Siglent SDS 1022DL strange behavior while scrolling a waveform

[tzok]

Today I've encountered some unusual behavior of my Siglent SDS 1022DL. I've tried to capture a single waveform. Capture has been done in single mode, with 5ms/div and on falling edge. After the waveform has been captured I've switched to 2,5ms/div and tried to scroll the image to the end of a buffer, and that is what I have seen:
https://youtu.be/nwD1TYIesJ0

[Siglent America]

Hi TZOK.
Engineering said that they were able to duplicate this so they will analyze the problem and fix it.
Thanks.

[tzok]

That is a great news. So maybe they could watch this videos:

This one they know, and all but the trigger offset have been somehow addressed. Yet in XY ADC offset is still applied when ADC is turned off (channel to GND), and incorrectly subtracted when channel is inverted:

The trigger offset is weird as it depends on trace position on the screen (moving trace up and down, changes trigger level), and also, even for trace offset = 0, the reported/displayed trigger level is shifted from the real trigger level. Same problem exists in LeCroy WaveAce 1001 firmware (video isn't mine, I've found it on YouTube):

[rf-loop]

This one they know, and all but the trigger offset have been somehow addressed.

Videos are from 2014 and LeCroy video is from 2012. I thing after then there have been some bugs repaired so these videos are totally obsolete.

Trigger feature.

This model have conventional analog trigger circuit and also very simple one. It is not at all as accurate what is today available in more advanced models full digital side trigger system with all its advantages over old analog trigger in conventional DSO. (Example  like SDS1000X or SDS2000X what have very good full digital side trigger engine, one really strong and important point in these models. In this kind of scopes group, top-class.)

This trigger vertical inaccuracy what you show is typical and normal. With this you need live. Also it is hardware "feature". There can not do so much with FW. There is not enough accurate and complex trigger circuit calibration system inside HW/FW for better result.  Also it reads in SDS1000DL models data sheet. (Trigger sensitivity and trigger offset).

Conventional analog trigger is DSO is very different if compare to full digital side trigger system.
In analog trigger system in DSO. User need know and understand that trigger level line over the digitized and displayed waveform is not at all same what happend inside scope. It need realize, it do not trig from image you see! When in TFT screen signal crosses rising or falling edge with your trigger level line there is NOT trigger point. Line is only some kind of help line where it is around.

Disclaimer: There is nothing personal, this is commonn:

Also in simple machines there is not selfcalibration accurately adjusted trigger pathway. There is not hardware for it.
After signal is splitted (even there it is not perfectly equal with signal what is going to ADC, motsly it is end buffer aux  out)  from analog signal what is going to ADC, there in trigger system is this signal handling (analog) and finally trigger comparator where it go to other gate and other gate there is analog trigger level. This level is produced by DAC and this is what level you adjust. Also this include more or less error.  After then there is (and there need be) comparator Hysteresis. Result is sum of several errors where also some errors are moving. (tem and ageing drift) and it need also note that noise what Trig circuit may see is perhaps different what you see after ADC displayed signal.  Also when go to different frequencies, trigger pathway  may have different frequency response and delay. 
Final result is, and this IS natural. sum of lot of different errors. Ideal circuits do not exist in any other place than basic schools theory books. After then we wonder why my trigger level displayed line is not exactly where I see signal level what crosses this line. "Know your equipment and its natural limits and there is less problems"

Simplified basic principles of digital and analog trigger system by R&S
First conventional analog (SDS1000DL/CNL/CML/CFL) and then modern full digital side trig (SDS1000X SDS2000X).

[tzok]

Thank you for explanations, yet I may ensure you that only ADC offsets in XY mode have been slightly corrected (in the oldest video the calibration offset is obviously added instead of subtracted after inverting a channel).
I may understand the trigger offset, but why the trigger level changes when I chance the trace position? I.e. if I set trigger level near the low level of the signal, but so the scope triggers correctly, and then I move the trace up, the scope stops to trigger. Is the trace position also done in the analog path?

[rf-loop]

I may understand the trigger offset, but why the trigger level changes when I chance the trace position? I.e. if I set trigger level near the low level of the signal, but so the scope triggers correctly, and then I move the trace up, the scope stops to trigger. Is the trace position also done in the analog path?

Analog shift is also (of course) done in analog side. .
From digital side (your position adjustment digital level) it go to DAC what produce analog voltage. Trace shift have one DAC and Trigger level have one DAC.  After then trace shift voltage go to input amplifier. DAC is not ideal, amplifier is not ideal. All they have more or less errors in aplifying factor and linearity.  There is not perfect circuits for correct all these in self calibration. We need accept errors, as long as they are in some range what are inside "acceptable" range. As we can see also in data sheet.

Trigger Sensitivity
For fixed gain ranges
1 Divisions: DC-10MHz
1.5 Divisions: 10MHz - Max BW

Trigger Level
Accuracy
(typical) applicable for the signal of rising and falling time ?20ns
Internal? ±(0.2 div×V/div)( within±4 divisions from center of screen)


Vertical input

DC Gain Accuracy
?±3.0%: 5mv/div to 10V/div in Fixed Gain Ranges
?±4.0%: 2mv/div Variable Gain Ranges

DC Measurement
Accuracy:
All Gain settings ?100mv/div
± [3%* ? |reading|+|offset| ? +1% *of |offset|+0.2div+2mv]

DC Measurement
Accuracy:
All Gain settings ?100mv/div
± [3%* ? |reading|+|offset| ? +1%* of |offset|+0.2div+100mv]