Update instructions
Very important!
Version5.1.3.13 optimizes self-calibration for the each channel. Perform a self-calibration once if the machine is updated from any version <5.1.3.13
After oscilloscope warm up least 30 minutes, do self cal and of course nothing connected to any input BNC's.
Yes, I read the instructions and performed the self-calibration after the update as described.
But anyway, I don't have an issue with the accuracy of the measured samples. The voltage readings seem ok on the measure/statistics display and on the grid. It is the cursor display and don't see the technical reason why the cursors should be affected by the self-calibration or by the accuracy of the analog front-end since they are just a display feature for reading plotted values.
After you note that you do not at all mean signal accuracy and you just mean cursor positions error I download your images and change image gamma so that I can clearly see your images display graticule on my monitor I can immediately see problem.
For look this problem there do not need any signal - this also keep thinking out from signal levels and focusing thinking just for cursors positions. Just enough bright display graticule and vertical (Y) cursors.
I have now seen it with my SDS1202X-E and I can confirm, there is problem in this FW 5.1.3.13
I will later try test more about it. This error do not exists with all setups. Need more investigations.
I will later continue more my tests and I will report it to Siglent after I have done test enough for sufficiently extensive documentation.
This is clearly FW error and it need repair in next possible FW where this fix can schedule. This is clearly FW bug and it need repair in next possible FW where this fix is possible to schedule.
Thank you for the good finding and the emergence of it
This error math/logic is quite simple and clear now.Error is that if channel offset is other than 0V and probe multiplier other than 1X there is error in Y1 and Y2 levels. Cursors Y Delta is not affected.
In FW 5.1.3.13 changelog we can see they have touched cursors.
1. Cursor values not correct if probe attenuation was not set to 1X
7. Corrected cursor measurements for active Zoom with FFT
10. Fixed decode threshold levels for 10X probe attenuation selection.
Somehow they have forget one thing - unbelievable mistake. They have forget channel offset voltage correction with probe multiplier for calculate cursors vertical levels.
Cursors have positions relative to display center line. If signal offset is 0 then vertical level in center line is 0V.
For cursors Y1 and Y2 voltage values relative to signal zero there need subtract vertical offset value from Y values for show correct Y1 and Y2 values. Delta Y value is not affected, it looks like always ok.
Error is now that when probe multiplier is what ever other than 1X they still subtract vertical offset value from Y1 and Y2 as it is if probe multiplier is 1X. Example:
1.
if set scope for 1X probe mult, vertical V/div 10mV and user set 0mV vertical offset (vert. center). Then cursors are set for Y2 +30mV and Y1 -30mV. Cursors display Delta Y 60mV and Y1 -30mV and Y2 +30mV. All is ok.
All is ok also if now change probe multiplier from 0.1X to 10000X. All is ok with this error only because vertical offset is zero. Subtracting 0 from Y1 and Y2 values do not show error. As long as Vertical offset is 0 it do not show this error what is now inside FW.
2.
Continue with same settings. Keep other settings untouched but change only probe multiplier to value 10X
Now cursors Delta Y display 600mV and Y2 300mV and Y1 -300mV. Just as told also in example 1.
Then, do not touch any other settings but adjust
channel Vertical position one div up to value +100mV.
Cursor Delta Y display now 600mV and this is ok.
Cursor Y2 display 290mV. Cursor Y1 display -310mV and this is wrong. Wrong because added offset is still as it is if probe multiplier is 1X.(in this example 2. offset is 10mV if probe mult is 1X)
This error continue without exceptions with all vertical offset and probe multiplier values.
It can try also using under 1X probe multiplier and result is fun but still follow exactly this explanation about error math/logic.