Siglent SDS1104X-E and SDS1204X-E Mixed Signal Oscilloscopes

[Performa01]

Attached is an old screenshot showing the phase error for all three Bode Plot channels with Ch.1 being the reference.

We can see that the actual skew is about 0.926 ns (40° phase shift at 120 MHz).
A skew less than 1 ns at a sample clock of 1 GHz for each ADC (500 MHz for each channel) isn't that bad after all - the only question remains why there is no deskew like during normal operation?

Of course I've brought this to Siglent's attention back then and the fact that it isn't fixed yet is either because of the very low priority of the Bode Plot or there are actually some technical reasons that we aren't aware of.

The 2nd screenshot shows the same scenario when viewed in traditional Y-t mode. Here we can see that the skew is well compensated (the remaining skew is due to the setup, not the scope).

[CharlotteSwiss]

slightly off topic, but not too much: I have a younger brother (siglent 1202x-e); it often happens that when I print to save the screen, the oscilloscope freezes (no key responds), to turn it off I have to unplug it; does anyone in here have this unpleasant problem?
thank you
char 
 

[blurpy]

slightly off topic, but not too much: I have a younger brother (siglent 1202x-e); it often happens that when I print to save the screen, the oscilloscope freezes (no key responds), to turn it off I have to unplug it; does anyone in here have this unpleasant problem?
thank you
char 
 

I do too, on my 1204X-E. There's been some discussion on that in this thread: https://www.eevblog.com/forum/testgear/siglent-technical-support-join-in-eevblog/msg3067512/#msg3067512

Haven't heard anything from Siglent on the matter.

[CharlotteSwiss]

slightly off topic, but not too much: I have a younger brother (siglent 1202x-e); it often happens that when I print to save the screen, the oscilloscope freezes (no key responds), to turn it off I have to unplug it; does anyone in here have this unpleasant problem?
thank you
char 
 

I do too, on my 1204X-E. There's been some discussion on that in this thread: https://www.eevblog.com/forum/testgear/siglent-technical-support-join-in-eevblog/msg3067512/#msg3067512

Haven't heard anything from Siglent on the matter.

thanks  blurpy for the testimony, so it's not just my 1202 problem ...
That Siglent pretends nothing is a good thing ...

 

I add: see my message below, did you see that a user solved it by returning to firmware .23?

[borjam]

In my experience memory sticks can be a hit or miss. I was going to upgrade the firmware on my spectrum analyzer now. The first unti I grabbed didn't work on the Siglent, but the second did work flawlessly.

[CharlotteSwiss]

In my experience memory sticks can be a hit or miss. I was going to upgrade the firmware on my spectrum analyzer now. The first unti I grabbed didn't work on the Siglent, but the second did work flawlessly.

sure, but when the Print problem occurs 4 times out of 5 ...
However, it seems that a user has fixed this bug by reverting to the previous firmware ..

https://www.eevblog.com/forum/testgear/siglent-technical-support-join-in-eevblog/msg3070454/#msg3070454

[blurpy]

thanks  blurpy for the testimony, so it's not just my 1202 problem ...
That Siglent pretends nothing is a good thing ...

 

I add: see my message below, did you see that a user solved it by returning to firmware .23?

I noticed. Not sure if the same apply to older firmware for 1204X-E, but possibly.

[rf-loop]

In my experience memory sticks can be a hit or miss. I was going to upgrade the firmware on my spectrum analyzer now. The first unti I grabbed didn't work on the Siglent, but the second did work flawlessly.

With over 10 years experience with Siglent (and some others) I have always used Kingston sticks.
Older times 1G, then 2G and later 4 and 8G. Last years always 4 and 8G USB2.0 sticks after I find that these need keep and care because more and more difficult to find.

Only just in true genuine FAT32 format with 4k cluster size. (even if some products may accept different but for avoid mess because all accept this) Mostly FW update files just alone in root and naturally never ever  more that just one partition. These sticks are so cheap that no reason for gambling.
And of course Never exFAT, never NTFS. 
I have also always formatted these in Windows 7 or previously older so that I can be sure it is genuine FAT32. Careful with sticks what have used for some other purposes, they may need deep format for remove all even hidden things and residues. There is some special tools for these.

With these my own rules only for me I have never meet any single problem and I have done these more than just several times and quite wide range of Siglent products. I will keep these bit older sticks reserved just for this purpose.

[nez]

Attached is an old screenshot showing the phase error for all three Bode Plot channels with Ch.1 being the reference.

We can see that the actual skew is about 0.926 ns (40° phase shift at 120 MHz).
A skew less than 1 ns at a sample clock of 1 GHz for each ADC (500 MHz for each channel) isn't that bad after all - the only question remains why there is no deskew like during normal operation?

Of course I've brought this to Siglent's attention back then and the fact that it isn't fixed yet is either because of the very low priority of the Bode Plot or there are actually some technical reasons that we aren't aware of.

The 2nd screenshot shows the same scenario when viewed in traditional Y-t mode. Here we can see that the skew is well compensated (the remaining skew is due to the setup, not the scope).

Well, quick update.

It looks like Deskew *IS* used in BodePlot II.....  but I have discovered that there are problems with Deskew that made it just appear by chance like it wasn't being used .

Deskew seems to work fine at 1 ns/div timebase, but is very broken at other timebases -- and those other timebases happen to be what Bode plots use while being generated.


For example:
Using the same ADC, e.g. Chans 1+2.

If I have two signals that are about 780 ps skewed, I can go to 1 ns/div timebase and adjust Deskew to get them to match perfectly.

However, when I then change to, say, 50 ns/div the phase measurement suddenly jumps back to what seems to be the orginal phase difference! (The Deskew value is still set)  Now, further adjusting the Deskew up or down doesn't affect the skew (visually or auto-measured) at all until 1 ns thresholds are reached -- at which point the skew measurement jumps another full ns to a new value and stays there until the next full ns higher/lower Deskew increment is reached.

If I change the Deskew enough to hit a threshold that it feels like changing the phase measurement at, *then* it will affect the Bode plot traces.

I see this weird Deskew problem below 50 ns/div as well.  20 and 10 ns/div are easy to see, while at 5 ns/div it is more subtle, but noticeable when changing from 0 to +10ps Deskew.

Hopefully this description makes sense.


I have been making a write-up to break down and describe the different issues I'm seeing, but it's not ready to post tonight (the testing is kind of tedious ).  I'll add this to the list.

[HendriXML]

This might be caused a low horizontal resolution and thus roundoff. The resulting issues depends where in the pipeline is calculated/processed (deskewed) what.
The coarse calculation of the fase difference at 50 ns seems also a result of this.
So it's probably not a threshold, but "steps" due to round offs.

As a side note: some time ago I was trouble shouting roundoff errors in the voltage output of the SAG1021. I eventually couldn't find the exact way the calculation where sloppy, but the effect could be brought into the light.
The bottomline is when doing (integer) calculations do it in the right order and analyze what's rounded up or down thoroughly.

[Performa01]

It looks like Deskew *IS* used in BodePlot II.....  but I have discovered that there are problems with Deskew that made it just appear by chance like it wasn't being used .

Deskew seems to work fine at 1 ns/div timebase, but is very broken at other timebases -- and those other timebases happen to be what Bode plots use while being generated.

Well, I've done rather deep investigations on many topics, some of them regarding phase measurements in general, also found (and reported) some related bugs back in 2018, but didn't further investigate the Bode Plot phase error between the two ADCs. Also, I  had used FW 6.1.33 and we cannot know what has changed since then.

For example:
Using the same ADC, e.g. Chans 1+2.

If I have two signals that are about 780 ps skewed, I can go to 1 ns/div timebase and adjust Deskew to get them to match perfectly.

However, when I then change to, say, 50 ns/div the phase measurement suddenly jumps back to what seems to be the orginal phase difference! (The Deskew value is still set)  Now, further adjusting the Deskew up or down doesn't affect the skew (visually or auto-measured) at all until 1 ns thresholds are reached -- at which point the skew measurement jumps another full ns to a new value and stays there until the next full ns higher/lower Deskew increment is reached.

This is the expected behavior. Basically, you cannot have a finer granularity than 1 ns with 1 GSa/s rate. For finer steps, we need interpolation. At 50 ns/div, we have the optimum setting with the highest waveform update rate, just because this is the shortest record length that doesn't need interpolation. Faster timebases require interpolation and consequently, the deskew can be performed with increasingly higher resolution.

If I change the Deskew enough to hit a threshold that it feels like changing the phase measurement at, *then* it will affect the Bode plot traces.

I see this weird Deskew problem below 50 ns/div as well.  20 and 10 ns/div are easy to see, while at 5 ns/div it is more subtle, but noticeable when changing from 0 to +10ps Deskew.

Hopefully this description makes sense.

Yes, it does. Hopefully my explanation makes sense too

For all measurements, the data is coming from the sample memory and not from the screen. Consequently, additional data buffers with interpolated data (at certain spots) would be required for accurate measurements below 2 ns resolution (with 4 channels in use).

I cannot check it right now, but I'm suspecting something. Maybe the data for measurements comes from the sample buffer for timebases >=50 ns/div and from the screen otherwise. Apparently that wouldn't be the case for Bode Plot right now…

[CharlotteSwiss]

Older times 1G, then 2G and later 4 and 8G. Last years always 4 and 8G USB2.0 sticks after I find that these need keep and care because more and more difficult to find.
Only just in true genuine FAT32 format with 4k cluster size.

sticks with a few GB are hardly found anymore, my smallest is 8GB.
But I saw that it was formatted fat 32 but at 8k allocation, now I try to 4k

[CharlotteSwiss]

I noticed. Not sure if the same apply to older firmware for 1204X-E, but possibly.

ops, I was confused that you had 1202 as mine ..

However in my opinion without having to go back to the previous firmware (updates are lost), Siglent can very well correct the problem with a subsequent firmware; you just have to want to, after all, customers spend money to buy their products.
Siglent 1202 is my first product of this brand, but if the engineers do not solve the Print button problem, I will not buy anything from this brand anymore. 
 

[borjam]

it's usually a good idea to stick to well known brands and to be careful where you purchase them.

There are lots of counterfeits of some of the big brands.

In the pro audio world, where SD cards are widely used it can be a real nightmare.

[Fungus]

it's usually a good idea to stick to well known brands and to be careful where you purchase them.

There are lots of counterfeits of some of the big brands.

In the pro audio world, where SD cards are widely used it can be a real nightmare.

It's a nightmare, the jitter produced by those cheap SD cards is unbelievable!

[borjam]

The jitter produced by those cheap SD cards is unbelievable!

SD cards are an atrocious invention. Most of them have optimizations for FAT filesystems. So if you try to format
them using something else you are doomed.

The spec was designed for writing in bursts.

I really miss Compact Flash cards. But of course CFs are much more expensive and the connector is much larger.

WAIT: you are not talking about sample jitter, right?

[Fungus]

WAIT: you are not talking about sample jitter, right?

I'm afraid I was.

I failed to see how any SD card could be too awful for audio use so I assumed audiophoolery.

[borjam]

WAIT: you are not talking about sample jitter, right?

I'm afraid I was.

I failed to see how any SD card could be too awful for audio use so I assumed audiophoolery.

No no no!!

Pro audio recorders stream audio to the card almost continuously (in blocks of course, the recordings are stored as files, nothing surprising here, so no jitter audio voodoo!) and they usually read it back.

It sounds silly, but the store and forget approach would prove disastrous if you spent a whole day recording only to find garbage when you go back to the studio. It's much better to detect a failing card right away.

This places quite a stress on a medium designed mostly for bursty writes (a photo now and then). So it's common to have problems (write timeouts, etc) with cards from who knows where and counterfeits.

Audio manufacturers use to keep a list of verified card models. And, again, better make sure to source them from a reliable supplier.

No audiofoolery here, keep calm and carry on! 

[zserg]

Has already someone botherd to exchange the Intensity/Adjust encoder with a detented encoder?

I've been wondering the same thing. 

I have done it right after purchasing bc this non-detent encoder was driving me nuts - pretty weird design decision given that they have detented ones for vertical/horizontal anyway.
In fact, I've even taken the photos to dump here, but I'm a lazy arse so that never happened. 
I can't remember the P/N of the original non-detented one, but it turned out to be some sort of Chinese unobtanium. After searching a bit through what ALPS/Bourns/Omron offer, the closest I was able to find dimension-wise was Bourns PEC12R-4220F-S0024.
To my taste it has a bit too much detent force, otherwise it has been working absolutely fine for the last half year.


UPD:
I still have the original part, it has no P/N, but LJV embossed on it.

[SMB784]

Has already someone botherd to exchange the Intensity/Adjust encoder with a detented encoder?

I've been wondering the same thing. 

I have done it right after purchasing bc this non-detent encoder was driving me nuts - pretty weird design decision given that they have detented ones for vertical/horizontal anyway.
In fact, I've even taken the photos to dump here, but I'm a lazy arse so that never happened. 
I can't remember the P/N of the original non-detented one, but it turned out to be some sort of Chinese unobtanium. After searching a bit through what ALPS/Bourns/Omron offer, the closest I was able to find dimension-wise was Bourns PEC12R-4220F-S0024.
To my taste it has a bit too much detent force, otherwise it has been working absolutely fine for the last half year.
(Attachment Link) (Attachment Link)

UPD:
I still have the original part, it has no P/N, but LJV embossed on it.

I may do this upgrade, as I would really enjoy a detented encoder on the intensity/adjust encoder.  Is there anything special that I should know about this upgrade before doing it?

[zserg]

I may do this upgrade, as I would really enjoy a detented encoder on the intensity/adjust encoder.  Is there anything special that I should know about this upgrade before doing it?

Not anything that comes to my mind straight away. Unsoldering the old one was pretty straightforward with the desoldering wick.
The encoder I've listed above makes an audible click when turning even at low speeds - one might find it annoying.
I'd prefer a "softer" one with lower turning torque, but this Bourns was the only option I was able to source locally quickly.

[jlo]

Hi, planning to buy my first oscilloscope, is the 1104X-E still recommended?

I'm studying to become an electrical engineer and will use it for audio projects with oscillators and filters as well as more digital automation projects with microcontrollers. As I learn more I want to be able to progress to more advanced projects.

[tautech]

Hi, planning to buy my first oscilloscope, is the 1104X-E still recommended?

I'm studying to become an electrical engineer and will use it for audio projects with oscillators and filters as well as more digital automation projects with microcontrollers. As I learn more I want to be able to progress to more advanced projects.

Welcome to the forum.

You can't go wrong with one.

Look at its competitors, who else offers 500uV/div sensitivity, two 1 GSa/s ADC's each with 14 Mpts memory support, Bode plot and web server capability.
As a first scope these will serve you well for many years.

[4x1md]

I like my unit very much. It was a very good investment in my home lab. In general it has a good hardware with software and UI/UX which can be much improved. During the lockdown in March I used it for working at home and the results I got were as good, as the results I got from much more expensive Tektronix and Keysight scopes I use at work. Still, compared to Tektronix and Keysight I have at work, it feels cheaper but you get what you pay for.

[Martin72]

who else offers 500uV/div sensitivity

When they were real, that´s another question.
By the way:

Today, we want to save a signal as matlab file to usbstick.
To be more precisely, the FFT.
Unfortunately, it saves only the channel, not the mathfunction from it - Did we miss something to do ?
Or couldn´t do the sds1104X-E it in general ?
If yes, that would be a showstopper.