we can't use it at all in XY mode on the SDS800X HD, right?

I didn't get this scope (or any DSO) for its XY mode, but it's a pity that we miss out on this bragging opportunity with the SDS800X HD...

[BRZ.tech]

Dear @pdenisowski
I frequently watch the videos on the R&S YouTube channel.
I think they are excellent THEORETICAL videos.
In my opinion, they need to be completed with practical videos, with real instruments, and without cuts and editing in the videos. It's you connecting the cables, the attenuators, pressing the buttons on the signal generators and the DSO/MSO, and showing the results “On Screen” from the DSO/MSO.
The idea is: if @pdnisowski does it, the user can also reproduce the tests step by step. It is in this condition that a “newbie” learns and keeps something in his mind.

About half of the 200+ videos I've made are about how to do things on our instruments.  I spent the first 20 years of my career in T&M demonstrating things on real instruments using real DUTs in real labs with real customers, so I like to think I'm pretty comfortable doing that   That said, I use the "narrated PowerPoint" method even for the practical videos for a number of reasons, one of which being that the "slides" can be used for in-person, classroom training and reference, whereas a video typically has to stand alone. 

Who watches your theoretical videos?
Maybe the “newbies” like me. Perhaps an “expert” does not have the patience to watch these videos considered “boring and long” and with lots of details. But “newbies” will really appreciate it. And the R&S concept will become much more popular, beyond the professional market.

The "theoretical" videos (I call them "technology" videos) are literally ten times more popular in terms of views, likes, and comments compared to the "product videos"  And as of last week, the "theoretical" videos had over 3 million total views. So somebody is watching them

And again, from my experience working with thousands of T&M customers: even people who are experts in one field often want a brief (~10-15 mins) explanation of a technology or measurement they may be unfamiliar with.  And they want it without chit-chat, advertising, jokes, a flashy intro/outro, etc.  That's my target audience.

This is not a personal criticism of you or R&S, it is a suggestion for consideration.

Not at all!  I very much appreciate any and all feedback.  Although my "formula" seems to work, I'm always looking for ways to improve and expand, so your post is very helpful.  Thanks!

Hi
Dear @pdnisowski,
Thank you for responding to my post.
I understand your and R&S's concerns:
“Education/training is very important to us here at R&S: we want our customers to be successful when using our instruments!”

I think your “narrated Power Point” method is very good. I have no criticism of his method and technical content presented in all R&S videos.
His voice delivery and word pronunciation speed are excellent, even the YouTube translator can keep up.

I even get worried:
“This is not a personal criticism of you or R&S, it is a suggestion for you to consider.
No way! I really appreciate any and all feedback. Although my “formula” seems to work, I am always looking for ways to improve and expand, so your post is very helpful. Thanks!"

I watched very carefully your video on “technology” for measuring “Slew Rate” in SMPS, which also applies to Operational Amplifiers. Very instructive, as you say, it is a “technology” video.

Here on the other side of the world, I am already thinking soon, as a “newbie”, to apply this “technology”, and do a test with the IC TL072, which is quite popular, check and be able to measure the (Slew Rate = 20 V /us), carry out tests at Various Frequencies, change the Signal Amplitudes, and check why the Slew Rate changes with the Amplitude Variation, and does not change with the Frequency Variation. And at what signal amplitude did TI or other, define (Slew Rate = 20 V/us) for the TL072.
And clearly I cannot abuse your kindness, or @Performa01, or other experts to do this.

@pdnisowski, there is one thing I disagree with you:
I really like knowing how the DSO/MSO makes its automatic measurements, but using the Cursors to measure the Slew Rate, or the Rise Time, I did it once, I learned how to use the tool, but on a daily basis, I prefer use the Instrument's automatic measurement, as generally the number scales are not whole numbers, and this is bad for locating where the 10% and 90% of the pulse are, it can cause a lot of inaccuracy in the reading and calculations.

Translation with Google Translate.
All the best.

[pdenisowski]

I really like knowing how the DSO/MSO makes its automatic measurements, but using the Cursors to measure the Slew Rate, or the Rise Time, I did it once, I learned how to use the tool, but on a daily basis, I prefer use the Instrument's automatic measurement, as generally the number scales are not whole numbers, and this is bad for locating where the 10% and 90% of the pulse are, it can cause a lot of inaccuracy in the reading and calculations.

Yes, automated measurements are generally always better than "manual" cursor based measurements.  Even though all of our oscilloscopes can automatically measure this, I included the cursor-based method because some scopes (and particularly, analog scopes) cannot.

The cursor method also allows me to use my old-school RPN calculator and turn it into an Instagram post

[tautech]

I'm still wondering how accurate the time keeping would be on these units once the clock is set?   

Preliminary results from a 4hr test between SDS814X HD running internal Linux time and SDS1104X-E on interweb NTP time.
SDS814X HD ~-3s

I'll run this overnight to get ~14hrs from start of test which should be accurate enough to indicate 24 and 48 hr accuracy.

Result for ~14hrs
SDS814X HD running on Linux time lost 7s vs SDS1104X-E on NTP time.

[BillyO]

SDS814X HD running on Linux time lost 7s vs SDS1104X-E on NTP time.

That's considerably more than 25ppm.  About 6 times.  It's also far worse than any watch or clock I have.

Maybe an enquiry to Siglent to find out where the OS timebase is coming from?

[tautech]

SDS814X HD running on Linux time lost 7s vs SDS1104X-E on NTP time.

That's considerably more than 25ppm.  About 6 times.  It's also far worse than any watch or clock I have.

Maybe an enquiry to Siglent to find out where the OS timebase is coming from?

Which fuels this thought....
Maybe it would have been prudent to manually Sync SDS1104X-E NTP time before taking the measurement.....

[ebastler]

That's considerably more than 25ppm.  About 6 times.  It's also far worse than any watch or clock I have.

Maybe an enquiry to Siglent to find out where the OS timebase is coming from?

Why does clock accuracy matter to you? Do you plan to keep the scope running 24/7? It is a bit of a power hog...