SDS1104X-E DC offset in AC coupling

[Deso]

Hi all,

I just joined the club of the SDS1104X-E owners. It replaced my good old DS1052E.
However, my joy lasted less than a minute until I spotted that there is a DC offset on all channels. I thought that it needs some time to get properly warm, so I left it for and more than 30 minutes (hours). BNCs started feel warm (39 C). DC offsets have improved but are not there yet. I switched to AC coupling and the DC offset became bigger. This behavior is also observed for the 4 channels. I ran the calibration (2 times) and the DC offset in DC coupling reset as it should. Switching to AC coupling shows a DC offset again. All the tests were performed without anything connected on the inputs, with connected and shorted probes and even with 50 ohm BNC terminator. No changes to the DC offsets in AC coupling.
I ran the very same tests with my old Rigol and there were no differences in the offset when switching between AC and DC coupling. Initially there was a large offset that was reset when the scope warmed up and didn't change for hours.
So my question is - is this a thing that I don't understand why it's happening, or is it a hardware issue?

Here are the scope details:
SW: 6.1.37R8
OS: 8.3
FPGA Ver: 2021-11-08
HW: 09-06

Thanks in advance

[tautech]

Welcome to the forum.

Install 37R9 firmware.
https://int.siglent.com/upload_file/zip/firmware/Oscilloscope/SDS1xx4X-E_V6.1.37R9_EN.zip

Release note:
Fixed the problem: The skew of the two ADCs is not accuracy for hardware of 09 version (System Status shows Hardware Version: 09-xx). The other hardware version has no this problem.

[Deso]

Thanks for the hint. I already updated the scope to 6.1.37R9. On cold scope and AC coupled channel it shows no DC offset. On DC coupling shows negative offset. After a few minutes the offset started to slowly increase to a positive values. I did another calibration as per the FW update instructions. It didn't help at all. Everything is as it was before the FW update. Even worse - the offsets became larger.

[Ringmodulator]

Please perform the self-calibration routine after proper warm up and report back.

Regards
Chris

[mwb1100]

And just in case: the probes should be disconnected when doing the self-calibration.

[Deso]

Please perform the self-calibration routine after proper warm up and report back.

Done. Still no joy. The offset in AC coupling became bit smaller, but the one in DC coupling also was changed.

And just in case: the probes should be disconnected when doing the self-calibration.

Yes, it was done in that way.

[Bionicbone]

Oh, I have one of these in the post, I will try to follow what you did here to check when it arrives today but it's my first scope so who knows lol.

[Bionicbone]

This is mine, but I have HW version 01-05 and it's been switched on for a couple of hours.

I'm not sure what this means to be really honest, are you expecting perfect zero?

[Deso]

Thanks for sharing.
I'm expecting to see no difference in the offset when switching between AC and DC coupling. I'm expecting small offset in general.
My old Rigol is showing no difference when switching between AC/DC coupling. I asked a friend with an Agilent 2 series oscilloscope to check his offsets and although there is some DC offset, it has a constant value between the AC and DC coupling setting.

[bdunham7]

I'm expecting to see no difference in the offset when switching between AC and DC coupling. I'm expecting small offset in general.

I don't know the exact details, but the input amplifier on these has a stage before the AC-coupling so that the input impedance is always 1M + 15pF (or so) regardless of which coupling you chose.  This first part will have some offset of its own and that will be blocked by the AC-coupling, so you do see some change in DC offset when changing coupling.  You should compare all three modes--DC, AC and GND.  On mine, the offset is about the same on AC and GND, but slightly different on DC.

Edit:  On mine, all three are a bit different.  DC has the most offset, AC has less and GND the least.  All offsets on all channels are 1 division or less.

As for the amount of the offset, you can read the specs and see what the tolerances are.  When comparing to other instruments, you need to make sure you are comparing at the same scale.  Do those other scopes have a 500uV/div setting?  Yours does look to have more offset than my older SDS1104X-E, but I don't think it is out of spec.  Offset may make it slightly annoying to do measurements on very small signals, but other than that it really isn't an issue.

[Bionicbone]

Did you see the OS update for hardware 09?

[Deso]

Did you see the OS update for hardware 09?

Yes. I'm on OS v3 from the factory.

I don't know the exact details, but the input amplifier on these has a stage before the AC-coupling so that the input impedance is always 1M + 15pF (or so) regardless of which coupling you chose.  This first part will have some offset of its own and that will be blocked by the AC-coupling, so you do see some change in DC offset when changing coupling.  You should compare all three modes--DC, AC and GND.

I've looked at Dave's images of the SDS1104X-E frontend and it doesn't look like they have anything active before the AC/DC coupling relay.
Moreover, they look extremely close to the Hantek DSO5000P schematic. Of course, there is some difference, but the basic idea and signal flow seem to be the same. So it looks like there are no active parts before AC/DC coupling.
What's more, the SDS1102X-E's specification lists an input impedance for DC coupling only.

As for the amount of the offset, you can read the specs and see what the tolerances are.

I couldn't figure it out: "Offset Accuracy: ±(1%* Offset+1.5%*8*div+500 uV): ≤1 mv/div"
Current situation is that I have 2 divs (1mV) error at 500uV/div out of 8 divs, which is 25%. The error is visible and measurable even at 5mv/div with AC coupling.

I  When comparing to other instruments, you need to make sure you are comparing at the same scale.  Do those other scopes have a 500uV/div setting?

No. But the tests were performed at their lowest possible setting. But to be fair - below in the attachments are the same measurements but this time at 2mv/div on both scopes. Still, the old one beats this one by a lot.

[bdunham7]

I've looked at Dave's images of the SDS1104X-E frontend and it doesn't look like they have anything active before the AC/DC coupling relay.
Moreover, they look extremely close to the Hantek DSO5000P schematic. Of course, there is some difference, but the basic idea and signal flow seem to be the same. So it looks like there are no active parts before AC/DC coupling.
What's more, the SDS1102X-E's specification lists an input impedance for DC coupling only.

Well, it would be interesting to know the exact setup here, perhaps someone can enlighten us.  However, the input impedance measured directly on mine is 1M in DC and about 1.2M in AC and GND.  So something changes but they are not just using a blocking capacitor like you'd see in many older CROs and perhaps other modern models as well.  Those would indicate open with a DMM.

I couldn't figure it out: "Offset Accuracy: ±(1%* Offset+1.5%*8*div+500 uV): ≤1 mv/div"
Current situation is that I have 2 divs (1mV) error at 500uV/div out of 8 divs, which is 25%. The error is visible and measurable even at 5mv/div with AC coupling.

No. But the tests were performed at their lowest possible setting. But to be fair - below in the attachments are the same measurements but this time at 2mv/div on both scopes. Still, the old one beats this one by a lot.

TBH, I have to agree that seems like a lot.  Certainly far more than mine. 

[Deso]

So, several days after my offset findings I hit another snag - in trigger setup menu, when switching to AC or LF Rejecting Coupling - the trigger marker line disappears. It is not above/below the current screen and it works, but is not visible at all. It shows only hint with the current threshold voltage. The channel coupling does not matter for this issue.
I thought a 5+ year old design should be clean of software bugs.

Another thing - the Trigger Source.. it stays always to the channel that have been selected even if this channel is not active anymore. It would be nice to automatically switch to the current active channel if the selected has been disabled.

[bdunham7]

So, several days after my offset findings I hit another snag - in trigger setup menu, when switching to AC or LF Rejecting Coupling - the trigger marker line disappears. It is not above/below the current screen and it works, but is not visible at all. It shows only hint with the current threshold voltage. The channel coupling does not matter for this issue.
I thought a 5+ year old design should be clean of software bugs.

That's neither a bug nor feature, just the way things work.  If you displayed a marker at the selected trigger threshold (say 0.0V) with AC-coupled triggering and then looked at an AC signal with a DC bias, you would then complain that the trigger event doesn't occur at the intersection of your signal and the displayed trigger marker.

[Deso]

Thanks @bdunham7. It makes sense. However, if we have Trigger with AC Coupling and channel with AC coupling then the line should be visible, but it is not.
The other thing unclear to me is with "LF reject" and "HF reject" - the first one shows no line the second shows the line. Why is that?

[bdunham7]

Thanks @bdunham7. It makes sense. However, if we have Trigger with AC Coupling and channel with AC coupling then the line should be visible, but it is not.
The other thing unclear to me is with "LF reject" and "HF reject" - the first one shows no line the second shows the line. Why is that?

LF reject is going to have the same issue as I mentioned above in some cases, even if both are AC coupled.  If both are just AC coupled, then perhaps you could show the marker without any error in the trigger location--I'd have to think about that and perhaps experiment a bit.  But most DSOs don't, AFAIK.

[Performa01]

When discussing topics like trigger coupling, we should first understand its uses.

Different modes of Trigger coupling give us the opportunity to filter the input signal before it is checked for the trigger condition.

DC is unfiltered, i.e. from DC up to more than the bandwidth of the scope
HF-Reject is DC to about 1.2 MHz
AC is from ~8 Hz up to more than the bandwidth of the scope
LF-Reject is ~1.2 MHz up to more than the bandwidth of the scope

So it should be clear, that DC and HF-Reject are basically the same, just with a different upper bandwidth limit, whereas AC and LF-Reject are basically the same as well, just with a different lower corner frequency.

For most use cases, DC-Trigger is fine – and you should know what you're doing when you choose something different. Why would you ever use AC trigger coupling at all?

Usually, we make that choice for the input channel coupling, and if this is AC, then the trigger is inevitably AC coupled as well. This leaves the rare cases, where we:

a)   don't have a (constant) DC offset, but a low frequency signal that is mixed with the signal we want to analyze and that appears as a constantly changing offset
b)   still want to use DC input coupling, so that we can observe the (changing) offset

and don't want to lose triggering at times.

I can't think of many use cases, where these conditions are met. But if you run in such a situation, the trigger filters, in particular HFRJ and LFRJ, help to achieve stable triggering even in tricky situations. Of course it would be even better if we had variable corner frequencies for this, but that would use too many resources, therefore such a feature is usually found in high-end DSOs only.

And maybe now it becomes clear why Siglent is reluctant to show the trigger level indicator in all the non-DC trigger coupling scenarios. This indicator is not stable, but might have to move quite fast – and it simply isn't fast enough for that. While the actual trigger level is adapting instantly, the trigger indicator lags behind a bit.

I know, the old Rigol DS1052 does show the trigger indicator in AC-trigger coupling mode, but many others do not. And wouldn't a fast moving trigger indicator rather be confusing while not giving any essential information? The main thing is, that with AC coupling, the trigger level is related to the signal offset, not an absolute screen position, as it would be with DC coupling.

On top of all that, trigger engines have evolved over time. Older scopes had an analog trigger, whereas modern DSOs like the SDS1104X-E have a fully digital trigger engine. The latter is far more powerful and more accurate, but not fully independent from the signal acquisition.

Hint: the EXT trigger found in all modern Siglent DSOs (except 4 channel X-E) is analog.

It might be helpful to study this document, especially reply #2, part 4 of the review:

https://www.eevblog.com/forum/testgear/siglent-sds1104x-e-in-depth-review/

[Bionicbone]

I was reading this, what surprised me was AC coupling should not be an automatic selection just because you are measuring AC.
It's an interesting read from keysights
https://blogs.keysight.com/blogs/tech/bench.entry.html/2018/10/18/when_to_use_ac_coupl-X0K7.html#:~:text=DC%20coupling%20allows%20you%20to,out%20all%20the%20DC%20components.

[tautech]

I was reading this, what surprised me was AC coupling should not be an automatic selection just because you are measuring AC.

Of course, doesn't everyone already know this ?
AC coupling removes the DC component of a signal, nothing more and nothing less.

I rarely use AC input coupling as there are is a rarely ever any need to. Measuring mV of ripple on a power rail is the only time it's common to use it.

[Bionicbone]

I am getting used to my first ever scope, have to admit I probably would have selected AC coupling for AC   

I have so much to learn, but that's the main reason for buying one.

[tautech]

I am getting used to my first ever scope, have to admit I probably would have selected AC coupling for AC   

I have so much to learn, but that's the main reason for buying one.

Fear not, we were all scope newbies once. 

What you do have that's really useful for the newbie is a easy method to grab screenshots and if you use some thought beforehand with an appropriate menu captured too !
Then bring it here and upload it whenever you see something you can't work out and many of us will jump in to help with conceivably many levels of understanding.

While this link is not for the 4ch X-E, SDS1202X-E the predecessor to SDS1104X-E, shares many but not all the same features so there is much in this thread to guide you about the feature set in your scope:
https://www.eevblog.com/forum/beginners/what-an-oscilloscope-recommended-for-a-woman-passionate-about-electronics/msg3104228/#msg3104228

[Deso]

When discussing topics like trigger coupling, we should first understand its uses.

Different modes of Trigger coupling give us the opportunity to filter the input signal before it is checked for the trigger condition.

DC is unfiltered, i.e. from DC up to more than the bandwidth of the scope
HF-Reject is DC to about 1.2 MHz
AC is from ~8 Hz up to more than the bandwidth of the scope
LF-Reject is ~1.2 MHz up to more than the bandwidth of the scope

So it should be clear, that DC and HF-Reject are basically the same, just with a different upper bandwidth limit, whereas AC and LF-Reject are basically the same as well, just with a different lower corner frequency.

For most use cases, DC-Trigger is fine – and you should know what you're doing when you choose something different. Why would you ever use AC trigger coupling at all?

Usually, we make that choice for the input channel coupling, and if this is AC, then the trigger is inevitably AC coupled as well. This leaves the rare cases, where we:

a)   don't have a (constant) DC offset, but a low frequency signal that is mixed with the signal we want to analyze and that appears as a constantly changing offset
b)   still want to use DC input coupling, so that we can observe the (changing) offset

and don't want to lose triggering at times.

I can't think of many use cases, where these conditions are met. But if you run in such a situation, the trigger filters, in particular HFRJ and LFRJ, help to achieve stable triggering even in tricky situations. Of course it would be even better if we had variable corner frequencies for this, but that would use too many resources, therefore such a feature is usually found in high-end DSOs only.

And maybe now it becomes clear why Siglent is reluctant to show the trigger level indicator in all the non-DC trigger coupling scenarios. This indicator is not stable, but might have to move quite fast – and it simply isn't fast enough for that. While the actual trigger level is adapting instantly, the trigger indicator lags behind a bit.

I know, the old Rigol DS1052 does show the trigger indicator in AC-trigger coupling mode, but many others do not. And wouldn't a fast moving trigger indicator rather be confusing while not giving any essential information? The main thing is, that with AC coupling, the trigger level is related to the signal offset, not an absolute screen position, as it would be with DC coupling.

On top of all that, trigger engines have evolved over time. Older scopes had an analog trigger, whereas modern DSOs like the SDS1104X-E have a fully digital trigger engine. The latter is far more powerful and more accurate, but not fully independent from the signal acquisition.

Hint: the EXT trigger found in all modern Siglent DSOs (except 4 channel X-E) is analog.

It might be helpful to study this document, especially reply #2, part 4 of the review:

https://www.eevblog.com/forum/testgear/siglent-sds1104x-e-in-depth-review/

You have done a great job with this review. Kudos to that.
I read the sds1004xe manual again and indeed found the information regarding LF and HF filtering for the trigger coupling.

Usually, we make that choice for the channel input and if we select AC coupling there, the trigger will inevitably be AC coupled as well.

I agree. However it is not labeled as such and if you manually switch to Trigger AC coupling - you have no level indicator. And really 99% of the time I won't touch this menu, but it should still display/work properly for those who care.

Back to main topic - While waiting for Siglent to resolve my issue with the AC coupling offset I spotted that there is a new firmware - V6.1.37R10 (EU: Release Date 04.03.23, NA: Release Date 03.30.23 ) which added "Optimized VGA configuration".
It looks like it might help me with my compensation issues. Unfortunately, not only did it not solve my problem, but it introduced a new one - interruption of the trace updating on specific interval. The trace is freezing for a fraction of the second. During one of my VNC video capturing even the oscope itself froze for a while. This issues is visible on the scope screen too.

And my input channels offsets are behaving strange. After firmware update and reboot, I waited for 40 min to warm a bit and I did calibration. The result shows no differences from the previous firmware.
Now I turn on the scope after several hours again to capture the trace drawing issue and the offsets are different, even after proper warm-up.

At the end my short straw seems to be shorter than I expected.

Here is a video with the new FW issue
https://drive.google.com/file/d/1YZMT3PAPBnjCt5um3wm5XVQ1QbDt7Kdh/preview

[pope]

What is the VGA configuration?

[BillyO]

What is the VGA configuration?

Variable Gain Amplifier.  They are used in the front ends of scopes (and possibly other places)..

https://www.ti.com/lit/gpn/lmh6518